Conceptual Rook Endgames

By

Jacob Aagaard

Quality Chess
www. qualitychess. co . uk
 First edition 2023 by Quality Chess UK Ltd

Copyright © 2023 Jacob Aagaard

CONCEPTUAL ROOK ENDGAMES

All rights reserved. No part of this publication may be reproduced, stored
in a retrieval system or transmitted in any form or by any means, electronic,
electrostatic, magnetic tape, photocopying, recording or otherwise, without
prior permission of the publisher.

Paperback ISBN 978-1-78483-194-3

Hardcover ISBN 978-1-78483-195-0

All sales or enquiries should be directed to Quality Chess UK Ltd,

Suite 247, Central Chambers, 1 1 Bothwell Street,
Glasgow G2 6LY, United Kingdom
Phone +44 141 204 2073
e-mail: info@qualitychess.co.uk
website: www.qualitychess.co.uk

Distributed in North and South America by National Book Network

Distributed in Rest of the World by Quality Chess UK Ltd through

Sunrise Handicrafts, ul. Szarugi 61, 21-002 Marysin, Poland

Typeset by Jacob Aagaard

Proofreading by Paul Brencher, Andrew Greet & John Shaw
Edited by Jacob Aagaard
Cover design by Kallia Kleisarchaki

Printed in Estonia byTallinna Raamatutriikikoja LLC

 Contents
Foreword by Karsten Muller 4
Preface 6
Introduction 8

1 Activate the Rook First 27

2 The Rook belongs behind the Passed Pawn 39
3 Shouldering 59
4 Rook vs Pawns 79
5 Passed Pawns 103
6 Promoting Passed Pawns 115
7 The Umbrella 131
8 Breakthrough 135
9 Check the Checks! 143
10 Checking Distance 163
11 Defence from the Front 171
12 Cutting off the King 175
13 The Kings Role 189
14 King Activity 207
15 Checkmate 217
16 Stalemate 229
17 Zugzwang 239
18 Mutual Zugzwang 249
19 Tempo Wars 263
20 Anticipation 273
21 Connected Passed Pawns 293
22 Active or Passive Defence 315
23 Slow Play 333
24 Lost Items 345
25 Unbelievable 359

Epilogue - Concepts to Live by 407

Name Index 410
Foreword by Karsten Muller
Practice makes perfect

Rook endings are the type of theoretical ending which occur most often by far. They are worth
studying as there are many positions which occur over the board regularly. Philidor’s draw,
Lucena’s win and Vancura’s draw are just the three most prominent examples. So there is already
a vast literature on the subject. Why add two more books to the collection?

Rook endings can be regarded as having two aspects. One theoretical and one practical. So Quality
Chess decided to cover each aspect in a separate book. Sam Shankland has a systematic style as
he has proven in his excellent books on pawn play {Small Steps to Giant Improvement and Small
Steps 2 Success).So he was a great choice for the theoretical work, the aptly named Theoretical
Rook Endgames. Equally, Jacob Aagaard’s creative genius, as demonstrated most recently in his
A Matter of Endgame Technique, makes him an ideal choice for the sister work on practical rook
endings, Conceptual Rook Endgames. Whilst the study of each book will be equally valuable to the
practical player, it would be preferable for the reader to start by obtaining a full understanding of
the theoretical aspects of rook endings from Sam’s book. These provide the fundamental building
blocks to the practical aspects of such endings as demonstrated in Jacob’s book, which provides
more advanced material.

In Conceptual Rook Endgames, Jacob Aagaard investigates certain important guidelines, such as
passed pawns should be pushed and the defender should exchange pawns. Numerous motifs are
also examined, one important example being zugzwang. It is well-known that this is a weapon
generally used by the attacker. Readers may be less familiar with the fascinating concept of
mutual zugzwang, where the side to play is in for a disappointment. This arises surprisingly often
in practice. These guidelines and motifs and many others are explained in detail with many well-
chosen examples. These examples clearly demonstrate that the real art of chess is not knowing the
guidelines. It is to develop an intuition for the application of those guidelines and the recognition
of when exceptions apply to them.

A particular feature of the work is the way in which it explores the difference between calculating
concrete lines and schematic thinking. It also examines when to use one or the other of these
techniques. This is an important practical skill to develop, both in rook endings and in chess more
widely. Jacob explains it in depth with many further excellent examples.

The book also benefits from the source of the examples chosen by Jacob. Several of the games
derive from the daily classes at his online academy, www.killerchesstraining.com. The beauty of
this is that the analysis has benefited from the input during those classes of a large number of
 Foreword by Karsten Muller

strong players, without the help of engine assistance. This brings a rich human element to the
understanding of these examples (complemented of course by the view of the silicon monster).
Furthermore, Jacob has included many games from very recent grandmaster practice. These prove
both that rook endings often occur in practice, and that they are difficult to successfully navigate
over the board. So make yourself familiar with the important concepts!

I hope that Jacob’s work will give you as much pleasure as it has given me.

GM Dr Karsten Muller
Hamburg, April 2023
Preface
Rook endings have traditionally been looked at from the perspective of fixed positions with a
few ideas on top. Rarely have the ideas been the main lens through which to look at this part
of the endgame. A few writers have done it, particularly Edmar Mednis and Mark Dvoretsky,
even though Mark also chose to look at the ideas as an afterthought. The ambitious concept of
looking at the most common endgame in chess from two directions at the same time - with
Theoretical Rook Endgames by Sam Shankland, and this book, which focuses on repeating ideas -
is perhaps a novelty in chess literature. It is destined that there will be a difference in opinion on
which book has the best approach. Be certain that those opinions will be strong and phrased in
absolutes. However, it is our opinion that looking at a topic from several angles is the best way
to approach it: to understand both the theoretical positions and to some extent memorise them
and the key ideas, and to understand the general themes and ideas and work on applying them.

This is not going to be a simple journey. From understanding an explanation to consistently

using it in your own games, there are quite a few stops. Understanding is not knowledge.
And knowledge is not skill. You will recognise the themes in games and exercises you see
and analyse, and play many games where they occur, before they become second nature.

All authors would love to say: read this book and you will play the rook endgame perfectly! But it
would be a lie. All I can promise is that you will be able to see and understand patterns and that
this will help you improve your decision making in rook endgames. Hopefully this will help you
at the moment of greatest importance!

A small note on the style of this book:

This book has two different identities that I have tried to make co-exist.

O n the one hand it is a part of the Grandmaster Knowledge series, which means that nothing is
dumbed down or skipped in order to make the book more accessible.

At the same time, this is a book about ideas. Ideas can be represented by variations, but are more
commonly supported by visuals, words and abstract concepts. Thus, variations have been kept
to a minimum whenever possible. Another word on the role of variations is that they are often
illustrative, and do not necessarily represent the only path forward, although sometimes they do,
but can also be the most logical and coherent path.
 Preface 7

The language of modern chess has changed from the days where games were analysed by hand and
conclusions were hard to come by and up for debate. In those days a lot of annotation markers (!,
? and so on) were given on the basis of if a move improved the position or made it worse. Today,
symbols are often used to indicate a clear change in the computer’s evaluation of the position.
I have chosen to follow something closer to the outdated approach, where symbols are used
to show if a position has become more difficult to play, or if the player has solved problems or
managed to put problems for his opponent. Thus, a question mark does not have to change the
evaluation of the position according to the computer, but simply indicates reduced chances of
a favourable result in a practical game between two humans; most commonly the players in the
game annotated.

Jacob Aagaard
Copenhagen, April 2023
Introduction
Structure

The basics are called the basics, not because they are simple or easy, as the word “basic” is occasionally
used, but a reference to foundational knowledge that anyone with aspirations in a given field will
need to master, in order to progress smoothly and get the most out of their talent and effort. In our
work on understanding rook endings conceptually, we will look at both basic ideas, and powerful
ideas/patterns/concepts/themes that are not necessarily foundational to understand additional
ideas.

No apologies

Like other books in the Grandmaster Knowledge series, this book is unapologetic when it comes
to complexity. I have not sought complexity and I have actively pruned all variation trees back as
far as it could be done without losing something essential. The ideas are the key points I want to
present, but I am not afraid of complexity, should it prove instructive or fascinating. This means
that I have often ignored a lot of false avenues that have been suggested in training with strong
players or at our academy, Killer Chess Training.

Rook endings come towards the end of the game and not in a vacuum. The players are tired and
have already gone through an emotional journey. I n this book you will see a lot of strong players
making mistakes. Noting that mistakes were made is not a negative commentary on the players’
abilities, but rather a commentary on the complexity of the game and focused on conveying an
understanding of how rook endgames work.

Many famous games and matches were decided in the rook endgame. The most famous rook
endgames in chess history are probably Alekhine’s win against Capablanca in their 1927 match,
Botvinnik holding the draw against Fischer in their 1 962 Olympiad game and Korchnoi’s win
against Karpov in their 1978 match. But there are also a lot of famous rook endings from the 21st
century. Magnus Carlsen has won a number of remarkable rook endings, but the most famous
rook ending of all time will probably have to be Gelfand’s botched conversion in the 2nd play-off
game in the 2 0 1 2 World Championship match in Moscow, where an unexplainable move (and
I have asked!) ruined the Israeli’s dream of the world title. Even more recently, teenage superstar
Alireza Firouzja won the following remarkable game in the 2021 Grand Swiss, on his way to
winning the tournament and qualifying for the 2022 Candidates tournament.
 Introduction 9

Alireza Firouzja - Evgeniy Najer This is where we will start our serious look at
this game. We are about to transition from the
Riga 2021
middlegame into the endgame.
l.e4 e5 2. 13 3.d4 xe4 4.dxe5 d5
30.&h2?
5. bd2 £lxd2 6.txd2 l,e7 7.c3 c5 8.1,d3
Only this is a serious mistake.
£lc6 9.0-0 JLg4 10.S e i ®d7 l l . h 3 th5
12.U4 ®e6 13.JLe2 0-0 14.®d2 JLg6?!
30. b4! would be less forcing and make up for
Releasing tension for no particular reason.
the mistake of a few moves earlier. If Black
White was better anyway, but Black still
plays forcingly, we can see the big difference to
needed to get the pieces into the game.
the game. 30...§xd4 31.®xd4 ®xd4f 32.§xd4
g5 33.Sd7 b5 34.Sd6 Sxe5 35.Sxa6 White
15.Sadl JLe4 16.£lg5 JLxg5 17.JLxg5
is winning. We shall see a position similar to
White has the advantage of the bishop pair,
this in the game, except with the pawn on b2,
as well as more active pieces.
which makes all the difference.

17...d4 18. fl ®g6 19.®f4 tc2 2O.Sd2

3O...Sxd4 31.®xd4 ®xd4 32.Sxd4 g5
Sae8 21.JU14 &h8 22.13
33.Sd7 b5 34.Sd6 Sxe5 35.Sxa6 1x41!
22.jkb5!? was stronger according to the
The big difference. Now White has to take.
computer, but it is not immediately clear to
these human eyes what the purpose of the
36.JLxc4 bxc4
move is.

22...a6 23.<t£2 Sd8 24.Scl JLbl 25.cxd4

xd4 26.JLxd4 cxd4 27.a3 JLa2 28.Sxd4?!
Entering unnecessary complications.
28. b4! would give White an improved
version of the game.

28...®b6 29.Scdl Sfe8

abcdefgh
We now see that the b2-pawn is a bad
backward pawn and that we are heading for a
typical endgame where the extra pawn is rarely
enough to win.

37.a4 Ec5!
There are other ways to hold, but Najer is
abcdefgh an experienced player, steering towards the
endgame he knows draws.
10 Conceptual Rook Endgames

38.Sb6 c3 39.bxc3 Sxc3 4O.a5 &g7 41.a6 point out that there seems to be a decline of
Sa3 42.&g3 Sa2 most players’ abilities in their 40s. Najer is
44 years old. His best years are behind him.
What appears to have happened in both this
game and the game against Caruana is that he
became fatigued. Remember: before the 4-5
hours of play that we see, there are 1-3 hours
of looking over your preparation and perhaps
finding something new to play. It is a full
working day, in a stressful job.

50.. .fl=®??
The explanation of this mistake can only
be psychological. Najer obviously made a
misevaluation somewhere, but where is not
abcdefgh clear.
43.f4l?
An important part of being a strong chess 50...§b2t!? was unnecessary, but draws.
player is to make great practical decisions. 51.Sxb2 f l = ® 5 2 . § b 7 t (52.a8=® ®f8t
In this case, Firouzja manages to put to his 53.&b7 ®f3t is an immediate draw.) 52...&g6
opponent some practical problems for the cost 5 3 . S b 6 t (Or 53.a8=® ®f8t 54.£a7 ®a3t
of only a pawn. with a draw.) 53...&g7 54.a8=® ®f8t Black
draws. 55.&b7 ®e7t 56.& b 8 ®f8f Black
43...gxf4f 44.&xf4 Sxg2 45. &e5 Sa2 draws.
46.&d6 £5!
The only move. Black needs counterplay. 50... 6.6! (or h6 or f6) was however the most
natural move in the position. What did Najer
47.&c7 f4 48.&b8 B 49.a7 £2 5O.Sbl miss in this line? 51.a8=® (5 l . S f l is not an
improvement. After 51...Sb2f, it is a draw.)
5 1 ...Sxa8t 52.&xa8 <±>g5 5 3 .Sf 1 <±>h4 54.Sxf2
&xh3 Is it possible Najer somehow believed
the rook would still be on the 1st rank here?
Without asking someone who surely is deeply
disappointed what he was thinking, there is no
way to know.

51.Sxfl Sb2f 52.&a8 Sb3 53.Scl &g6

53...Sxh3 5 4 . § b l Sa3 55.Sb5! is the
classical way to win such a position. The
h-pawn is going nowhere.

abcde fgh 54.Sc7 h5 55.Sb7 Sxh3 56.Sb6f &g5

Perhaps the most famous moment of the 57.&b7 Sa3 58.Sa6 Sb3t 59.&c7
tournament. At this point it makes sense to 1-0
 Introduction 11

The following endgame was from the same 60.. .g5?

tournament on the same day, six boards down. Sarana misevaluates that the pawn on h6
It is a big brutal fight with mind-boggling would be less weak than the pawn on g7, or at
complexities. It cannot be diminished to least that it would matter less. This is wrong, as
verbal ideas; variations are needed to show the after the h6-pawn falls, the h5-pawn becomes
mechanisms that govern this endgame. a passed pawn. This matters a lot in a race.

David Navara - Alexey Sarana 60...6.7 was more prudent. 61.&d4 (61.§b7t
&d6 62.§xg7 §xa5 63.Sg6 &c5! would give
Riga 2021
Black active counterplay.) 61...e5t 62.&c5
c3 63.&b6 (63.§b7t &e6 achieves nothing.
If 64.&b5??, Black has 64...§c4! 65.&xc4 c2,
when the c-pawn cannot be stopped.) 63...Sc4
64.§bl c2 65.Scl §c3 Creating a safe distance.
66.a6 Sb3t The rook goes to b2/a2, forcing a
draw.

61.h5! &e7 62.&d4 &d7 63.Sb7t &e8

64.Sa7
I have long been speculating on where the
rook is best placed in situations like this one:
beside the pawn or in front of the pawn. In
abcdefgh most cases it is probably in front, as the ability
to advance the pawn disappears if the rook
White is pressing, but Black should be is next to it. But it depends a lot on which
able to hold with accurate play. This is an row the pawn has reached. The further up
improvement book. The goal of writing it is to the board, the more the advantages of being
help people play the rook endgame better and beside the pawn come into play. Maybe from
win, and thus enjoy chess more. Thus, I have the 6th rank onwards. It really is a complicated
tried to structure the book in a way so that one question to answer! In this position, the rook
idea follows another and they build on top is better placed in front.
of each other, slowly expanding the readers’
understanding of this fascinating part of the 64.. .e5t
game. But it would be a mistake to believe
that rook endings are all about the basics.
Rook endings can be as complex, beautiful
and perplexing as anything in chess. I hold a
profound affection for the aesthetics of rook
endings and this book is about that too.

59.d?c3 Sa4 6O.h4!

The traditional positional move. The plan is
to play h4-h5 and fix the g7-pawn. Although
this is the strongest play, it is not enough to
force an advantage.

abcdefgh
12 Conceptual Rook Endgames

65.&c3?
Navara commits a miscalculation and the
position tips back over the edge from winning
to barely holdable for Black.

The winning line was beautiful, but not

especially difficult:
65.&d5! c3 66.&e6
66.Sc7 <±>d8 67.Sxc3 <±>d7 68.Sc5 Sd4#
66...<±>d8 67.Sd7t <±>c8 68.Sdl Sxa5

abcdefgh
66...&g7?
Sarana walks straight into the zugzwang,
which will be explained below.

The way to avoid it was:

66...<±>f7l!
abcdefgh Only this will make the draw.
67. g3!
The only moment of resistance.
The trickiest try.
69.<±>f7!l
67. a6 always fails if Black is not put in
The target is the h6-pawn. You can see the
zugzwang. Below we shall see the importance
ghost of 60...g5? at this moment.
of the Sa6-a7-a8 manoeuvre.
69.&xf6? §c5 is an immediate draw. White
67...6e7I!
cannot get the king to d6 and cannot attack
Another only move.
e5 and h6 at the same time, meaning that the
67. . .6e6? fails to 68.a6 g4 69.a7 gxf3
rook can defend the c-pawn and whichever
70.§e8t, showing why the king belongs on
pawn White attacks.
e7.
69.<±>f5? Sa6 70.<±>g6 f5t’! 71.&xf5 Sa5
67... g4 68.fxg4 &g7 needs to be refuted
draws in exactly the same way.
by bringing the rook to the 6th rank, thus
69...Sa6 70.<±>g7
showing why pushing the a-pawn too soon
White takes the h6-pawn and wins.
is mistaken. 69.Sa6 &f7 70.g5! fxg5 71.g4
&g7 72.§g6t &h7 73. a6 White wins.
65...&A8! 66.Sa8t
68.a6
After 66.§h7 §xa5 67.&xc4 &g8 68.§xh6
68. g4 &e6! Only now the king can go to this
&g7 69.Sg6t &f7 the rook will never escape
square. 69.a6 &f7 70.a7 &g7 With another
from the corner.
draw.
 Introduction 13

72.§f5!
Forcing the rook to a poor square.
72...§c5 73.Sf7t <±>g8 74.§f6 <±>g7 75.Sg6t
<±>h7 76.Sa6 <±>g7 77.Sa3 Sc7 78.<±>b4
White is about to win the c-pawn. Even here
it is not obvious to all that White is winning,
although I doubt either of the players would
have been in doubt.

abcdefgh
68.. .g4!! 69.fxg4
69.a7? gxf3 would win for Black now. There
is no check on e8.
69...617
This is stunningly a draw. White cannot get
the rook to the 6th rank, which would be the
way to make something of the advantage.
abcdefgh
67.Sa7t«
67. a6? would also fail, as it did in the game 67...&g8 68.Sa8t &g7?
(see below). Black will not go for 67...&f7? We can see here that neither player
68.a7 &g7 69.g4!, and Black is in zugzwang. understand the mutual zugzwang issue.

There were a number of convoluted wins, but 68...6f7! was right.

all of them are parodies of:
69.&B2 &f770.Sa7t?
67.g4! 70. g4! and we are back on the horse.
Black is in zugzwang again. 70.. .6g7 71.&c3 £f7 72.a6 &g7 73.a7 with
67...<±>(7 68. a6 <±>g7 69.a7 zugzwang.
Black has to make a serious concession.
69...6h7 7O.Sf8 Sxa7 71.Sxf6 Sc7 70.. .6g8?
7O...‘i?f8! gives the flexibility Black needs.

71.Sa6?
< >
White would win after 71.Sa8t g7
(71...&f7 72.g4! is back to where we were)
72.g31. This waiting move works when White
has not advanced the a-pawn, and has the
option to play Sa6, as we saw above.
14 Conceptual Rook Endgames

71...&f7 72.Sa8&g7? Black is obviously under obligation to

72...Bb4f 73. &c3 §a4 would make the produce a move. If he moves the king, White
draw. will play Sf8 and take on f6, with a winning
position. But Black here has an additional
option.

75...g4! 76.frg4
White has no alternative to taking the free
pawn, something that will ruin his pawn
structure.

76.. .6h7 77.SA8

White has no way to make progress after
77.g3 &g7 78.g5 feg5 79.g4 <±?h7-

77.. .5xa7 78.Sxffi

abcdefgh
73.a6?
Navara misses the last chance to win the
game.

73.g3! <±>h7
73. . .6f7 74. g4! and we are back on the right
track.
74.&c3 g4
74... 6g7 75.g4 Black is in zugzwang again.
75.fxg4 <±>g7 76.Sa6! <±>F7 77.g5!
We have already seen how this wins above.

73...&h7 74.£c3 &g7 75.a7

abcdefgh
78...Sd7I!
The saving move.

Perhaps Navara was counting on 78...§g7

79.&xc4 Sxg4 80.&d5, when White wins.

79.Sc6
79.Sf5 Sd3t! 80.<±>xc4 Sd4f and Black
holds.

79...Sd3t 80.&b4 Ee3 81.Sxc4

White is two pawns up, but will not create
real problems for Black.
abcdefgh
 Introduction 15

81...&g7 82.&b5 83.Sc6f &g5 that the engine would not sacrifice a piece for
84.Sg6t < >h4 85.Sxh6 &xg4 86.He6 Sxe4 compensation, leading to a devastating loss in
87.h6 &g5 88.h7 3h4 89.Sxe5t £<6 9O.Se3 19 moves.
Hxh7 91.Sh3
Vi-Vi In 1 997 I also won a game against John Emms
based on computer preparation. Since then,
About using engines engines have become stronger and stronger
and there have since 2005 been fears of engine-
Young players today are often referred to as supported cheating at the highest level. How
the “computer-generation”, which both makes founded in reality these have been in over the
sense and is a bit puzzling. board chess is hard to say; but I gather it is
limited to lower levels (like my own normal
I personally learned to play chess in 1979, GM-level in some high-profile cases), mainly
but after losing to my mother - and crying - with using phones in the toilet.
the set was packed away and it was not until
1984 I got into the game again. I played with The biggest influence on chess from computers
a friend at school and briefly there was a chess has been through opening preparation, but also,
club in the school, although this only lasted a the ability to analyse and understand games.
few sessions. My obsession with the game soon Yusupov once told me that it was obvious that
become full-blooded and I saved up to buy a there would be so many young grandmasters,
chess-playing computer; one of those with real as the information they were receiving was
pieces and blinking lights indicating the moves much higher quality than the information he
the computer wanted to do. My first few years had received when he was young. But using
in chess was playing against a computer and engines is a skill that also has to be learned.
slowly progressing to the point where I could And more so as the computers have become
beat the 8th and top level on the machine, stronger. Especially with Stockfish 15 there
which I guess was rated between 1200 and was a shift, when most evaluations in the
1400. Which is the point. The computers were endgame are either +156 or 0.00. When you
bad. are analysing your games with an engine, it is
impossible to see the difference between moves
In the 1990 World Championship match, they that equalise (meaning that both players have
still played with adjournments. It was briefly the same chances in a practical game - 0.00)
rumoured that Kasparov had had assistance and positions that hold with only the most
from the Deep Blue team in finding the win, amazing continuations (where no one will ever
which the IBM team said was untrue - and the hold them - also 0.00).
machine was not able to find the win anyway.
In this book I have tried to use the computer
In 1996 Deep Blue lost 4-2 to Kasparov in in this way: I accept with gratitude all the
a high-profile match, where the computer information it has given me, but without
honestly looked feeble. It returned in 1997 believing for one moment that I would have
to win a highly controversial match, where been able to understand or find most of it on
Kasparov resigned a position where he had my own. I believe it is professional to ensure
been handed a chance to draw. Also, Kasparov that the chess in the book is correct and well
played a highly risky bluff in Game 6, hoping analysed; but at the same time also remember
16 Conceptual Rook Endgames

that chess is a game played by humans, with 44...Sa8!

limitations. Therefore, you will see moves with A good practical decision.
question marks and exclamation marks that
do not change the evaluation according to 45.Seb3 Sxa5 46.Sxb7 Sa7 47.Sxd7t Sxd7
the computer. My goal is to help you play the 48.&e3
endgame better; and to do this, I have tried to
explain ideas, strategies and themes that will be
useful when making decisions without digital
interference.

The following example between a seasoned

English Grandmaster and the biggest Scottish
talent in a generation is a good example of this.
I had a training session with Freddy after the
game. We had both looked at the game with
a computer, but we came to wildly different
conclusions.

Mark Hebden - Frederick Waidhausen Gordon abcdefgh

Internet 2022 48...<&e7
48.. T5? looks tempting, but loses to a
beautiful sequence. 49.Sc3! fxe4 50.§xc6 Sa7
51.Sc3! The rook is not allowed to come to a3
(and d3). 51...Se7 After the rook has occupied
this passive spot, White regroups the rook to
f4. 52.Sc2 Se8 53.Sf2f <±>e6 54.<±>xe4 <±>d6t
55.<±>d3 Sg8 56.Sf6t <±>e7 57.Sa6 With many
moves remaining, Stockfish gives White an
1 1.49 advantage...

49.Sb8 Sa7 5O.Sg8 Sa3t 51.£f4 &f7

52.3c8 g5t?
A bad decision, which leads to Black having
abcdefgh two weak pawns and an awful responsibility to
defend them.
42.. .dxe4?I
I do not like this move. I prefer improving
52...§a6 with passive defence would still have
the king and keeping the tension with:
held the game with a minimum of discomfort.
42...6e6! and 43...&d6.

53.hxg5 frg5t 54.&xg5 Sxg3t 55.&f4

43.fee4Hd7 44.a5
To me, Black is under a bit of pressure. To
Stockfish it is all 0.00.
 Introduction 17

abcdefgh
58...Sg6!
abcdefgh Also a difficult move, not found in the first
55...Sg4t? attempt.
A simple blunder, after which the game is Freddy once again tried 58...§h6?, but after
lost. looking at 59.d 5 for a bit, he realised that
White is winning. The pawn comes to d6
Freddy said that “If I only had played: and the king goes to c6. The rook on h6 is
55...5g6?l entirely out of the game.
I would have made the draw.” “OK,” I said, 59.d5 cxd5 6O.exd5 Sg5t!>
“show me” and proceeded to put the position 60...&f8! is a Philidor draw. The text move is
on the board. First off, I gave him a check. less practical, but still interesting to analyse.
56.Sc7t 61.&e6
Freddy immediately went to the wrong Another challenge. This time Black has two
square (f8), losing. drawing moves; neither suggested by Freddy.
56...<±>g8!
56... 6f8? 57.Sh7 We played around a few
moves before he accepted that it was hopeless.
Having changed his move, I gave Freddy the
problem of finding the next move.
57.&f5 Sd6!
Freddy chose 57...Sh6, when after 58.e5!,
Black loses the h-pawn. One line goes:
58...h4 59. <±>g5 h3 60.<±>xh6 h2 61.<±>g6!
and White wins.
58.<±>e5 abcdefgh
61...Sg31!
The natural thing is to put the rook on g l ,
but in the main line below, this would lose.
So, the rook is best placed here.
Also 61...Sg6t?! 62.&f5 Sa6 63.&e5 Sa5!
64.&e6 Sa3! 65.d6 Se3t 66. d7 §d3!
holds, which is still not easy to determine
for a grandmaster.
18 Conceptual Rook Endgames

62.d6 Se3t 63.<±>d7 Sd3! 64.Sc5 h4 This book is many things. It is built around
We now see that the pawn would be simple ideas that can be understood by all and
defended on the 3rd rank. To see more about replicated. The foundation is deeply analysed
this sort of thinking, go to the chapter on games, hopefully without mistakes, but
Anticipation (page 273). realistically, with few mistakes. Although a lot
65.<±>c7 &f7 66.d7 <±>e7 67.Se5t <±T6 68.Sh5 of the material in this book is advanced, the
Sc3t 69.<±>d8 h3 concepts explained are not. Nothing would
Black holds. be more complicated than navigating only
by calculation. By using these themes and
Having seen that 55...Sg6 holds on the showing how they work with carefully chosen
computer is not enough. We need to examples, I hope to give the reader an innate
understand why it holds - not just know understanding of how rook endgames work.
the moves. We need to understand if it is There are beautiful and amazing games and
repeatable. On which foundations it relies. variations in this book. I love rook endings.
The themes and ideas. And a good dose of They are no less beautiful to me than the
realism concerning our own abilities to find or most inspired attacking games. Still, a lot of
remember moves. variations have been omitted or cut short,
to save trees, but also to spare the reader
Incidentally, I prefer the other drawing line unnecessary headaches. I know from working
Black has: 55...Sc3! with students that you can always put in more
detail and everyone would want something
else covered, but are indifferent to the rest. In
the case of this book, I have tried to remove
everything that was not a part of the story of
the game, or the point I wanted to make. I
hope what remains is a clarity concerning the
themes and concepts.

General Concepts

Before we go into the ideas and concepts of

abcdefgh rook endings, I want to briefly mention a few
When after 56.d 5 &e7!! 57.dxc6 &d6 Black concepts which I believe anyone reading an
will be able to draw against the sole e-pawn advanced book on rook endings is already to
with reasonable ease. some extent familiar with, but which it makes
sense to remind you of.
56.&f5 Sgl 57.Sc7t &e8 58.Sxc6
White wins. Endgames are mostly about promoting pawns.

58...E4 59.Sh6 S h i 60.&e6 61.d5 &g7 This may seem obvious, but it is often forgotten.
62.Sfi6 h3 63.e5 Sei 64.SE2 &g6 65.Sh2 The opening is often a fight for space, activity
Se3 66.d6 &g5 67.d7 Sd3 68.Sxh3 Sxh3 and structure, while the middlegame is often
69.d8=®t &f4 70.®ffit a fight either for technical advantages (like
1-0 winning material or inflicting damage to the
 Introduction 19

opponents structure), or based on dynamics, The defender frequently

mainly mating attacks. In the opening and wants to exchange pawns
middlegame, the way you win games is either
to deliver checkmate, or collect advantages I wrote a longer chapter about exchanges in the
to exploit in the endgame; broadly speaking. endgame in A Matter of Endgame Technique,
Meanwhile almost all endgames are decided which often relates to multi-piece endings and
by the advance and promotion of pawns. This other material distributions. One foundational
leads to a few logical conclusions: idea is to give up your minor piece for the last
pawn and draw automatically, a piece down.
Passed pawns must be pushed Obviously, this does not work in rook endings.
But we frequently see that exchanging pawns
Such is the old adage. Obviously, there are benefits the defender nonetheless. Especially
some limitations. Again, intelligence on behalf in situations where the defender loses the rook
of the receiver of such advice is expected. The and tries to win the rook back with a passed
further up the board a passed pawn is, the more pawn. In such situations, it is not good to have
valuable it is. If it is on the 5th rank, it has other liabilities (see Mazur - Malaniuk, page
potential. If on the sixth, it requires attention. 93). Or if there are situations with multiple
If it is on the 7th rank, it dominates the position weaknesses, where it would be good not to lose
(unless effectively blocked, as we shall see in too many pawns in the first example in this
many examples - but this still demands the full introduction, as well as Navara - Bartel, page
time attention of a piece). Of course, a passed 364 and others, where you will see this theme.
pawn can be pushed too far, but in the about
1000 games I have analysed in preparation for Schematic thinking
this book, this is rare and often for extreme
reasons, while a world class grandmaster not Schematic thinking is the process of thinking
pushing the passer is something that frequently backwards from a goal position, rather than
catches my eye. Throughout this book we shall thinking ahead in traditional “I go there,
look at many themes related to pawns, passed he goes there” fashion. It can be used when
pawns, promoting passed pawns and so on, we know where we are headed (could be a
because queening pawns is the common path number of places; Swedish endgame legend
to success in endgames. Ulf Andersson would think up 15 fortresses,
before you had realised it was your move).
While I am generally sceptical about evaluating
positions without having a concrete solution,
it can be incredibly useful to think about
which positions would be good for us, try to
work them out in advance, and to use logic to
do so. Another way to think schematically is to
see what our problems are, find ways to solve
them without variations, and only then apply
variations.
20 Conceptual Rook Endgames

Alexander Baburin - Sam Collins based on his games and experiences. I gave it as
an exercise to both Boris and Sam Shankland,
Armstrong 2018
and both struggled. Sam took an hour
before he solved it, with some half-hearted
suggestions from time to time. Boris solved it
in a day, returning to it for ten minutes here
and there, before he had to deal with kids,
life and other stuff again. But impressive was
GM John Shaw, the chief editor at Quality
Chess, who solved it in ten minutes, while
reading his morning emails and sipping
coffee. Whether solving is generally easier with
coffee and emails is an experiment for another
time. What I believed worked was John’s less
concrete approach (partly based on lack of
abcdefgh skill) of thinking logically and schematically.
But before we discuss John’s thinking, let’s see
In the first example, White is in desperate how other suggestions have failed. The starting
trouble, being a full pawn down. Baburin, sequence is so logical that we can call it forced.
still Ireland’s only grandmaster at the time
of writing, lost to the man who will likely be Gilberto Hernandez - Boris Gelfand
Ireland’s next grandmaster, seemingly without
a fight. Merida (variation) 2003

48.Hc6t? ±d5 49.Bc5t &e4 5O.Hc7 SBt

51.&g2 Bxa3 52.Bb7 Sb3 53.±£2 h4
54.Bxe7t ±xd4 55.Bh7 h3 56.±e2 Hc3
0-1

Thinking about the position logically, we can

see that White has two major issues. The first
is that he is a pawn down. The second is the
weakness of the a3-pawn. Ideally he would like
to exchange it, but S a l and a3-a4 is met with
...b4, while a3-a4 immediately will just lose
the pawn. It is not too great a leap from there
to find: abcdefgh
63...Bc5
48.Sbl! d6 49.Sb4! and 50.a4!, with a The ideal placement is behind the passed
holdable ending. pawn (see Chapter 2).

I found the following perfect example of 64.Bc7 Bc3t

schematic thinking when working on Decision For some reason everyone wants to win the
Making in Major Piece Endings, a book tempo; but as White is not doing anything,
ghost- and co-written with Boris Gelfand, this is not really necessary.
 Introduction 21

65.±£2 Bc4 66.±g3 69...Sc8 7O.Se3t &d6 71.Sf3 Sc7 72.&g3

The black rook is optimally placed, aiming &xc6 73.Sf6t & d 5 74.&f4
at both white pawns and tying down both the Black cannot easily make progress. White
rook and the king. The next step is to improve holds.
the king and choose the best pawn structure. To
be able to do this, it makes sense to understand 68.Hc8 ±6
what we are heading for. More below! Black is ready to check on the 3rd rank and
bring in the king. Sadly, White has a defence.
The most common suggestion here is to
remove the pawns from the 7th rank first. 69.c7! Bc3t 7O.±g2 ±g4

66. ..g6 67.&h3

White can at least wait.

abcdefgh
71.Sg8!
White holds in this theoretical endgame.
abcdefgh Especially if he has read the excellent
67.. .16 explanation by Sam Shankland in Theoretical
The most logical continuation. Rook Endgames.

Before coming upon the strongest Returning to John’s schematic thinking, there
continuation, Sam tried regrouping the rook - were a few elements, all constructed from a
a bit out of desperation. This also gives White base knowledge of the endgame theory. First of
a chance to regroup his rook, which could be all, he wanted to avoid the theoretically drawn
better placed. endgame with the f- and h-pawn. So, holding
on to his g-pawn was essential. Secondly,
67...Sa4 68.Sd7! Sa8 he knew that the endgame with the h- and
68...Sa6 makes more sense, but after 69. Sc7 g-pawn advancing was winning. So, looking at
Sa8 70.&g3 Ef8 71.£f4, or 7O...Se8 71.£f4 the white defensive idea, he understood that if
Se7 72.Sb7!, White is doing fine. The king his pieces were differently placed, the position
is coming to g5 next, unless Black blunders was winning easily. Essentially, he wanted to
and allows 73. c7!, winning. play ...f6 and ...g5 without having to worry
69.Sd3! about the g-pawn. And he did not want to
Activating the rook. play ...g6 first, for reasons we have seen above.
22 Conceptual Rook Endgames

Understanding that White is only waiting, he 72.. .±h7!

improved his position to the maximum before Knowing your endgame theory is helpful.
acting. The king needs to leave the g7-square for the
rook. This did not have to happen on this
66...±fi6! 67.<£h3 &g6 move, but knowing it means we can do it right
away.

73.c7 Bc3t 74.±g2 h4 75.&gl g4 76.Sa8

White can of course wait and hope Black
will put the king on g7, but it is unlikely to
happen.

76.. .Bxc7 77.Ha4 Hg7

Black wins.

If I had shown you this example without telling

you that two top players had spent an hour
each finding the winning continuation, would
abcdefgh you have seen it as entirely obvious? I think
68. ±g3 f6 you would. Logic has that effect on people;
Played without dropping or moving the once we understand it, we cannot emotionally
g-pawn. return to the emotional place where we did not
understand it.
69.&h3
White can only wait. The last game of this introduction is once
The key point with keeping the pawn on g7 again one of the most remarkable rook endings
is that after 69.Sc8 &f5! 70.cl Sc3t 7 1 . £ g 2 I have seen - as well as played between Navara
g4 72.Sg8 Sxc7 the rook defends the pawn. and Sarana. This time David won, but only
after remarkable events.
69...±h6 7O.±g3 g5 71.hxg5t feg5 72.Hc8
David Navara - Alexey Sarana

Belgrade 2022

l.d4 £if6 2.c4 e6 3 . & B d5 4. c3 tb4

5.®a4f c6 6.e3 0-0 7.®c2 Be8 8.Jkd2
JkfS 9.a3 e5 10.dxe5 xe5 ll.cxd5 xf3t
12.gxf3 xd5 13.h4 xc3 14.Jkxc3 ®d5
15.®a4 td7 16.®f4 JU6 17.tc4 txf4
18.1xd5 c6 19.te4 th6?
I presume this was meant as an ambitious
decision, as otherwise it is just plain stupid
(and grandmasters are usually only stupid on
an advanced, not a plain, level).
abcdefgh
 Introduction 23

The foundational rule on exchanges is that you

want to exchange the opponent’s good pieces
for your inferior ones. In this case, this means
19.. J e5 and Black would have been fine.

20.0-0-0 Had8 21.h5

Black is under some pressure.

21.. .JU6 22.Bxd8 Hxd8 23.f4 6 24.tc2

td5 25.Sgl te4 26.tb3t td5 27.txd5t
Bxd5

abcdefgh
Our key moment of interest. What is the
first thing you notice? For me it is the distant
passed pawn Black would have in a pawn
ending on both sides of the board and by
default win all pawn endings. The second I see
is that the pawn ending will not happen. White
is planning to take on g7 and f5. We can look
ahead and see that White is quite happy to
lose both the f2-pawn and b2-pawn in return
for taking the f-pawn, as the connected pawns
in the centre would be powerful. Especially if
abcdefgh they are supported by the king. The outlook is
This need not be a critical moment of the bleak and we have to start being creative.
game, as White could have improved his
position with 28. c2 and slowly advancing. 3O...±h8?
The bishop on h6 is still stupid and White is An understandable move, which sadly
better, but if it is enough to win is far from happens to be losing.
clear. Black has a plan of ... 17, ...Sd7 and
...g6, to bring the bishop back in the game. The question that could have saved the game
White can delay this by putting his rook on was: how will White actually win this position
h 1 , but then it will also be sort of stuck there. if Black does nothing? He will obviously want
It is certainly possible for White to develop to take on g7 and f5, but we are not forced
active plans in that direction too. But on the to give up the f-pawn. So, the white king will
other hand, he has a tempting option now that have to come up through the centre. But this
he must either take or forever let go. Navara can be blocked:
decided to take it.
3O...c5!!31.hxg7
28.txg7I? txg7 29.h6 Hd7 3O.Hg5 The natural continuation.
31.&c2 is met with 3 1 ...b5.
If White plays 31.b4, Black will take on
b4 and create a passed pawn with ...b6 and
24 Conceptual Rook Endgames

...a5, creating enough counterplay on the 31.hxg7t Sxg7 32.Bxf5 Bg2 33.Bf7?
queenside. An attempt to run through with An innocuous mistake, allowing a truly
the king in the centre would see the pawn unbelievable defensive opportunity. White had
running and supported by ...Sf7-a7 (or a lot of winning continuations. I like 33. f3!,
...Sf6-a6), forcing the white rook to go to a l which is exactly what Navara did when faced
to stop the pawn. Then the h-pawn would with this position again on move 37.
become a major asset and White will feel
lucky he is not losing.
31.. W ! 32. &d2 b5 33.&d3 a6

abcdefgh
abcdefgh
33...Hg7?
Black has set up a fortress. The rook will be
Black would also lose after 33... h5? 34.Sf5
able to oscillate between 17 and f6 and if the Sh2 35. &d2 Exf2t 36.&d3 Sh2 37. e4 and we
white king ever makes it to g3, Black will be know what is happening.
able to play ...h6 and either win the pawn
ending, or reply to Sg5-h5 with ...Sf6 (if on
The strongest continuation was:
17) and ... xg7. White can also attempt to 33.. .6g8!! 34.Sxb7
break through in the centre with the king,
34 .Sf5 Sxf2 is a wildly different situation
but it holds considerable risk and little
than the game. Black is no worse. The white
chance of reward.
king is passive and the black king is in the
34.b4!? c4f 35.&d4 Sd7t 36.&e5 c3 37.Sgl
game. If White goes e3-e4, Black has ...h5
c2 38.Scl Sd2 39. f3 &xg7 40.&xf5
with equality.
Black will draw, unless White blunders.
34.. .h5 35. d l !
41.e4?
The best try.
The king needed to go to the g-file.
A big part of me wants 35.Sxa7 to be the
41...gg2
main line, as it holds the key idea of the
Now it cannot. Black wins.
defence. But it is so short and perfect!
35...h4 36.Sa5
3O...Sf7? does not work. White will be able to
quickly invade through the d-file, as he does
not have to spend a move on hxg7 just yet.
 Introduction 25

Generally, in this book I have avoided

long variations, but here we will make an
exception and give the variation till the end.
39.gg4 h3 4O.gg3 &h6 41.f6 gxb2 42.gxh3t
£g6 43.Sf3 <&£7 44.e4 S b 5 45.&d2 Se5
46.&e3 Se6 47.&d4 Sxf6 48.gxf6f &xf6
49.a4 e6

abcdefgh
36...gg7!! Rook behind the passed pawn,
which you will see a lot of soon enough and
which was the one concept you definitely
knew about already. 37.gh5 gh7 38.gg5t
Sg7 With a draw by repetition. Ironically,
this position would be winning for White, if
the king was on b l , as gc5-cl-hl would be
abcdefgh
too many extra pawns.
35...h4 36.Sb8t 5O.a5 &d6 51.a6 &e6 52.&c5 &d7 53.e5
e6 54.<&xc6 <&xe5 55. b7 d6 56. xa7
&c7 57.£a8 &c8 58.a7 &c7
Stalemate!

34.gffit Hg8 35.HI7 gg7 36.HB

No draw!

36...Sg2 37.B! g£2 38.Sg5 Hxf3

Black cannot allow Sg5-g3 followed by the
f-pawn advancing.

abcdefgh 39.&d2 H£2t 4O.±d3

36.. .£h7!!
36...<&g7? would be a mistake. The f-pawn
will quickly give a check (after 37. e2!),
winning a tempo that will force the king off
the defence of the h-file.
37.f5!?
Trying anyway.
37. e2 h3 would leave White with only
38. gb7t h6 39.gb8 with a repetition of
moves.
37.. .Exf2 38.gb4 gh2

abcdefgh
26 Conceptual Rook Endgames

4O...h6
40...5xb2 41.e4 would be totally winning
for White, as the black king is cut off and the
connected passed pawns will pound down the
board with tremendous force.

41.Hgl a5 42.&e4 Hxb2 43.6 b5 44.K S£2

45.S$?e5 b4 46.axb4 axb4 47.e4 b3 48. S$?e6
b2 49.e5 c5 5O.±f7 Hg2 5 1 . B h l &h7
51...5h2 52.Sxh2!

52.e6 c4 53.e7 Be2 54.e8=® Hxe8 55.&xe8

c3 56.f7 c2

abcdefgh
57.Sxh6t!
Queening the pawn also wins, but this is
simpler.
1-0

You will see versions of the ideas covered in

this introduction throughout the book. They
are foundational for all endings and important
to not just know and understand, but to feel
in your bones.
Chapter 1

Activate the Rook First

A sound endgame idea that has been about for around a hundred years is that with few pieces on
the board, it is important to get the most out of your most powerful piece. In the rook endgame,
this is the rook (the rook can dominate the king, the king cannot dominate the rook), which is
why both Mednis and Dvoretsky were talking warmly about the idea of “activate the rook first”,
which is probably the smartest thing said about rook endings. The first thing you should look at
in a rook ending is how to get your rook to work more.

Before we go to concrete examples, we should probably define active and passive. Active means:
attacking, threatening, proactive, with more choice and ability to dictate events. Passive means:
defending, protecting, reactive, with less choice and freedom. Active is the first choice of all, but
at times things are worth defending or defending against. In a lot of rook endings, it is worth
giving up a pawn in order to get active. While in others it would be a horrible mistake. We shall
see more examples of it being clever in this book, but this does not mean that you should freely
throw your pawns into the fire. Rather, you should look for active options, even when they cost
material, and consider if the outcome of the variations you see is better than the alternatives.

Avital Boruchovsky - Erez Kupervaser

Israel 2022

abcdefgh
Black could have held the draw in two ways.
28 Conceptual Rook Endgames

The first would have been 49...Sg6! followed 51.h5! would have secured an important
by ...Sg4 and the activity of the rook would tempo down the road. Black has no activity
make the rest of the game easy for Black. of his own and can either choose to seek
a different avenue with the rook, or wait. If
Another would be 49...Sf8, with the idea he waits, White wins because the h-pawn is
50. e3 d6 51.Se5 Sc8! with the idea further advanced in the coming race: 5 1 ...Sb3
...c5-c4, making the draw with a considerably 52.Se5t &d6 53. &e4
smaller margin. In both cases, the draw was
secured by the rook doing all the lifting.

49.. .5a6? 50.&e3! Sa3

Following up on the previous move is
the logical move, offering White the most
challenges.

50...5g6 would come too late. After 5 1 .Se5t

d6 52. e4 White has managed to push the
black king backwards, making it too late for
abcdefgh
Black to seek activity.
Threatening Se5-d5t followed by e4-e5,
and Sd5-d6f. Active play comes too late. The
following variation is a logical way for White
to play. 53...Eb4f 54.£f5 Sd4 55.Se6f &d7
56.Sxh6 Sxd3 57.Sa6 c4 58.h6 Sh3 59.h7 c3

abcdefgh
51.6?
Jumping the gun! One of the fundamentals abcdefgh
of endgame play is to understand when your
6O.Sa3 d6 61.Sxc3 Sxh7 62.Se3 White
opponent is unable to improve his position
wins, as the black rook is too close to be able to
and you thus should be strengthening yours to
give checks from the front (see Defence from
the maximum before you take the game from
the Front on page 171).
the build-up phase to a time of consequences.
 Chapter 1 - Activate the Rook First 29

51...Sa6 52.±f4 Sal 53.±g4 Vincent Keymer - Nodirbek Abdusattorov

Internet 2022

abcdefgh
53...SA? abcdef gh
We will talk about putting the rook behind
Black is in desperate trouble and about to
the passed pawn later. Here it is done a bit too
lose a pawn. In the game the Rapid World
automatically, as all it achieves is to attract the
Champion of the time had limited time to find
white rook to the same position.
the way to activate his rook.
Black had the chance to block the f-pawn in
56...Sc3?
an oddly convenient way from the f6-square.
Trying to place the rook in an active place,
53...Ha6! White is unable to make significant
but failing.
progress. And if White plays 54.&f4, Black
can keep the rook active with 54...Hal!. Black
57.Exb4Ed3
will draw after something like 55.He8 Hflf
56.<&g4 < ?d4! 57.Hd8t < ?e5! and although
White can still pose Black some problems, the 8
draw is not far away.
7
54.Hf4 gglf 55.±h5 ±e5 56.K! 6
Black resigned. 5
1-0
4
3
2
1
abcdefgh
58.Eb5?
A logical-looking move, but this misses the
chance to win an important tempo.
30 Conceptual Rook Endgames

58.Hb6! would have threatened e4-e5. After

58...&f7 59. b4 Hb3 60. b5 White is winning.
He will play 61.Hb8 and advance the b-pawn.
Black can scoop up the b-pawn, but either
the white rook will ravish the kingside, or a
pawn ending will occur, where the black king
is desperately far away on the queenside.

abcdefgh
64.h5!
The simplest way to exploit the material
advantage is to push the passed pawn and head
for a winning pawn endgame.

64.Hxf4? would be greedy and misplaced.

64...<&e5 65.Hg4 f5! would allow Black to
abcdefgh escape with a narrow draw in a theoretical
58...Sc3? endgame.
Passive play, giving White another chance.
64.. .5b3
Abdusattorov was given a second chance to White concluded the game in convincing
activate the rook with: 58...Hd4! 59.&f3 Other style.
concessions are also concessions (see what I did
there?). 59...Hd3t 6O. xf4 Hh3 and Black is 65. h6 66.e5 &e6 67.Sh5 Sb8 68.h7
back within the drawing margin for reasons Sh8 69.exft> &xf6 70.&B <&g6 71.£g4 B
elaborated below. 72.Sh3 &g7 73.Sh2 &g6 74.£xB &g7
75.&g4 &g6 76.Sh3 77.Sh6t &f7
59.SB6 ±f7 6O.Sb7t &e6I? 78.&g5 &g7 79.f4 &f7 8O.Shl &g7 8 1 . 6
Trying to activate the king, but this comes 1-0
at a price.
60...<&g6 61. b4 did not look appealing If we return to the starting position, we may
either. The black king is out of play. spot that a combination of tactics and strong
dedication to keeping the rook active would be
61.Sb5! Sd3 62.Sxh5 Sxb3 63.SB Sc3 enough to make the draw. But only just!
 Chapter 1 — Activate the Rook First 31

56...BI! 62.e6
62.exf6 Ef4f followed by another check and
...Hf4(t), the moment the white king goes to
the d- or h-file, as well as f l . White may be
able to push the pawn to b5, but thereafter he
will be unable to make progress without giving
up one of his passed pawns, which defies the
purpose.

62...Sf4t 63.±g2 Se4 64.b4

abcdefgh
57.&g3
White is not interested in exchanging pawns.

57...Scl! 58.Sxb4 Sglf 59.&xB g g 4

Black wins the h-pawn and although he is a
pawn down, he has reasons to be happy. His
rook is active and soon he will have a passed
pawn of his own with which he can create
counterplay.
abcdefgh
64...&g5!
60. b6 Sxh4! Black will play ...f5 and ...< ?f6, securing the
Black is relying fully on activity. draw after probably a good deal more moves.
The star behind this result would be the active
61.e5 &B! black rook.

The next example, from the World Rapid

Championship, was presumably played with
limited time on the clock. White chose to play
passively and lost without a fight.

abcdefgh
32 Conceptual Rook Endgames

Maxime Vachier-Lagrave - 79...Sxg6 80.&xg6 £4 81.&15 13 82.&e4 £2

Rustam Kasimdzhanov 83.Ehl &g2 84.&e3 fl=®
0-1
Berlin (rapid) 2015

Activating the rook would have made the draw.

61.Sa2!
Black can try a number of things, but none
of them appear as critical as...

61...Sg2
...although Whites path to a draw is quite
simple.

62.Sa7t
The king cannot go to the queenside. White
abcdefgh takes the f7-pawn and also picks up the black
h-pawn, after which Black will be on the
61.Shl? &d7 62.&d4 &c6 63.Sclf &b5 defensive.
64.Sblf &a4 65.&c4 &a3 66.Shl &b2
67.&c5 Sd3 68.&c4 SB 69.&c5 &c3 63.Sa8f &g7 64.&B
70.&d6 <£d4 71.15 gxf5 72.&e7 &xe5
73.&xf7 Sa3 74.Self &f4 75.g6 Sa7t
75... h2 76. g7 Ha7t was perhaps a bit more
accurate, but the game is good enough.

76.<&xe6 h2 77-Shl Sa6t 78.&17 &g3

79.Sel

abcdefgh
Black cannot avoid losing the h-pawn, if he
wants to get the king out of the corner.

64...Sb2 65.&g3 h2 66.Sal

White picks up the h-pawn in good time
abcdefgh before the black king can enter the game.
 Chapter 1 - Activate the Rook First 33

Brandon Jacobson - Christopher Yoo As in the previous example, Black was able to
give up a pawn in return for activating his rook,
Charlotte 2019
reaching a simplified version of the endgame,
where the black king would control the white
d-pawn and the b-pawn would control the
c-pawn.

69...Sd8! 7O.±xfi6 e4
After 7O...He8? 71.£17! Ha8 72.&e7, Black
would lose the fight for the d6-square and have
to deactivate the rook in order not to lose on
the spot, bringing into question why he gave
up a pawn in the first place.

71.±e5 Sa8!
abcdefgh
This fight between two young American
IMs, both close to achieving the grandmaster
title, ended abruptly with a transition into a
pawn ending in White s favour.

69...±b7? 7O.a5 ±c7 71.Sxb6 Sxb6

72.axb6f &b7
Black has no moves, once he sees: 72...&xb6

abcdefgh
With the exchange of pawns, now or in a
few moves, Black will reach an endgame where
his active rook will be enough to hold an easy
draw.

abcdefgh
73. d6 c6 74. &e6 and the pawn queens.

73.c5 &a6 74.d6

1-0
34 Conceptual Rook Endgames

The next two games are more complex, and the

role of active and passive rooks is crucial.

The first sees the great Russian Grandmaster

Rakhmanov outplay an IM in the open
Russian Championship.

Bogdan Belyakov - Aleksandr Rakhmanov

Sochi 2020
abcdefgh
8 White would lose after normal play such as
7 47.&e4 f7, but can make a draw based on
rook activity with: 47. e3! &f7 48.Hd4! < ?g6
6 49.He4 with a coming exchange of pawns.
5
46.g6t &g7 47.Se4 ±fi6 48.Sc4 Se5t
4
49.&B c5 5O.Sg4 ±g7
3
2
1
abcdefgh
44...Sa5! 45.Sg4?I
A natural move, which could have been the
first step towards losing the game.

45.Hc4! would have been stronger, harassing

the c-pawn. After 45...&d6 46.Hg4 <&e5 47. g6
Ha8, active play with the rook leads to the
exchange of pawns needed to secure the draw. abcdefgh
48.Hc4 c5 49.Hg4 &f5 50.ggl! Hg8 51.Hbl
with a draw on the horizon. White is under a lot of pressure and fails to
find the only move.
45...±f7I?
Another way to make life difficult for White 51.Sgl?
would be: 45...He5t!? 46.&f3 Hf5t 51.Hg3!! This move looks so weird, it can
damage your iris and pupil if you stare at it too
long. The point becomes clear after: 5 1 ...Hd5
52.&e2 Hd4 53.Hb3 Hb4
 Chapter 1 - Activate the Rook First 35

65...e2!? 66.Ke6 &f2 67.&xa5 c3 would also

win.

66.&xc4

abcdef gh
We can see that the king is not in the way
of the rook remaining active with 54.He3!»
holding the draw. This would not work with
the rook shuffling on the 1st rank, where the
white king would be in the way.

51...Sd5 52.&e4 Sd6 53.2g3 c4 54.Sc3 abcdefgh

Sc6 55.&d4 66...Se4f!
See Chapter 9, page 143, on the importance
of checks in rook endgames. Here the white
king is pushed back, as the other choices of
losing the a-pawn, or blocking the retreat
route for the rook, are unattractive.

66.. .e2? 67.®bl would allow White to escape.

67.±d3 e2 68.Sg6t
White could have tried 68.Hbl!?, when
Black needs to win a tempo getting the king
closer to the queenside, where the final race
focused on catching the white a-pawn will take
abcdefgh place.
55...&xg6
Good enough to win, but making it a bit of
a task.

A waiting move would have made the win

much easier. 5 5 . . .Hc5! would be a good choice.

56.Eg3t &f5 57.&c3 e5 58.Sf3f &e4

59.Sfl Sc5 6O.Ebl Sa5 61.Self &f3
62.&b4 e4 63.Sflf < >g2 64.Sft> e3 65.Hxb6
Se5 abcdefgh
36 Conceptual Rook Endgames

Here 68...e l =®? would throw away the win. Piotr Murdzia - Radoslaw Psyk
After 69.Hxel Hxel 70. c4! White is in time
Poland 2022
to make the draw. This sort of race we shall
look more at in Chapter 2.
Correct is: 68. .. f3! 69.Hal White loses a
tempo. (69.Bel would lose to 69...Hxa4, as
the e-pawn is defended by tactics.) 69...el=®
70.Hxel Hxel 71.&c4 < ?e4 72. a5 Hal Black is
in time and wins.

68...&B 69.Bfi6t &g4 7O.±xe4 el=®t

71. ±d5 ®a5t
Black won this theoretical ending, which is a
good deal more difficult to do against the best
defence than most would think. Personally, I
have won two out of the three times I have
abcdefgh
reached it.
Another situation where activating the rook
would have saved the draw - yet only with
precise play. There is a lot wrong with the
white position.

To start somewhere, he is down two pawns, the

b5-pawn is under threat, (although it cannot
be taken immediately, it would be catastrophic
for White to get forced to defend the pawn
passively from b l ) and the two connected
black pawns have a majestic potential. But
there are also some plusses.

abcdefgh First of all, the strong pawn on c6, which, if

even just for the moment, is safely guarded by
72.&e6 &g5 73.Sf8 ®b6f 74.&d7 ®b7t
the b 5 -pawn. The king is better placed than
75.&e6 Wc6t 76. >e7 <&g6 77.a5 ®c5t
that of the opponent too.
78.&e8 Wc7 79.a6 &g7 8O.Sfl ®c8f
81.&e7 Wc5t 82.<&d8 Bd6f 83.&e8 We6f
All in all, there is still time to act, if only it is
84.&d8 ®xa6 85.Scl Wa5t 86.&c8 <&f7
done now. Throughout this game we shall see
87.Sc6 &e7 88.&b7 &d7 89.Sb6 Bd5t
how White gets in trouble because he does not
90.&b8 ®e4 91.Sb5 &c6 92.Sb7 ®e8f
make his rook work up to its full potential.
93.&a7 ®d8 94.Sh7 Wa5t 95.&B8 ®b4f
96.&a7 ®a3t 97.&B8 Bb3t
54.<£g4?
0-1
Trying to activate the wrong piece!
The next example is a good illustration of how
big the drawing margin can be in rook endings.
 Chapter 1 - Activate the Rook First 37

54.Hal! 56...e4! 57.hxg6 d3! would exploit the

This was the only move. The b-pawn remains position of the white pieces to eliminate the
immune for the moment and if allowed, white queenside pawns. White is missing the
White will use this to his advantage and play X...Hxb5 Y.Hcl resource, due to Y...d2.
Ha6xb6 next, when Black’s extra pawn on
the kingside is too far back to play a decisive 57.hxg6?
role in events. 57.Hgl!
54...&f7 With the idea 58.&f4 or 58.&f3 with a gain
Stepping out of the 8th rank check of tempo, would have saved the game.
(followed by c6-c7-c8=® under the right 57...e4
circumstances). This remains the critical try.
55.Sa8! &e6 57. . .g5 58.< ?f5! with Hal-a7t on the cards
55...d4 56. &f3 would also hold. The king also gives White enough counterplay.
is meant to keep an eye on the black pawn 57. . .Hc3!? Trying to control the white
combo. king. 58.Hel gxh5t (58...&f6 59-Hflf
with repetition) 59.&f5 White has enough
counterplay to make the draw.
58.&f4 e3 59.Hxg6t £f7 6O.Hg2 &e7 6 1 . £ e 4
Hxb5 62.< ?xd4 < ?d6 63. xe3 Hxh5 64.Hg6f
<&c7 65.&d4 b5 66.Hf6 £b6 67.Hg6 Hh4f
68.&d5 Hhl 69.c7t &xc7 70.&c5 Hh5t
71.&b4 <&d7 72.Ha6 &e7 73.Hb6 £f7

abcdefgh
56.Hb8!
White wins the b6-pawn and secures a draw.
The key variation is:
56.. .d4 57.Hxb6 &d5 58.Hb8 £e4 59.He8 d3
60.<&f2!
The king comes down to block the pawns
and Black should take the draw before less is abcdefgh
offered. 74.&c3!
Keeping the king close enough to be able to
54.. .g6 55.h4 exchange rooks on h5, as well as to hold the
5 5.Hal no longer works, as White is lacking pawn-down rook ending.
the &f2-el manoeuvre at the end. 74.Ha6? would lose to 74. ..&g7 75.Ha7t
<&g6 76.Ha6t &f5 77.&xb5 Hh3!! 78.&c4 h5
55.. .d4 56.h5 &g7? 79.&d4 h4 8O.Ha8 Hf3 and the white king is
Caring about the wrong pawn. cut off for long enough to give Black the win.

57. ..±xg6 58.H£8 e4 59.Sd8

38 Conceptual Rook Endgames

abcdefgh
71...He8!
abcdefgh Sooner or later the rook has to focus on
59...±f6? pushing the strongest pawn.
Played to avoid an annoying check on d6, 72.gg6 b4 73.Hg4 b3 74.ggl
but loses an important tempo. 74.Hb4 Hb8 and Black wins.
74.. .6xc6 75.Hb 1 Hb8 76.&xe3 &c5 77.&d3
59...h5t! £b4
This was the winning move. The reason is Black wins by a tempo; not that the win was
simple: in the lines that follow, the pawn is close!
no longer hanging on h6.
6O. f 4 e 3 61.Hd6t 6O.±f4??
After 61.<&f3 Black’s only obligation is not Presumably short on time, White missed the
to panic. With 61...Hc4! he is able to keep random chance 6O.Hxd4, when the h6-pawn is
his two pawn advantage. More moves can be hanging in the key lines.
made, including mistakes, but with decent
play, Black should win. 60.. .e3 61. £3 &e5 62.Sb8 &d6 63.Sxb6
61...6.7 62.&f3 Losing fast, but White was going down
62.Hxd4 loses to 62...e2 63.He4 Hc4!. slowly otherwise.
62...Hxb5 63.Hxd4 Hc5
Black has achieved control, but converting 63...5Bt 64.±e2 65.±d3 Sd2f
the two pawn advantage can still go wrong, 66.±c4 e2 67.c7t &xc7 68.Se6 d3
as the e-pawn is closely defended by the 0-1
white king. The following line is not the
only win, but it makes a lot of sense.
64.Hd6 e7 65.®h6 He5! 66. e2 < ?d8
67.Sh7 b5 68.Hb7 h4 69.Sh7 h3 7O.Hxh3
&c7 71.Eh6
Chapter 2
The Rook belongs
behind the Passed Pawn
DrTarrasch’s most famous quote is that “all rook endings are drawn”, which must have been said
in jest and is low on content. While it is true that rook endings have a higher drawing tendency
than pawn or minor piece endings, they have a lower drawing tendency than opposite-coloured
bishop endings and queen endings. I am a far greater fan of his second most famous quote, that
“the rook belongs behind the passed pawn”.
Not only is this strategically sound advice, it is also economically phrased. Saying nothing
about attacking, defending or other circumstances, it manages to encompass all scenarios and
be on point at the same time. The advice is correct, as endgames are about promoting pawns,
no pawns are easier to promote than the passed ones and if the rook is to play a role, it needs to
cover as many squares on the pawns path as possible, without being in the way. And if you are
the defending rook, you do not want to see yourself squeezed by an advancing pawn, you would
rather want to see the opponent’s rook lose activity in front of the pawn, as we saw in some
examples in the previous chapter.

I. Krikheli
Magyar Sakkelet (no. 1742) 1986

abcdefgh
40 Conceptual Rook Endgames

1.5e7! 5...&d2 6.Sb7 &c3 7.&a5

White would not make any progress after White catches the pawn and wins.
l.&b7? &c3! 2.Se3t &c2 3.Se4 &c3. 1-0

1...6a3! Martin Jogstad - Ludvig Carlsson

I...&c3 2.Sc7t! &d3 3.Sb7 &c4 4.&c7 b4 Uppsala 2022
5.&b6! and the white king arrives on the right
side of the pawn. We will discuss this more in
the following chapter on “shouldering”.

2.Sa7t
Harassing the king, putting it in the path of
the pawn, thus winning time.

2.. .6b3
Notice that if there was a file to the left of
the a-file, White would still win. The key point
here is the tempo won.

3.&b7 b4 4.&a6!
abcdefgh
White is using shouldering again. If the
black king was to go to the a-file, White would After 37.Sb5? &f7 38.&E2 &e6 39.&xf3
play 5.&b5t and win a tempo. *&d5 40. <&e3 Sa4 the advantage was gone.
4.£l?b6? &c3! would allow Black to make the White played on for another 20 moves but got
draw. nowhere.

4.. .<&c2 If White had placed the rook behind the pawn
with 37.Sb3! instead, he would have won
8 effortlessly.

7 As stated, the basics are foundational, not

6 simple or easy.
Understanding the core idea of the basics,
5
such as “the rook belongs behind the passed
4 pawn”, is already a bit more than remembering
3 a catchphrase.
Chess is not a dogmatic game and we have to
2 understand why and when, which is done from
1 seeing a lot of examples. Several times in this
book we shall see the rook going behind the
abcdefgh passed pawn, instead of giving a check, being
5.Sc7t! a crucial mistake. Once we understand it, it is
An important trick to have in your bag. The still difficult to implement it into our games.
black king is moved away from the b3-square There are ideas in this book grandmasters
and White thus wins a tempo. will encounter for the first time, but every
 Chapter 2 — The Rook belongs behind the Passed Pawn 41

man and his dog knows the rook belongs 44.&c3?

behind the passed pawn! Still, implementing A passive move, which is rarely the way to
it requires imagination and attention to this go in rook endings. White was winning after
idea at surprising moments, in order to make 44. c5!? and some complications, but even
the most of it. In the next example, Black is a stronger was:
pawn down, but his pawns are more advanced
on the kingside and because the pawn is on h3, 44.§d6t! &f5
he is able to create counterplay. 44.. . £7 would be a big admission. White
can now play 45. c3, transfer the rook
Grzegorz Nasuta - Damian Lewtak to the 2nd rank and shoot the king to a5,
where it will threaten the a6-pawn, as well as
Czech Republic 2022
help the queenside pawns get going.
45.c5 Se2
45... g4 46. c6! Taking on g4 first would be
a massive mistake, as it would give Black
access to the g4-square. 46... f3 47.gxf3 gxf3
48.c7 Sc8 49.Sd8 Sxc7 5O.Sf8t White wins
in the long run.

abcdefgh
43...Se8?
Too ambitious. Black is trying to win the
g2-pawn and in this way make the most
out of the advanced f-pawn. But this is time abcdefgh
consuming and White could have exploited
46.Sd5t!
the extra time to push his pawn.
Pushing the pawn also wins, but as Black
would not want to give up the g-pawn, this
43. . .g4! 44.hxg4 (44.Sxh5?? loses to a basic
makes a lot of things easier.
breakthrough trick: 44...f3! 45.gxf3 g3!)
46...&f6 47.&d3 Sxg2 48.c6
44...5xg4! The threat of ...f3t wins a useful
White will follow up with Sd5-c5 and the
tempo. 45.&e4 Sxg2 46. xf4 Sxa2 With
black rook will have to fight to make it back to
the exchanges of pawns and the rook placed
c8, when he will have won back a pawn, but
perfectly behind the white pawns, Black will
have lost all quality of his position and soon
be able to make a draw with ease.
be faced with an advancing b-pawn, as well as
threats to his pawns on the 5th rank.

44.. .g4 45.hxg4

No silly excursions to h5 please.
42 Conceptual Rook Endgames

45...hxg4 46.c5

abcdefgh
5O.a4?
abcdefgh A natural-looking move and a mistake many
46...g3? grandmasters would make. Hopefully, this
This feels incredibly slow. Black is trying to book will save some of them from making it
balance two objectives: keeping the rook on in the future.
the 8th rank for defensive purposes, and to
create a passed pawn, either by removing the 50.&b3! with the idea Sd3-c3 was winning. It
g2-pawn, or by pushing ...f3 at an opportune is hard to come up with variations, but here is an
moment. But the lack of energy behind this attempt. 50...&e4!? 51.Sc3 & d 5 Black is using
move gives White a chance to regroup. the king to block the c-pawn, which is a key
strategy discussed later, in Chapter 13. 52.&a4
46...Sel! 47. &d2 Sgl would have given Black &c6 53.Sf3! Sf8 54.&a5 Sf5 55.a4 and White
sufficient counterplay to hold the draw. There will win slowly. At the right time, White will
are lines where both sides promote, and lines transfer the rook with S f l - d l ( o r cl), winning.
where neither do. It is visually obvious that
White is not making progress, while Black is. 5O...Sh2?
And the threat of coming behind the passed Black returns the favour.
pawn with a check is the key reason for this.
(After 47.Sd2 &e5 Black also holds. Whether A draw was available if Black had placed his
a queen or a pawn ending a pawn down, the rook behind the passed pawns. The main line
draw is not difficult.) is highly instructive and the main reason why
this example should be studied, but as you
47.Sd2I? may have noticed, not the only one.
47.Sd4 with the idea of Sc4 followed by an
advance of the queenside, while the king will 50...Shi! 51.b5 axb5 52.axb5 S e l f !
run to fl to stop the counterplay, was perhaps Black gives three annoying checks, forcing
the most principled winning method. the white king to where it would like to be
for offensive purposes, but far away from its
47...Sg8 48.Sd3 &e5 49.b4 Sh8 defensive duties. The draw is almost poetic
in its simplicity thereafter.
 Chapter 2 - The Rook belongs behind the Passed Pawn 43

53. &b4 S b l t ! 54.&a5 S a l t ’ 55.&b6 Sa2 Bartosz Socko - Daniel Sadzikowski

56.c6 Sxg2
Bydgoszcz 2021
The black pawns are not worse than the
white pawns, which is the reason why a draw
can be found in all lines. Like this one:
57.Sc3 Sh2 58.c7 Sh8 59.&a5 g2 6O.gcl f3
61.b6f2 62.b7

abcdefgh
White is completely winning, but also
abcdefgh completely out of time. All the remaining
For no reason at all, I am stopping here. moves by both players were made in 9 seconds
Stockfish says 0.00, while most of us would or less.
be entirely uncertain about everything at the
board. But chess is a practical game, where 35.a5?I
making good decisions is everything. And Not a major mistake, but allowing the
Black is happy to be here, in this total mess, black rook to come behind the passed pawn
and not in the game, where he lost without a needlessly, making life awkward for the rook
fight. on a7.

51.c6 Sxg2 52.c7 Sgl 53.&b2 &e4 54.c8=B The principled winning idea was to put the
55.Wc2f white rook behind the pawn instead: 35.&c3!
1-0 ge2 36.Sd7 Sxf2 37.Sd2! and 38.Sa2 was an
easy win.
The next example starts within our theme, but
then spins out of control in a direction that 35...5a4 36.a6?
could have worked well in Theoretical Rook Proceeding down the same path.
Endgames as well.
White was still winning after: 36.&c3 &h7
37. &b3 g a l , with the idea 38. &b4 ga2
39.gc7 gxf2 4O.gc3 gxg2 41.a6 ge2 42.a7
ge8 43. & b 5 and White wins by a few tempos.

36.. .£h7 37.&c3 &g6 38.&b3 gal

It was Dvoretsky’s counting that this type of
endgame has a drawing margin of two tempos.
44 Conceptual Rook Endgames

Meaning, you can lose one tempo and still

draw. Whether this is also the case with this
particular configuration I cannot say, but if so,
Black did manage to lose a tempo over the next
10 moves.

39.h4 Ba5 40.&b4 Sa2 41.<&b5 Sb2f

42.&c6 Sc2t 43.&b6 Sb2f 44.&c5 Sc2f
45.&b6 Sb2f 46.&c6 Sc2f 47.&b6 Sb2f
Black could have claimed a draw based on
three-fold repetition, which it is not easy to
spot in a rapid game where you do not write
down the moves. abcdefgh
54.Sb4t??
48.&c7 Time trouble causes insanity.

8 54.hxg5 fxg5 55.Sb5 would have won a lot of

tempos and later the game.
7
6 54...£g3 55.hxg5 fitg5 56.Sb2 g4 57.&b7
5 h4 58.&c6 h3 59.gxh3 gxh3 60.&d5 h2
61.Bbl &g2 62.&e4 hl=® 63.Bxhl &xhl
4
3
The rook is no less useful behind the opponents
2 pawns, as previously explained. The following
1 examples will show various situations where
the defending side missed the chance to make
abcdefgh the most out of Tarraschs advice.
48...Bx£2?
It is not wrong to take the f-pawn on
principle, but timing matters. The white rook
is well placed on a7 to assist the a-pawn.

48...Sc2f 49.&b8 Sa2! would force White to

misplace the rook with 5O.Sa8, if he was to
attempt to make progress, but now 5O...Sxf2
is a tempo faster and will make the draw.
We shall see this mistake (not checking the
king to a worse square) repeated frequently
throughout this book.

49.Sb7 Ea2 5O.a7 $6 51.&b8 g5 52.a8=®

Bxa8f SS.&xaS &g4
 Chapter 2 - The Rook belongs behind the Passed Pawn 45

Ivan Saric - Alojzije Jankovic 51...Sa4!

Black eliminates both white pawns and
Croatia 2014
makes the draw.

One of India’s seasoned grandmasters takes

one of their many young grandmaster talents
back to school with this beautiful display
of endgame technique after Black’s initial
mistake.

Abhijeet Gupta - Aditya Mittal

Kanpur 2022

abcdefgh
In the game, passive defence failed quickly:
49...Sb8? 50.a7 Sa8 51.Sa5! &g5 52.<&f3
&h4 5316 1-0

Better was:

49.. .5g4! 5O.a7

50.&f3 Sg8 51.a7 is different from the
game. Black can play 51...g2! and use the
doomed pawn to win a vital tempo.
abcdefgh
6O...Sdl?
50...g2 51.Sxg2 Going behind the white passers, but as the
You would expect Jankovic to have seen this black king is passive and Black is a pawn down,
during the game and despaired, missing: this alone will not be enough to defend.

61.a4 &e7 62.a5 c4 63.Sc6

The white rook is perfectly placed. It assists
the a-pawn and contains the c-pawn. White
won slowly and safely.

63.. .Sal 64.a6 &d7 65.Sh6 &c7 66.&d4

Sa3 67.15 SB 68.&e4 Sfl 69.Sh7t &b8
7O.Sb7t &c8 71.SI7 &b8 72.&d4 Sal
73.SI6 Scl 74.Sc6 Sfl 75.Sc5 &a7
76.&xc4 &xa6

abcdefgh
46 Conceptual Rook Endgames

To play like this, Black would have had to

find:

62...Sd8t!!
With the idea of ...Sd4f, securing the draw.

White can try a few things, but if they involve

the black rook on f4 cutting ofF the white king,
they are not too threatening. Please notice how
the two top principles of rook endings come
together here. The rook seeks to be behind the
passed pawn, but prefers activity to all else.
abcdefgh
Sometimes it can be difficult to apply the most
77.Se5! &b6 78.&d5 &c7 79.&e6 &d8
basic concepts, because they do not happen at
80.&f7 &d7 81.16 Sgl 82.Sd5t &c6
the edge of the board and they have dynamic
83.Sd2 Hfl 84.<&g7 Bglf 85.&AB Bfl 86.(7
components.
Bgl 87.Sd4 &c7 88.<&e7 Self 89.&16 Bflf
90.&g6 S£2 91.Sd5 Daniel Naroditsky - Jorden van Foreest
1-0
Internet 2021
Black should have aimed to put the rook
behind the passed pawn for counterplay:

6O...Sd8! 61.a4 Sc8

White would now seek to block the c-pawn
with the king.

62.<&d3!

abcdefgh
Van Foreest played the natural-looking
move, which would be the right move in 90%
of positions with this check on h6. But in this
case, the highest objective should be to get the
rook behind the passed pawn.

64...&b7? 65.&d3 Ba4 66.8 Sb4 67.&e3

abcdefgh Hxb6 68.Hxb6f &xb6 69.&f4 *&c5 70.&g4
1-0
 Chapter 2 - The Rook belongs behind the Passed Pawn 47

It was far stronger to keep the rook active, White lost this miserable-looking position
behind the passed pawn. As in previous after attempting a passive defence.
examples, giving a check is a big part of why
this works. 5O.Sal? &c5 51.Sa2 Sd4f 52.<&c3 Sh4
53.Sal Sh3t 54.£d2 &d4
64.. .6c5! 65.b7 Blacks domination is total. The rest of the
Another key point is that after 65.&d3 Sb4 moves were quite unnecessary, but more than
66.&e3, Black can just wait. White is not one person has dropped a rook with a mouse
threatening to do anything. But the only way slip...
to wait is to centralise the king. 66... &d5!
55.&e2 Sh2f 56.&B Sd2 57.2hl Sa2
65...5c4f! 58.Sh4f s£?d3 59.Sh5 &c4 6O.Sh4f d4
This check is crucial. 61.&e4 Se2f
61...Sxa4 was perfectly fine too.
66.&d3 Sb4 67.Sh7 &c6
67...&d5 68.&e3 h3 The h-pawn is a
62.&B Se8 63.Sh5 d3 64.Sh7 d2 65.Sc7t
distraction that will cause the b-pawn to fall,
&b3 66.Sd7 &c2 67.Sc7t £dl 68.Sd7
after which the game is immediately drawn.
&el
68.&e3 h3 69.8 &d5 70.&12 h2 71.&g2 0-1
Sb2f 72.&hl &e5
As said, when you are a pawn down, passive
Black holds.
defence is unlikely to work. Barring the few
In the next example we shall see the full power exceptional exceptions, it is better to seek
of an active rook, placed behind the passed active counterplay if you don’t have time to
pawns. Passive play in positions where we are think. And if you do have time to think, you
down material works so rarely that it becomes need to be sure that the passive defence works,
spectacular when it does. Here, too, it does not before choosing it instead of active defence.
work. 5O.Sb8! Sxa4 51.Sa8
51.Sd8f &c6 52.Sc8f &d7 53.Sa8 Sal
Vincent Keymer - Le Quang Liem
54.&d4 is similar.
Internet 2022
51...5al
After 51...&e5 52.Se8f, there is nowhere
for the black king to escape to.

52.&d4
Threatening Sa6t, winning the d-pawn.

52... 5 d l f 53.&e3!
53.&c3 S e l f and 54...Sc5 would allow
Black to reorganise the pieces and stabilise the
position.

53...5al 54.&d4!
abcdefgh
48 Conceptual Rook Endgames

67.. .5a5t? 68.&g6?

White missed a not too tricky win: 68.&f4!
Sa4f 69.&f3 Sa5 7O.Sd8f and it is all over.

68.. .Sa4
Black had the chance to avoid the queen
ending with: 68...Sa8!? 69.Se7 Sa7!, when
Whites only try is: 7O.Sxe2!? &xe2 71.h6,
when Black has to play a few accurate moves to
make the draw. 71...Sa6f! 72. & h 5 &f3 73. h7
S a l ! 74. g6 &f4 75.g7 &f5 With a well-known
theoretical draw.
abcdefgh
Black is unable to make progress. 69.&15
More challenging was 69. h6! Se4, which
The next example also comes from the Rapid leads to a queen and pawn vs queen ending
World Championship. Black was one of the that is objectively drawn, but still unpleasant
big winners of the tournament, as he was to defend.
the coach of Nodirbek Abdusattorov, when
he won the tournament and for that reason 69.. .5a5t? 70.&f6?
received a big cheque from the government of 70.&f4!
Uzbekistan.
Later he had a big score on Board 4 in the 70.. .5a6f 71.&g7 Sa8 72.Sxe2 &xe2 73.g6
2022 Olympiad as part of the sensational &£3 74.&h7 &g4 75.h6
victory by the Uzbek team. At this time the Vl-Vl
“elder” (read in his 20s) Uzbek national got
away a bit easier than he deserved... Black could have held a draw, without relying
on assistance from the opponent, with a
Milosz Szpar - Jakhongir Vakhidov beautiful accurate sequence.

Warsaw (rapid Wch) 2021 67.. .5h4! 68.h6 Shi

We notice that because of the rooks
placement behind the pawn, White is
temporarily stalled in his advance.

The queen ending after 68...Sh3?! 69.&g6

Se3 7O.Sxe3t &xe3 71.h7 e l = ® 72.h8=®
®b 11 is a draw, but since it is not immediate
perpetual, we should know that this ending is
terribly difficult to hold in practical play and is
best avoided.

abcdefgh
 Chapter 2 - The Rook belongs behind the Passed Pawn 49

69.&g6 Sean Kelly - Kushal Jani

69.g6 leads to a little image all should have
Internet 2021
installed on their “hard drive”. 69...Sh5f!
and Black is immediately fine. But amazingly,
he also draws with 69...Sxh6!? 70.g7 S h i !
71.Sxe2 S f l f 72.£e6 Sgl!.

69...el=® 7O.Sxel Sxel 71.&h7

A dangerous moment for Black. White is
threatening to advance the g-pawn with deadly
effect.

abcdefgh
Black played the most natural move and
survived by sheer luck.

68...£b3?
68...Sb8? also would not work against the
advancing pawns. 69.g5 Sh8 70.d?g3 &b3
71.&g4 &c4 72.&h5 &d5 73.g6 and the
abcdefgh black king is arriving too late.

71...Se7t!
69.&g3?
An important disruption.
These are the things that happen in blitz
games. White missed the immediate win with
72.&g6
69. h7! S b 8 70.g5 and we are heading for a
After 72.d?g8 <4’e4 the king arrives in time.
queen vs rook winning ending.

72.. .£e4 73.h7

69...&c4 70.&f4 &d5 71.&5 S b l 72.h7
73. *£6 Sh7! is not progress.
Shi 73.&g6 &e5 74.g5 &f4 75.&g7 &xg 5
76.h8=® Sxh8 77.&xh8
73.. .5e8 74.&f7 Sh8 75.g6 &£5
Vi— Vi
Black arrives just in time.
Black draws by getting the rook behind the
The next game was played in the blitz
pawns with an incredibly important idea that
championship on Killer Chess Training.
often seems to escape players. Obviously, at
the moment there is no space, so it has to be
created.
50 Conceptual Rook Endgames

68.. .5b2f!! 69.&g3 Bbl But Urkedal had not seen it before and should
White can keep blocking the rook’s access have lost at once.
to the g l / h l -squares, but only by repeating
moves. Prantik Roy - Frode Urkedal

Helsingor 2018
7O.g5 S h i !
The rook goes behind the pawns without
delay.

70... 5 g l f ? would lose a tempo and do nothing

else, except to show that Black is holding by
said tempo. 71.&f4 S h i 72.&f5 &b3 73.&g6
(73.g6? Sh5t=) 73...&c4

abcdefgh
49...b2?
I hope there is a small flash of unpleasantness
from you when seeing this move, remembering
how the rook loses the last bit of activity, as
abcdefgh seen a few times in Chapter 1 .
74. &h7! This zigzag movement is another
small image you should pay attention to. 5O.Sb7t?
74. ..&d5 75.g6 &e6 76. g7 White wins. White missed a direct win. 5O.f6f
51.Sb7t and the advancing f-pawn just wins.
71 J&g4 &b3 72.&B &c4 73.&g6 <&d5
The king arrives just in time. 50L.&e8 51.Sb8f &e7?
Aiming to repeat, but walking into a deadly
74.&h7 &e6 75.g6 Sgl 76.g7 f7 77.&h8 check. The king should have gone to one of the
Sg2 78.h7 Sxg7 light squares.
Stalemate.
51...&f7! is the most natural move. After
Once we have seen the above example, it 52.e6f &f6, White has no advantage. For
becomes possible to see the next one; which example: 53. &g2 (53.Sb7? S h i and Black
is part of the aim of this book: to provide you wins.) 53...&xf5 54.e7 S e i with a draw.
with a lot of snippets and ideas and images that
you can hopefully remember on some level, 52.Kf &e6 53.Sb6f!
when you are deep into your own rook ending. This is the difference. With the f-pawn
advancing, the black king cannot hide behind
the pawn.
 Chapter 2 - The Rook belongs behind the Passed Pawn 51

53...&f7 54.Sb7t &g6 55.f7 £g7 56.e6 52.. .h4! 53.&£2

Shi 57.SB8 After 53.&g2, the most natural is 53...Sb2f
1-0 54.&h3 Sf2, with a draw. But it should be
noticed that 53...Sei!? 54.Sxb3 Se5! 55.Sf3
You already know how this goes, although Se4 also holds. The white king cannot escape
there is a bit more to it than just giving the the box: 56.&f2 Se5 57.Sf4 h3! 58.&g3
check. Se3t 59.&h2 &e8!, for example, and Black is
holding a draw by the smallest margin.
49...Sb2f! 50.&B S b l 51.Sb7t
53...5b2f! 54.&e3 Sbl 55.&e2 Sb2f
56.&d3 S£2 57.&e4
8
7
6
5
4
3
2
1
abcdefgh

Spotting the difference between two moves

abcdefgh
is not always easy. 57...h3!
The h-pawn is a far stronger asset, as the
White would win after 51...&e8? 52.e6 h4 white rook cannot keep control of it from
(52...b2 53.&g2!) 53.&e2 Sb2f 54.&d3! behind as effectively.
S b l 55.f6, and the pawns cannot be stopped.
58.Sxb3
Also after 51...&d8?, the pawns cannot be 58.&e5? h2 59.&f6 will not lead to mate.
stopped. 52.e6 h4 53.<±>e2!Sb2f 54.&e3 S b l Black wins after 59...Sxf5t«
55.e7t &e8 56.f6 and there is no saving check
on f l . 58...h2 59.Sh3 &e7 6O.Sh6 &d6
White cannot make progress.
52.e6
After 52.f6 b2! Black is saved by having the The next idea is equally simple, but important
resources ...Sei or ...Sfl, since the king is on to have seen before, as you may otherwise miss
the 3rd rank. it.
52 Conceptual Rook Endgames

Daniel Hausrath - Volker Bub 6O.Sxc2f £xc2!

Usually, we need to take with the rook, to
Dortmund 2022
keep the king as close as possible. But because
of the geometry of the chess board, the black
king is not really further away from the action
on c2 than on d2, but it is vital that the rook
can get in behind the white pawns as quickly
as possible.

61.g5 Sh3 62.&f5 *£><13 63.&g6 &e4

64.1?h7 &£5 65.g6 &g5
Black makes the draw in the nick of time.

When you have a pawn on the 6th rank, you

rarely lose these pawn vs pawn positions, as
the pawn is so close to promotion it provides
abcdefgh counterplay. But it is still possible to make
Black lost quickly after: mistakes.

58.. .6d2? 59.Hxc2f Sxc2 6O.h7 Sc8 61.g5 Silvia-Raluca Sgircea - Daniel Alsina Leal
&d3 62.g6 Sf8f
Benasque Open 2021
Black resigned without waiting for the white
king to approach.
1-0

Black could have drawn the game by using

tactics to get the rook behind the white pawns.

58...5c3I!
Threatening ...&d2t.

59.&f4
59.&g2 is an obvious step in the wrong
direction, but Black still has to find a way to
exploit that. Simplest is 59...Sc6! winning a
abcdefgh
pawn, securing the draw.
My good friend Raluca had been defending
59.. . d2 well, but here misplaced her rook and lost.
59...Sc4t!? curiously also draws, although
57.Sa8?
it is a poor move technically. The line is still
57.Sh6! would make an effortless draw.
instructive. 60.&f5 Sc5t! 61.&g6 Sc6f
The rook will be able to fulfil the dual role of
6 2 . £ h 5 &d2 63.Sxc2f Sxc2! 64.g5 &e3
defending the a-pawn and preventing Black
65.g6 £f4 66.g7 &f5! 67.&h4 Sh2f 68.&g3
from promoting the g-pawn. It is hard to come
Sxh6 69.g8=® Sg6f with a draw.
up with a winning try for Black.
 Chapter 2 - The Rook belongs behind the Passed Pawn 53

57...Sa2f? rank. Also, with the king blocking the pawn, it

57...Sa5! with the idea 58. &e3 Sg5I, can be a bit surprising that the rook still has to
heading for g7, is a winning idea, as executed support the promotion of the pawn.
on the next move.
Jakhongir Vakhidov - Sam Shankland
58.&el?
Chennai Olympiad 2022
White had the chance to run with the king.

After: 58.<±>d3!Sa5 (58... fl 59.§f8f’)

abcdefgh abcdefgh
59.&c4 The king comes to assist the a-pawn, 48.&d7!
securing the draw. The key is that something 48.&b6? &e6 would lead to an immediate
like 59...Sf5 6O.Sd8 is a simple draw. And draw.
after 59...Sg5 60.a7 Sg7, White can defend
against the threat of ...Sf7 with 61.&b5! Sf7 48...Sxb2 49.& d8 a5 5O.d7 Sc2 51.Sh8 b4
62.&a6, freeing the rook to come out of the
corner.

58...§a5!
Regrouping the rook and winning.

59.a7 Se5t 60.&d2 Ee7 61.&dl Sh7

62.&e2 &h2 63.Bg8 Bxa7
0-1

In his aforementioned fabulous performance

at the 2022 Olympiad, Vakhidov missed out
on a win in a rook endgame, presumably
missing the chance to put the rook behind the abcdefgh
passed pawn. In this case this was a bit harder, 52.a4?
as the visual pattern is one of putting the rook The winning continuation went like this:
behind the pawn from afar (say the 1st rank),
and not right behind the pawn on the 6th 52.axb4 axb4 53.Sh6f ! &g5
54 Conceptual Rook Endgames

The only move that makes sense. The king Jon Ludvig Hammer - Vincent Keymer
goes for the g3-pawn.
Helsingor 2018
Online commentators were all focused on
53...&g7, which is a truly bizarre move,
but notably the first choice of Stockfish 15.
White wins immediately after 54.Sb6 Sc4
55. e7 Se4f 56.&d6 Sd4f 57.&c7.

abcdefgh
We start this rook endgame just out of the
opening, which has gone awfully wrong for
Black. Our theme will show up soon enough.
abcdefgh
21.Sc7t &e8
54.Sd6!!
Keymer understandably did not like the look
The key move, placing the rook behind the
of 21... g6 22.Sxd8 Sxd8t 23.&c3 where
passed pawn. Black is totally lost. White will
Black will have to attempt to defend passively,
play &e7xf7 on almost anything. The race is
with little chance of success.
lost by a mile.
However, 54.Sb6? would lead to a draw after
22.Sd3!
54...<±>g4!55.&e7 §c7 (or 55...§e2f) 56.&d6
Forcing Black to improve the position of the
Sxd7t 57.&xd7 &xg3, and the f-pawn will
white king.
secure the half point.

22.. .5xd3t 23.&xd3 Sd8f 24.&c4!

52...b3 53.Sh6f &g7 54.Sb6
The king might need to go to d5 urgently,
54.Sd6 b2 is a tempo-down version of the
thus this is the best move.
same draw.
24.. .5d2
54.. .b2 55.Sb5 &g6 56.&e7 Se2f 57.&d6
Black is forced to go on an adventure, as the
Sd2f 58.&c6 Sc2f 59.&b7 Sd2 60.&c7
pawn ending after 24...Sd7 25.Sxd7 &xd7
Vi-Vi
26. &d5 is obviously hopeless.
Combining the two main ideas that the rook
25.Sxb7 Sx£2 26. a4 a5
needs to be active and that it is best placed
behind the passed pawn, gives us the additional
idea of trying to prevent the opponent’s rook
from coming behind the passed pawn.
 Chapter 2 - The Rook belongs behind the Passed Pawn 55

26...Sxh2 27. b4 is hopeless. Everything 29.&b5 Sb2 30.&xa5 Sxb3 31.g4

is talking against Black. His pawns have not
exited the pit stop, his king is crashed into the
barrier, and the rook is so far away we could
consider it lapped. Leaving all the motorsport
metaphors behind, the point is that we can see
that White is going to win this race. Black has
to try something else, thus 26...a5.

abcdefgh
31...Se3?
Later in the book we shall discuss the role
of the king in rook endings (Chapter 13, page
189). Had Keymer read that chapter, it is likely
he would have found 3 1 ...&d8!, which should
draw in many ways. One of them is 32.Sxh7
abcdefgh Sg3!?, which is perhaps a bit too concrete, but
27.b3? which absolutely works.
27.§xg7? fails to 27...§c2f! 28.&b5 Sxh2!
29.&xa5 Sxb2 3O.Sxh7 f5 , when we will soon 32.Sxh7 Se5t?I
find ourselves in a basic drawn theoretical Often checks are able to ruin the
ending. coordination of the opponent’s pieces. But
they can also be helpful, which is the case here.
The winning idea was: 27. b4! axb4 28.&b3!! Black could have tried 32...Se4!?, when White
and the black rook cannot easily get behind wins narrowly after 33. &b5! &d8 34.Sg7!.
the passed pawn. A small key point is that a
check on f3 can be answered with &xb4, which 33.&b6 Se6f 34.&b7! Se5 35.Sh5! Se7t
explains why White needed to lose the pawn 36.&c6 Se6f 37.&c7 Se7t 38.&d6 Se4
first. 28...Sfl 29.Sxg7 With the ambition to 39.a5 Sxg4 40. a6 Sa4 41.Sh8f &£7 42.a7!
have extra material, as well as an overwhelming Finishing with a classic. Black resigns.
positional advantage. 29...S b I f 30.&a2 S d l 1-0
31.Sxh7 Sd3 32.&b2 Sa3 33.Sa7 White
wins. First the a-pawn is advanced, then the The final example is taken from a blitz game.
h-pawn. Often blitz games between strong grandmasters
will see them putting the pieces on good
27.. .5xh2 28.Sxg7 Sc2fl? squares, but they have no chance of picking
Although this should not lose, I prefer up the finesses and tactics of the positions.
28. . .h5! to liquidate pawns. The draw cannot If we remember that we are not evaluating
be far away. the players’ performance, there can be
56 Conceptual Rook Endgames

a lot of value in looking at high-quality later, but at the same time controlling the
blitz games. b2-square.
48.&a3!’
Jeffery Xiong - Wesley So 48.&c3? f4 49. b6 fee3 5O.fee3 Sh8 5 1 .&d4
Sd8f and Black holds.
St. Louis 2020
48..T4 49. b6
Black can try both active and passive defence.
I prefer active defence 8.5 out of 10 times.

abcdefgh abcdefgh
45.S18? a) Passive defence is too slow, but still requires
A wasted tempo. There is no reason why the some accuracy to overturn.
rook needs a checking distance in this position.
49...§h6 5O.Sb2 fke3 51.fee3 §h8 52.b7 § b 8
It will not surprise you that the winning idea 53.Sb3
was to bring the rook behind the passed pawn.
But I have to say that the winning sequence is
rather brilliant and precise.

45.Sc7! &f3 46.Sc2 Sh6 47.b5 Sh2

White is in control. But winning the position

still requires a bit of finesse. If the white king
storms up the board, the black king will
make it to c2, dislodging the white rook,
before returning to attacking the e-pawn.
So far, so good. The white king now has to The winning line for this reason looks like
find the right square to avoid being checked this:
 Chapter 2 — The Rook belongs behind the Passed Pawn 57

53.. . >e2!54.&b2
54,&a4?! &d2 55.&a5 is wrong, although
after 55...&c2, White can still backtrack and
win as in the main line with 56. &a4.
54...6.2 55.&bl £ d l 56.Sb2!
Dominating the black king.
56... e l 57.&C1 £fl 58.<±>dl £ g l 59.&el
& h l 60.&d2! &g2 61.&c3t &f3 62.&d4

abcdefgh
5O.fxe3!
The trick is that Black holds after 5O.b7?
Sh8! 51.Sc8 Sxc8 52.bxc8=® exf2, where
the e4-pawn is a visual distraction. It will go
and we will have a basic draw.
5O...Sxc2 51.b7
White wins in the long run.
abcdefgh
45.Sd7? does not work. It is easy to see that in
Not really mutual zugzwang, as White to variation b above, Black would have 51...Sd8.
play would be happy to play Sb2-b5 with the It is the only difference, but boy does it
idea Sf5t and Sf7, followed by eliminating matter...
the e4-pawn.
62...5d8t 63.&c5 S b 8 64.Sb3 45...Sb7?
The white king is too close for Black to Likewise Black wastes the chance to activate
successfully give up the rook. his passive rook (in front of the passed pawn)
64...6e2 65.&b6 with 45...Sh6!, which combined with ...Sh2,
White wins. ...&f3 and ...f4 will now be a tempo faster
and hold effortlessly. 46. b5 (Against 46.Sa8,
b) Active defence fails faster, if you look at the Black has 46...&f3 47.Sa2 Sh2 48. b5 f4!,
length of the variation. But like the TV chef, which works as always.) 46...S h i 47. c2 Sh6
the roast has been cooked in advance: 48.&c3 S h i 49. b6 S b l 5O.Sf6 &g5 51.Sd6
& g4
49...fxe3!?
46.Sd8?
The right idea, but the wrong file. Later in
this book we shall talk about the importance
of keeping the rook on the adjacent file to the
pawn you are trying to promote (see page 353
for example).

46.Sc8! worked again.

58 Conceptual Rook Endgames

46...&B 47.Sd2 f4? 51...&g2 52.Se2 <&B?

47...Sh7! holds as explained above. The type of mistakes you expect from a blitz
game. With only seconds on the clock, it can
48.exf4 &xf4 49.Sd5 Sf7 5O.b5 &f3 be difficult to spot finesses such as 52...Sf4!
with the idea of 53...&fl would have drawn.
As always 53. b6 drops the pawn to 53...Sf6.

53.Se3t &f4 54.&b4 SK 55.&a5 SB

56.&a6 Sc5 57.b6 Sc2 58.b7 Sa2f 59.&b6
Sb2f 60.&c7 Sc2f 61.&d7 Sb2 62.&c6
Sc2f 63.&d5 Sd2f 64.&c4 Sx£2 65.Sb3
Sc2f 66.&b5
In the final position the rook is also behind
the passed pawn, even if its main job is to
shelter the king from checks, it also prevents
the black rook from getting behind the pawn.
1-0
abcdefgh
51.Sd2?
5 1 .Sd4! would have been a powerful
prophylactic move. After 51...Se7, White
can play 52.Sd2, having forced the rook to
an inferior square (the e-pawn is not a passed
pawn!). And after 51...Sf4, White wins by
pushing the b-pawn.
Chapter 3

Shouldering
The principle of shouldering (at times also called bodycheck, choosing to compare it to ice hockey
over football) is simple. A king is helping the advance of a pawn, by at the same time controlling
squares ahead of the pawn to support it, and interfering with the path of the returning king,
trying to catch up with the pawn. When explained with words, this may sound complex and
convoluted, but when illustrated with examples, it becomes easy to understand. A version of the
following illustrative position once occurred in an engine game, but I prefer to show it in the
classical form, with White to play.

Basic Shouldering

abcdefgh
l.&g5 &c3 2.&fo!
2.f6? would allow the black king to get close to the white pawn. 2 . . A d 4 d.tfegb &e5 4.£7 &e6
and Black wins.

2...&d4 3.&e6 S e l f
60 Conceptual Rook Endgames

6O...Sxa2 61.f4 &c3 62.6 &c4

8 Black is in good time.
7
63.A6 Self 64.&B &d5 65.f7 66.&g6
6
&e6
5 0-1
4
59.f4? would also lose, but only to 59...Sa4l,
3 cutting off the white king along the 4th rank.
2 If White wants to use the king to help the
pawn, he will have to travel via f3-g4 to the far
1
side of the pawn, losing, as he would not be
abcdefgh able to prevent the black king from coming in
4.&d6! on the opposite side of the pawn.
Continuing to keep the black king at bay.
But White could draw without any drama
4...Sal 5.f6 Sa6f 6.&e7 &e5 7.f7 Sa7t with classic shouldering tactics.
8.&e8 &e6 9.f8=£>f!
White draws, as long as he keeps the knight 59.&d4!
close to the king. The king stays close enough to the f-pawn to
be able to defend it if needed. And at the same
What is remarkable is that this simple idea is time, enough in the way of the black king to
often messed up by grandmasters. ruin Black’s ability to bring the king back.

Jan Michael Sprenger - Sergey Fedorchuk 59...&b2

Germany 2020

abcdefgh
abcdefgh 60.SR1!
59.&e4? &b2 6O.Sxa2f 6O.Sxa2f? would lose to 60...&xa2!, as the
6O.Shl al=W 61.Sxal Sxal! also wins. The white king is now cut off and would have to go
white pawn is not far enough up the board to to e4 to help the pawn advance.
make it.
 Chapter 3 — Shouldering 61

6O...al=® 61.Sxal Sxal 62.f4 Sfl 63.&e5 43.. .5d5?

&c3 64.f5 &c4 65.16 &c5 66.&e6 We have already established that when you
White makes the draw by a tempo. are a pawn down, it is dangerous to defend
passively. Of course, there are exceptions,
At its core, shouldering is an easy principle and but usually they are based on concrete
can be used almost without thinking in rook considerations, as we shall see in Chapter 22.
endings - which can be useful in time trouble.
You try to keep the opponent’s king away, if 43.. . 6 g 5 ! would have activated the king and
you are doing the shouldering, or you try to given a highly desired exchange of pawns. I
get your king close, if you have the rook. It is will not confuse the matter with a variation.
possible to find complicated examples (I recall
a very difficult one that Kramnik got wrong 44.H4?
over the board from Grandmaster Preparation Preventing ...&g5, but giving Black another
- Endgame Play), but this is not truly desirable chance to put the rook behind the passed
from an educational point of view. So, the pawn.
following examples in this chapter will be
focused on the difficulties good players have Passed pawns must be pushed! The endgame
faced in implementing it in practice. Not is about queening or otherwise exploiting the
to put them on the spot, but to provide the power of passed pawns more than anything.
repetition that will make the principle stick. So, when you get the chance to put a pawn on
the 7th rank, you take it. 44. b6! followed by
The next game is between two strong Danish b6-b7 would have won.
grandmasters, from when they were juniors
and both defending European Champions in 44.. .5d4 45.g5t
their age group ( u l 8 and u l 4 respectively). 45.Sc6f!? driving the king backwards was
After an exchange of knights, White has an a better try, but with accurate play, Black will
extra pawn and is threatening to take another keep the balance.
one.
45...6e6 46.§xc5 Sxh4 47.&£2 Sb4 48.&f3
Jesper Thybo - Jonas Buhl Bjerre

Svendborg 2018

abcdefgh

abcdefgh
62 Conceptual Rook Endgames

48.. .5b3t?!
Black has a chance to do something and
8
instead treads water. 7
6
48...6d6 49.Sc6t &e5 would have drawn
instantly. 5
4
49.&e4 Sb4t 50.&d3 &d6 51.§c6f &e5
52.b6 &f5 53.&c3 S b l 54.&c4 &xg5 3
55.Sc5t 2
1
abcdefgh
59.Sb4t?
Overthinking. White was winning with
direct play:

59.&c6! g5 60.&c7 Sxb7t 61.Sxb7!

The rook needs to get behind the pawn as
quickly as possible.
61...g4
It is too late for Black to change strategy.
After 61...&g3!? 62.&d6 &f4 63.&d5 g4
abcdefgh
64.&d4 &f3 65.&d3 g3 66Wt White
White has gained a few tempos from Black’s wins, as the king will come close to the pawn.
dithering, but the position is still within the 62.&d6 g3 63. &e5
drawing margin, if shouldering is applied. The most elegant way to win, but prosaically
putting the rook behind the pawn also
55...6h4? works.
The wrong side of the pawn. Showing a lack 63.. .g2
of knowledge.

55...6f4! would be the same endgame, just

with the king on the correct side of the pawn.

56.Sb5 S e l f 57.&d5 Sc8 58.b7 Sb8

abcdefgh
64.&f4!gl=® 65.Sh7#
 Chapter 3 - Shouldering 63

59...£h3! 61...g5 62.Sb5 &h3 63.&B g4f 64.&£2

White’s idea is that after 59...&g3?, he can g3t 65.&gl &g4 66.&g2 &f4
play 60.&e5 and take the pawn, without the Going to the queenside to secure the draw.
black king being able to join in.
67.§b4f &e5 68.&xg3
6O.Sb3t Vl- 1/!
60.&c6 is worse now. After 60...g5 we can
see that the black king was given a free tempo. Mostly in this book, we will start at the critical
moments, but it is good to remember that the
game does not start there, but that a lot of play
comes before them and they are not earmarked
when they arrive. So, with this example, we
start at move 41, just after the time control,
where White is a pawn down, but with a lot
of decent ways to defend. What is interesting
is to see how Black improves his position by
manoeuvring the rook to e6 and only then
activating the king.

Bjorn Alilander - Can Ertan

Sweden 2020
abcdefgh
60...&R4?
Keeping the king on the wrong side of the
pawn in order not to obstruct it. This is not
stupid, but what all would do if they do not
understand shouldering. These players are both
hugely talented and smarter than most people
you will ever meet. But without knowledge,
you will make this sort of mistake.

60...<±>g4!I was the correct move. The king

wants to be on the left side of the pawn, where
it can keep the white king away. 61. c6 g5
62.&c7 Sxb7t 63.Sxb7 &f3! and Black makes abcdefgh
the draw, as the white king cannot join the
41.&e4 &e6 42.Sc8 &d6 43.Sf8 Sfl
action.
44.&d4 Sdlt 45.&e4 &e7 46.Sc8 &d7
47.Sf8 &e6 48.Se8f &fi6 49.Sc8 Sd6
61.£e4?
50.&f4 &g6 51.e4 3ft>t 52.&e5 3e6f
This leads directly to a draw with no chances
53.&f4 SfiGt 54.&e5 &g5 55.Sc7 Se6f
of winning at all.
56.&d4 ft.
61.<ic6! would transpose to the note above.
64 Conceptual Rook Endgames

White has to keep the balance between

defending the e-pawn and attacking the
c-pawn.

59.&xc4? Sxe4f 60.&d3 loses to 60...Sei!,

which will be explained in depth later (one of
few theory positions included in this book).
(60...f5? would allow White to defend from the
front [see Chapter 1 1, page 171] with 61.S h i !
&g4 62.Sglf &h3 63.Sfl Se5 64.&d4! and
holds.) 61.&d2 Se5! The white king is now
far from being able to harass the rook. 62.S h i
abcdefgh f5 6 3 . S g l t &h4 64.Sfl £g4 65.Sglf &h3
Black has made progress and we have now 66.Sfl &g2 Black wins.
entered what Boris Gelfand refers to as the
zone of one mistake. His argument is that 59...c3
White is also holding after 59...Sc6 60.&c3
getting your opponent into a situation where
he can make fatal mistakes is a big achievement &f4 61.Se7!.
on the path to winning. It is hard to disagree.
60.SR1!?
6O.Sc7? &f4 would lose, but White could
57.SR7!
The only way to defend against ...&f4. have played: 6O. xc3! Sxe4 61.S h i ! f5
62.&d3!, holding as above.

57...c5t 58.£<B
6O...c2 6 1 . 3 c l &f4 62.Sflf &g3
58.?bxc5 Sxe4 would see the king cut off.

58...c4t

abcdefgh
The critical position, where White was able
abcdefgh to hold only by shouldering away the black
59.£d4! king.
 Chapter 3 — Shouldering 65

63.Scl? Victor Torres - Andres Carlos Perez Javaloyes

63.&e3!! was the holding move. 63...Sc6
Santander 2011
(63. ..f5 64.Sf3t ! is a remarkable way to survive
with checks.) 64.Scl < >g2 65.&e2! White
survives by keeping the black king at bay. The
position is quite incredible.

63...&B
The black king comes in for the kill.

64. ?d3 Sd6f 65.&c3 Sc6f 66.&d3 Hd6f

67.&c3 &xe4

abcdefgh

With the white pawn reaching the 7th rank,

the likelihood of a rook vs pawns ending
becomes highly likely. With two flank pawns,
Black will continuously have to consider
carefully where to run with the king.

51.d 7 § e l t
Black gives checks until the white king hides
abcdefgh in front of the pawn.
Black is winning. The white king is cut off.
52.£d6 Sdlt 53. e6 S e l f 54.&d6 S d l f
55.&e7 S e l f 56.&d8
68.8elt &B 69.&xc2 6 70.8flt &g4
71.3glt &h3 72.Sfl Sd5 73.Shlf &g2
74.Sh5 &g3 75.Sg5t 76.Sgl &e3
77.Sg5 &e4 78.&c3 Sd8 79.Sg2 f4 8O.Sh2
&e3 81.Sh7 Sc8f 82.&b2 B 83.Se7t
&d2 84.Sd7t &el 85.Se7t &fl 86.Se6 £2
87.Sf6 Se8 88.&c2 &e2
0-1

The following game was not played between

top players, nor was everything perfect, yet it
had a lot of instructive moments. The game
was in the last round, making it plausible that
the players were reaching a stage of minor abcdefgh
exhaustion. And it was already quite a long
game by the time we come in. Which pawn should Black push?
66 Conceptual Rook Endgames

56...R5! Instead, 59...Sdl! would have held the game.

It is surprising that the h-pawn is the first to 6O.Sxh5 &c6 6 1.Sh6f & b7! White would not
make a go for it. The reason is non-obvious. It be able to make progress without allowing the
is a decoy! black king and pawn to advance freely.

Thus 56...a5? would lose in beautiful fashion.

57.Sc5! a4 (Also after 57.. .Sal 58.§e5 Black
would be lost. And after 57...&b6 then 58.Sd5!
with the threat of &d8-c8 would win on the
spot.) 58.Sc4! Separating the king and the
a-pawn. 58. ..a3 59.2b4f <&c6 60. &c8 White
queens the pawn and the a-pawn is already
cut ofF from the herd. After 6O...Sdl 61.d8=®
Exd8f 62.<&xd8 <±>d5 63.&e7 &e5 64.Ea4 &f5
65.Sxa3 h5 66.§a5t White will win quickly.

57.Sc7t
The point in shipping the h-pawn down the abcdefgh
river can be seen in the following variation.
57.Sb3t &c6 58.Sd3 &b7! Preventing the 6O.Sh7?
white king from escaping. 59.Sd5 h4 6O.Sd4 6O.Sh8! would have won in a very instructive
h3 6 l . S b 4 f &c6 62.&c8 S d l 63.d8=B line. 60...a5 61.Se8Sd5 62.&e7 &c6 63.d8=®
Sxd8f 64.&xd8 Exd8 64.Exd8 a4 65.Sa8 & b 5 66.&A6 £b4
67:& d 5 a3 68.&d4 &b3 69.&d3 a2 7O.Eb8t
&a3 71.&c2 a l = t 72.&c3 &a2 73.Sh8
& b l 74.Sxh5 and White soon wins the knight
as well.

60.. .5g5?
60.. .a5! 61.Se7 Sd5 62.&e8 &c6 and the
a-pawn saves the day for Black once again.

61.Se7?
6 1 .&e7! Sd5 62.d8=B Exd8 63.&xd8t and
abcdefgh 64.Sxh5 would cut off the black king on the
5th rank, which almost always wins.
64...a5 Black makes the draw, as the king
is already close to the pawn and can help it 61...5g8f?
advance: 65.Sh4 &b5 66.Sxh3 a4 And so on. 61...a5! drew as above.

57...&b8 58.Sc6! 62.Se8 Sg7 63.SH8?

Sneakily threatening Sb6f and &c7. 63 .Sf8! was a much better square for the rook,
as can be seen in this variation: 63... c6 64.Sf6f
58...&b7! 59.Sh6 Se5? &b5 65.&e8, and White wins two tempos
An understandable mistake with a tricky compared to the next note. Black has to take on
refutation. d7, not only leaving the white king closer, but
 Chapter 3 - Shouldering 67

also without the benefit of him making a move 65...Sg8f! It is important to keep the king
first. All because 65...a5 66.d8=® Sg8f can be as far away as possible. Later, we shall look
answered with 67.Sf8!, and wins. into this topic in greater detail. 66.d8=®
Sxd8f 67. &xd8 a5 68. &d7 a4 69.Sxh5t &c4
70.&c6 a3 7LSh4t &c3 72.&b5 a2 Black is
in time to make the draw.

64.&e7
Black resigned.
1-0

N. Griva & A. Zhukov

sp.hm Shakhmatnaya Kompositzia 2018

abcdefgh
63...Sg5?
Black is losing the plot - and possibly also
on time.

Black had a chance to make a draw with:

63... &c6! It is important to threaten to steal
the d-pawn. White has to give up a tempo.
64.Sh6f & b 5 The king needs to make it to
the 5 th rank, or he will be cut off. Imagine that
the pawn reaches a3. White will play Sh3 and abcdefgh
Black has to be able to play ...&b4! to support
There are situations where the king and the
the pawn, or it will be lost. 65.&c8 (65. eS
rook can seem to be in each other’s way. And
a5! 66.d8=® Sg8f transposes.)
as tempos should not be carelessly discarded,
knowing how to get the pieces to collaborate
is useful. In the following study, the first
blitz tries of the hand fail, even if the way
to coordinate the pieces is not so surprising.
This example could easily be in the chapter on
Rook vs Pawn, or in the chapter about cutting
off the king, but we are keeping it here to put
a torchlight to the shouldering aspects of it.

3.Sa5?I
Premature. White allows the black king to
join the game, at the cost of only a pawn.
abcdefgh
68 Conceptual Rook Endgames

3...&c6 4.§xa4 b5 5.Sal <±>d5! better from the c-file than the d-file, which
Shouldering. Black draws. is only the case if it is attacked. Which means
protecting the pawn from d4 or d3 is better.
A far more natural try is: Now 7-<±e5! b4 8.Ea8 <£c4 9.&e4 &c3
10.&e3 b3 1 l.Sc8f and White wins by a
3.<±>fi5?! tempo.

7.§a8 b4
The draw is achieved.
The solution is to bring the king back
without interfering with the rook’s domination
of the black king.

3. g6!

abcdefgh
But the king and the rook do not cooperate
well like this.

3...B6I!
The key move. Taking away the a5 -square
from the rook is the best use of Black’s turn.

4.Se5
abcdefgh
The necessity of manoeuvres is usually a bad 3...b6!?
sign. The composer’s main line. The point of the
white position is that it is flexible and the king
4.<±>f5 is met with 4...<±>d5!> with classic can still choose to go the fast path, if Black was
shouldering. to throw away a tempo:

4.. .a3! 5.Se3 b5 6.Sxa3 <±><15! 3...&c6 4.<±>F5! &d5 The shouldering attempt.
The standard move. Keep shouldering! Black has no chances if he lets the white king
come close. 5-<±>f4T The black king is trapped
6...b4? 7.Sa5! would see the king cut off on between shouldering the white king and
the 5th rank, which is a death sentence. This defending the pawns. After 5...&d4 6.Sa5 b5
stuff belongs in Theoretical Rook Endgames. 7.Sxb5 &c3 8.Sa5 &b3 9.&e3 White is in
time.
6.. .6.5? is a more likely mistake, forgetting
about shouldering, reacting o n a human 4.&h5!
emotion that the king defends the pawn The not-so-surprising point.
 Chapter 3 - Shouldering 69

4. f5? would fail to 4... <&d5 5. f4t d4 and 39...Sd2

we can see the benefit of 3...b6, taking away With this move Black accepts that he will
the a5-square from the white rook. lose the rook sooner or later, but hopes that the
king will be able to create counterplay on the
4.§f5? is also slow. There is no space for kingside before it happens.
manoeuvring in a race. 4...a3 5.Sf3 b5 6.§xa3
&d5 and Black makes the draw. The passive try was also difficult to refute:
39.. T5 4O.Sc3!
4...&c6 Renewing the check on the e-file.
4...a3 5.Sg3 is similar. 40...a5 41.&c7! Sd2 42.Se3t &f7 43.d6 a4
44.d71?
5.&g4!
I think everyone would play this move, but
The king is in time. Move mores can (and
there were more wins, most spectacularly:
should) be played, but the outcome is certain
44. d7! is sneaky and more beautiful, but
bar any silly slips.
besides noticing the idea of transferring the
1-0
rook along the 8th rank after a check, we
shall not pay too much attention to it.
The following example starts with White
44...5.2f 45.&d8 Sc6
having a positional advantage through his
more active king and strong central passed
pawn. But quickly it becomes clear that the
pawn will queen and claim the black rook in
the process. The race that will decide the game
is all relating to the counterplay Black will be
able to create in the meantime.

Swapnil Dhopade - Tejas Bakre

New Delhi 2017

abcdef gh
46.g4!l
A classic break.
46...fxg4 47.Se7t
47.f5? is impatient. 47...gxf5 48.§e5 &f6
49.Sd5 Sa6! 50.<&c7 Sa8 would allow Black
to escape with a draw. The king will help the
f-pawn advance.
47...6f8 48.f5 gxf5
Otherwise fxg6, §e8t and c7 will allow
the white rook to get to a8 and the d-pawn
to promote.
abcdefgh 49.Se5
White wins.
39.&c6!
Fighting for the squares in front of the pawn.
70 Conceptual Rook Endgames

39. ..£d8?! 40.&d6! Sd2 41.Sc6! will allow 51.SB6?!

the rook to go to a6 or c7, in either case with Making things harder than they needed be.
an easy win.
51.d8=® could be calculated. 51...Sxd8
4O.Se4f &d8 41.Sa4 &e7 42.Se4f &d8 52.<&xd8 g5 53. e7 gxf4 54.gxf4 <&xf4
43.SB4 &e7 44.Sb7t 5 5 .&f6 with an easy win.
44...<&e8 45. d6 <&f8 46.§a7 wins easily.
51...Sxd7t 52,<&xd7 &xg3
45.d6 Sxh2 46.&c7 Sc2f 47.&d8 Sd2
48.d7 &f5 49.&e7 &g4

abcdefgh
53.Sxg6f?
abcdefgh Swapnil misses out on the final chance to
50.SB3! win the game. It is based on shouldering, so
A natural move, defending the backstopping although it is a series of only moves, they are
pawn. White also had a direct win with all not impossible to find in a practical game.
natural moves after: 50.d8=®? §xd8 51. xd8
&xg3 52.Sxf7 h5 53.&e7! h4 54.Sh7 h3 53.&e6! £xf4 54.&f6! a5 55.&xg6 <±>g4
55.&f6 a5 56.&xg6 a4 57.f5 h2 58.f6 a3 56.Sb11 f4 5 7 . S g l t &h3 58.Sfl <±>g3 59.&f5’
59.f7 a2 6O.f8=@ al=@ 61.@d6f and mate to f3 60.<&e4 f2 61. 63 White wins.
follow. But in time trouble, no one would trust
their own accuracy in such a variation. 53...&xf4 54.Bxh6 &e4 55.Be6f st?d3!
56.Bd6f &e4 57.Be6f &d3 58.Bfi6 &e4
50...B 59.Bxa6 f4 60.&e6 6 61.Ba4f
But now it is time to queen the pawn. Things Vl-1/!
will never get as favourable as now.
 Chapter 3 - Shouldering 71

Shouldering with a rook each

The following examples are truly complex. The

shouldering is happening in positions where
both players have a rook, and thus should
be understood in a somewhat different way,
although it is still about keeping the opponent’s
king at bay.

The following game is a roller coaster ride,

with a beautiful combination, botched in time
trouble, followed by a deeply complicated
rook endgame, where shouldering will occur abcdefgh
in a profoundly different version than we are 36...Sh2?
used to.
36...bxc5! was the winning move. Black
takes a pawn, realising that White cannot do
Peio Duboue - Matthias Bluebaum
anything with the tempo he has been given.
Chalons-en-Champagne 2021 The transition into the rook ending afterwards
would be deadly.

37.&gl Sag2f 38.&fl Sc2

Finding there no longer is a forced win,
Bluebaum tries his luck in the rook ending.

39.®xc2 Bxc2 4O.cxb6 Bb2

4O...jLxc4f?? 41.bxc4 with the threat of
c4-c5 would win for White. 41...§xc4 42.§bl!
for example.

41.<±>gl?!
41.§d3 Jkxc4 42.bxc4 §xb6 43.Sd6! f5
abcdefgh 44.&f2 g5 45.&f3 draws.
31...®xe3!
41...jLxc4 42.bxc4 Hxb6 43.Bd6 Bb2!
A wonderful queen sacrifice.

32.fxe3 Sxg2f 33.&fl la6f 34.Sc4 Sd£2f

35.&el Bxa2 36.&fl
72 Conceptual Rook Endgames

49...f4!
A strong practical decision. Black needs
passed pawns to win. Going from a pawn up
to pressuring a weak pawn could be a step in
that direction.

49.. . S h i 50.c5 S c l 51.Sc8 followed by c5-c6,

with the intention of snatching a pawn on the
kingside in exchange for the c-pawn, looks like
a draw too.

5O.exf4 <±>F5 51.S£8

abcdefgh I dislike this decision intensely. I feel that
44.&fl! there are more chances of making mistakes in
The h-pawn is traditionally the least useful a situation where White allows Black to have
pawn and White is right to let it go. two passed pawns and relies on the c-pawn
44.Sxc6 Se2 and wins. being a counterweight. Put simply, passed
pawns are more likely to promote.
44...c5 45.Sd5 Sh2 46.Sxc5 Sxh3
Black is a pawn up and definitely has winning I personally would see getting the king to f3
chances in practice. Analysing an endgame like as the path to a draw. After 5 1 .Se3! S h i 52.c5
this with an engine can be highly misleading, S c l 53.&f3 f6 54.c6 Sc4I? 55.&g3 Sxc6
as it will tell you when the position changes 56.Sa3 I cannot find a way for Black to pose
evaluation, from drawn to lost, for example, White with serious problems.
but is far less helpful in telling you when your
position becomes more difficult. I have tried 51...K 52.&d2
to speculate below, but this is always up for This is the idea of the previous move.
discussion. 52.c5! Sc3 53. &f2 Sxc5 54.&f3 would be a
way to keep the balance.
47.&e2 f5 48.Sc8 &fi6 49.Se8
I intuitively like this move. 52...SB 53.c5 Sxf4

abcdefgh abcdefgh
 Chapter 3 - Shouldering 73

54.c6? 61.Bf8
This is a real mistake. White should not Because of the check on f l , this is now the
allow the black rook to get behind the c-pawn only try.
so easily. Especially with the king being passive. 61...g5 62.<&d6 &g3 63.Sf5
After 63. c7 f3 64.c8=@ Sxc8 65.Sxc8
54. <&d3 Sfl 55. c4 would still give White £2 66. S c l , the easiest is to shoulder with
enough counterplay. The black rook does 66...&f4!.
come behind the c-pawn, but at least the white
king is active. (55. d2? <&e6 and Black wins)

54.. .5c4 55.Sc8 g5?

The most natural-looking move in the world
and not for a second one you would suspect
was a mistake. Once you see the defensive idea,
you probably get confused more than inspired.

55.. .6f4!? looks natural, once you understand

the defensive idea, as it takes away the e3-d4
abcdef gh
path for the white king. But it also involves
giving up the rook and racing the two passed 63...g4 64.Sc5 B d l f 65.&e7 S a l 66.c7 Sa8
pawns down the board, winning by one tempo. 67.c8=® Sxc8 68.Sxc8 f3 69.&e6 £2 7O.Scl
Looking at it during the game must have been <±>h2
challenging. Black wins.

55...&g4! 56.&d3
This has the same issue. It does feel slightly Played after more than 4 minutes, after
more natural to me, as it prevents all which White resigned. Prematurely, as the
Se4-c4 ideas, due to ...fl=@ being with position still holds.
check. But honestly, it is a matter of
experience and taste and not objective at all. 56.. .5cl 57.&d2!
56.<&d3 S c l 57.&d2 Sc5 58.&e3 f5 59.&d4 57. d4? looks natural, and was probably
S c l 60.&d5 f4 what White was looking at that made him
resign prematurely. But after 57...&f4 58. <&d5
g4 59.<&d6, Black would have to find 59...f5!
6O.Sg8 g3! to win the game. It may look easy
when you see it, but there are other options
that look reasonable, just like 55...g5? did.
For example, 59...g3?, where White holds after
6O.Sg8!, with the core idea 60...&f3 61.Sf8!.

57.. .5c5

abcdefgh
74 Conceptual Rook Endgames

The king is headed for the faraway corner,

8 as first discovered by Botvinnik. The endgame
7 is still difficult, but objectively within drawing
margins.
6
5 The following example is simpler.
4
Radek Kalod - Vitezslav Priehoda
3
Czech Republic 2021
2
1
abcdefgh
58.&e3!!
Shouldering. This is a race between the
white rook and black pawns, just with a lot of
additional elements in the mix.
58.<±>d3 <±>£4 59.&d4 Sc2 60.<±>d5 g4 is
hopeless.

58...&g4
The black king is kept at bay after 58. ..g4
59.c7 Bel 60.&d2! Sc6 61.&e3!. abcdefgh
54...Bxc2??
59.&d4 Bel 60.&d5 f> 61.Be8 f4 62.Be4!
A simple move, played automatically.
&g3
62.,.&f5 63.Sc4 §xc4 64. <&xc4 e6
55.&f4 &d6 56.&e3
65. b5 only makes things easier for White. White drew.

63.Bc4 Bxc4 64.&xc4 £3 65.c7 £2 66.c8=®

Black would have won by shouldering.
fl=®t 67.£b4!
54... e6!
would have the double effect of shouldering
the white king (54... db? 55. f5! holds
for White) and putting White in zugzwang,
unable to hold the blockade along the 5th
rank.

55.Ba2 &e5

abcdefgh
 Chapter 3 - Shouldering 75

The game was a disaster for White.

52.f4?
52.§d8? also does not work. Black has many
ways to win, for example 52...§a4 or 52...§c2f
53. e3 £ c l and Black wins the race. We
maybe cannot calculate this clearly, but we can
see that it is slower than the alternatives, which
is enough to discard it.

52...Sc5
Going for the Lucena position. 52...§c2f
abcdefgh 53. <& d l §c3 was another good way to
Black is winning in many ways. The best continue.
technical approach is to keep White tied down
and improve the position as much as possible, 53.Sa7 Sd5t 54.&e3 &c2 55.Sc7t &b3
which here means advancing the c-pawn. 56.Sb7t &c3 57.Sxb2 &xb2 58.&e4 Sd8
59.B &c3 6O.fi6 Sffi 6 1 . & B &d4
56.Sb2 c5 57.Sa2 Sc3 58. >g4 c4 59.Sb2 0-1
Sa3
Black wins the c-pawn at the right moment. White makes a draw by creating a passed pawn
and using the king for shouldering.
The following example is mind-blowingly
complex, if you simply look at the variations 52.B!
and allow yourself to get confused. But with
a knowledge of shouldering, it was possible to
find the draw for White.

Sarasadat Khademalsharieh -
Mykhaylo Oleksiyenko

Doha 2014

abcdefgh
52...§c2t
I assume that White believed that she would
be lost after this move.

52...§d4f is less of a challenge. After 53. e3

§b4 54.<&d2! §b3, Black has made a small

abcdefgh
76 Conceptual Rook Endgames

improvement, but White will still be able to get 57...&al!

herself a passed pawn with: 55.g4!, when the 57...<&a2? would allow the white king back
draw is available, as long as White remembers into the game: 58. c2! §c4f 59. d3 Sc8
to run for the kingside with the king at the 60.§b7! (60.§b6? would help the rook to get
critical moment. to a better rank. 6O...Sc6! 61.§b7 b l = @ t
and Black wins narrowly.) White holds after
both 6O...Sc7 61.Sb8, and 6O...bl=®t
61.§xbl <&xbl 62.<&e4!.
58.&c2

abcdefgh
53.&d3!! abcdef gh
A mixture between prophylaxis and
shouldering. The power of this move cannot 58.. .5a5!!
A fantastic waiting move. The rook is also
be fully understood without looking seriously
OK placed on a6, but not a7.
at the alternatives.
58...5a2 59.&d3! Sa4 60.&c2 would give
53. d l ? §c3 is hopeless. Black another chance, but not tell him what
to play.
But it is more puzzling that White is lost after: 58...§a7? fails after: 59.Sb4!! §c7t 60.<&d3
53.&e3? & c l 54.Sb8 b l =®t 6 1 . S x b l t & x b l 62.g4 &b2 63. <£>e4
This would allow Black to cut off the white <&c3 64.&f5 d4 65.gxh5 d5 66. h6 and
king with: White draws.
54...Sc4!! 55.&d3 Sa4 58...5a6’? does work too, but is less clear cut.
Black wins after: 59.Sb4
56.Sc8t £ b l 57.Sb8 59. g4 Sa2 60.<&d3 hxg4 61.fxg4 §a4! would
cut the king off at the 4th rank and decide
the game.
59.. .5c5t 60.&d3 b l = ® t 6 1 . S x b l t £xbl
62.g4
After 62.<&e4 <&c2 63. g4 <&c3 White is
making no meaningful progress.
62.. .<&b2 63.gxh5
63.&e4 c3 Black wins.
63.. .5xh5 64.&e4 c3 65.f4 c4 66.f5 c5
67. e5 c6 68. 66

abcdefgh
 Chapter 3 — Shouldering 77

abcdef gh abcdefgh
68.. .6c7! White should not jump the gun with
68.. .5xh4? would be a mistake, as after 69. f6 55. e4?, when Black wins with accurate play:
Se4t 70.&f5 S e i , White has 71.f7 &d7 55...Sxg3 56.Sc8t £ d l 57.Sd8t £ e l 58.Sb8
72.f8= f ! surviving. Sg2 59.f4 £ d l 60.&f3 Sd2! 61.f5 <±>cl
69.f6 62.&f4 Sd8! An important tempo-winning
White cannot keep up with the black king move you can find a number of times in this
and defend the f-pawn at the same time. book. 63.Sb6 b l = ® 6 4 . S x b l t &xbl 65.&g5
69.. .6d8 Sh8 66. f6 c2 67. g6 d3 68. g7 Sa8
The black king arrives in front of the pawn White is lost. The black king is returning too
and wins the game. fast.

53...6C1 54.SB8 b l = ® Rather, he should keep harassing the black

54...5g21? could be a small trap. king with 55.Sc8f & b l 56.Sg8!, when Black
is not making progress.

55.Sxblf xbl 56. e4!

White draws.

56. g4? hxg4 57.fxg4 Sg2 58. g5 Sg4! on the

other hand, would see the king cut off on the
4th rank, which is a death sentence.
Chapter 4

Rook vs Pawns

We have already considered rook vs pawn to a great extent in the previous chapter on shouldering.
But in this chapter we shall take a greater look at this material constellation, as it is a frequent
occurrence. Rook endings often see a pawn being promoted, winning the rook, but with the
game going on. At their core, rook vs pawn endgames are races (exceptions at the end) with lots
of finesses individual to each position, but also a lot of recurring themes and ideas. The two big
themes are shouldering and checks (more about checks later in Chapter 9, page 143), but also
about keeping the king as close as possible. While it is possible to talk at length about this in the
abstract, it is more meaningful to look at practical examples and show the key points along the
way.

The first position came from Irish G M Alexander Baburin and illustrates the finesse and check
part.

Baburin

Chess Today 20 1 9

abcdefgh
80 Conceptual Rook Endgames

l.£cl! Elisabeth Paehtz - Jan Gustafsson

Not the first move you would think of
Germany 2015
(which would be l.&c2), but actually it is
the principled move, with the king aiming to
come in front of the black pawn, which sounds
irrelevant, until you combine it with the check
on the 2nd rank. White could also reverse the
order of the first two moves, but it would come
across as a bit odd, even if it works.

I.&c2? f3 2.Sa8 f2 3.Se8t f3! Black would

make a draw.

1...B 2.Sa2f!
Winning an important tempo. Black has to
keep the shouldering with:
abcdefgh
2.. .£el 103.£el?
But now the king is misplaced. But going But in the centre the king is exposed. Most
to the 3rd rank would allow the white king to players’ intuition is created more from opening
approach. and middlegame considerations, than for
positions with a minimum of pieces on the
3.Sa8 £2 4.Se8f board. In the middlegame or early endgame,
White wins. we would try to centralise the king, but in this
endgame, the king is better off on h2, where
There are generally speaking two defensive it is less of a target and less in the way of the
strategies in chess and the same is the case rook’s ability to operate from the side (see the
in rook endings: passive and active. Without chapter on Checking Distance, page 163).
making it into a general case, passive defence
is rarely successful in queen and rook endings. The correct way to defend the position is thus
The heavy pieces are too strong and can inflict 103. g l ! £f4 104.Sc8 S a l t 105.&h2 Sfl
too much damage. But there are exceptions. 106.Sc2, where Black does not have a way to
For example, the next position, which belongs break through. With the king on d2, the rook
in a book on endgame theory, and not on cannot defend the pawn from the side.
concepts (that comes a bit later). Part of the
reason for White needing to play passively is 103...£f4 104.Sc8 Salt 105.£d2 e3t
the lack of targets for active counterplay. 106.fxe3t £g3 107.Sg8f £h2!
The practical move.
107.. .££2 won faster according to the
tablebases, but I prefer Gustafsson’s technique,
where a few moves are spent getting the king
to the preferable fl -square, making the win
technically easier, which is a different thing
than “moves to mate”.
 Chapter 4 - Rook vs Pawns 81

108.Sh8t &g2 109.Sg8f &fl 110.e4 £2!? 112.&dl?

1 10...§el was more natural. 112. d3? would also lose. After 112...&el
1 13.Sf8 §a4! the king is cut off at the 4th
I l l . e 5 Sa2f? rank and the pawn is unable to provide any
An understandable mistake. The win was far counterplay. We will talk more about cutting
from easy: off the king in Chapter 12, page 175.

1 1 l...§el 1 12.§g5 §e2f! These small checks 112. 63! followed by giving up the rook is
are so helpful; we will have a full chapter on however a draw. It is difficult to see what Paehtz
them (see page 143). 1 13. d l §a2 The rook was afraid of, but at the same time, anyone
is manoeuvred to the best possible square, with practical experience of defending endings
which happens to be in the d-file. 114.e6 where there are so many plausible moves and
§a6 1 15.Sg6 (H5.Se5 looks clever, but after so little time to consider them, will know the
115...Saif 1 16.&d2 S e i , the outcome is the difficulties she faced. The point of this book
same.) 1 15...Sd6t 116.<£c2<£el 117.e7Sd2t is to give the reader a lot of experience with
1 18.&C3 Se2 And Black wins. typical situations and thus the tools to make
this type of decision correctly and confidently
Our key moment of interest. in their games.

112...Sa5
Black wins.

113.e6 Sd5t H4.£c2 Se5 115.Sg6 & e l

0-1

The goal for the side with the rook is almost

always the same: to get both the rook and the
king close to the scene of action, in order to
stop the promotion of a pawn. It does not
greatly matter if the weaker side has one or
more pawns. If the king and rook are both
abcdefgh there, they can deal with a lot, even if it is not
It is useful to develop a good feeling for when always trivial to do so.
the race for counterplay with a pawn against
a rook is likely to be successful, so you can
make these type of decisions confidently, even
if tired after a long defence and under time
pressure. Paehtz certainly considered eloping,
as she spent 1 minute and 38 seconds on this
decision and was down to seconds afterwards.
But the hope that passive defence can hold
here turned out to be misplaced.
82 Conceptual Rook Endgames

Sebastian Donoso Diaz -

8
Gonzalo Ivan Recabarren Gajardo
7
Antofagasta 2016
6
5
4
3
2
1
abcdefgh
50.&B h3 51.Sc3t
Taking on h3 first is a transposition.

abcdefgh 51...&xg2
51...&h2 is refuted most simply by taking
Black made an effortless draw after: on h3 and playing &f5-g4, but 52.g3 is cute.

49.Sc2? h4 52.<&xg4 h2 53.Sc2f <&gl 54.<&g3 h l = f

Now the white king will not get back in
55.&B
time.
And the knight is trapped and lost.

50.&e6 h3 51.gxh3 gxh3 52.&B h2 53.Scl The next study is a lot about shouldering, but
£g2 it starts off with a deep finesse, which matters
*/2-¥2 in the coordination of the pieces in the race for
the 8th rank.
White would have won easily if he had brought
in the king: R. Reti

49.&e6 h4 Shakhmatny Listok #369 1929

Losing the g-pawn does not matter.
Remember that Black will not be able to
advance both of his passed pawns at the same
time, so those two connected passed pawns are
not that valuable.

49...<&xg2 50.&f5 g3 51. 4 h4 52.&g4 and

the check from c2 will decide against 52...h3,
while 52...&h2 loses to 53.§h7!.

abcdefgh
 Chapter 4 - Rook vs Pawns 83

The most natural attempt is to advance the 4.c4 Sd8t 5.&c6

e-pawn. 5. e6 §c8 6. d 5 b2 Black wins by a
tempo.
1.e4? Se2 2.<&c4!
White needs to activate the king and wins a 5...Sc8f!
few tempos by giving up the e-pawn. As said, this is key.

2.c4 §xe4 3.&b4 & b 2 White is lost because 6.&b5 &b2 7.c5 &c3 8.c6 <&d4
of the pin on the 4th row. Black will continue Black wins.
...< )c3-d4-e5.
An even weaker attempt is:
2...Bxe4f 3.&d5
l.c4? Bxe2 2.&b4
2.c5 Se4 cuts off the king at the 4th rank
8
and wins instantly.
7
6 2...&b2 3.c5 Sc2!

5
4
3
2
1
abcdefgh
3...Se8I!
The instinct is to place the rook behind the
passed pawn, but no advice is so good that it
should be followed in all circumstances. But it
is easy to see someone missing this option; try
abcdefgh
to choose the best move at move 1 . To me, this The rook usually belongs here.
is the move that makes the study.
4.<&b5 &c3
3...§el? 4.c4 S d l f looks tempting, but if you Basic anti-shouldering tactics. Black wins.
have truly understood the previous chapter on
shouldering, you will know where the white The surprising solution is to save the pawn,
king belongs. 5. c6!! &b2 6.&b6 and the but not push it to the max.
c-pawn is running down the main strait to the
finish line. We now see the point of 3...Se8!!, I.e3! & b l
which is to win a tempo on the white king, The college try.
when giving the check from c8, forcing the
white king back to the 5th rank, winning a
vital tempo.
84 Conceptual Rook Endgames

Reti’s original main line went like this: l...Se2 The next study comes from the simple
2.c4 §xe3t 3. b4 &b2 4.c5 and the pawn question: what would happen if we moved
draws. The key comparison with l.e4 being the pawns over a row? It turns out that the
the lack of a pin on the 4th rank. position is winning and for this reason, to
follow study-convention, I have made the first
2.£c4 £c2 3.e4 Se2 4.&d5 £d3!? move and reversed the colours, as well as not
Trying to get the king around the e-pawn, looked up if Reti did this study too, to avoid
only to face a new problem. disappointment.

After 4...Sd2f 5. c4!? Black is clearly not Aagaard

making progress. But going to c6 draws too.
Original 2023

After 4...&xc3 5.e5 White draws by a tempo.

abcdefgh
abcdefgh 1.£c8! <&d5
l...d5 2.§a6f &c5 loses to simple
5.c4!
shouldering. The white king needs to be on the
The most accurate way to play. White is
other side of the pawn. The f6-pawn does not
shifting his front runner, to prevent Black
matter. 3. d7 d4 4.&e6 d3 5.Sd6 &c4 6.&f5
from being on the winning side of the pawn.
and White wins.

5.e5 is a bit less accurate, as you will have

2.<&d7
to defend knight against rook, which is not
2.Sf7? fails to 2...&e4, when Black is in
too difficult, but can still go wrong on rare
time. (2..T5? is tempting, but after 3.Sxf5t
occasions. 5...&e3 6.c4! <£>f4 7.c5 &f5 8.<±>d6
&e4 4.Sfl d5 5.S e l f ! White wins by a tempo,
Hd2f 9. 67! Shouldering to benefit the
as in the Reti study above.) A key point is that
c-pawn. White makes the draw, promoting to
after 3.§e7t Black only draws with 3. ..* 5!!,
a knight.
when the white king is shouldered.
5...Bxe4 6.c5 Bd4f 7.&e6
2...B 3.&e7
White draws.
3.§a5t? &e4 4.&xd6 f4 would allow Black
to run away.
 Chapter 4 — Rook vs Pawns 85

3...&e5
8
7
6
5
4
3
2
1
abcdefgh
10.&b4!!
abcdefgh To me this is the beautiful move. White is
4.&f7!! aiming to come back to the 1st rank to block
This is the key move. The white king is the black pawns. The curve on the route is
preparing to leap in either direction, while irrelevant to its effectiveness, but even after
space is made for the rook to give the disruptive close to 40 years with chess, I still struggle with
check. this optically.

4.. .d5 5.Se7t &f4 10...d4

The best try to keep the shouldering going. 10...f3 11. c3 would also allow the white
king to reach the 1st rank.
6.&e6
The king goes opposite, to avoid the Il.&b3 &e3 12.&C2 £3 13.&dl
White wins. Black will be out of moves soon
shouldering.
enough.
6.. .6e4 Delaying giving up the rook
Black is just lost. For example: 6...d4 7. d 5
d3 8. < )c4! d2 9.§d7 &e3 10.&c3 Winning. Focusing on the rook vs pawn endgame as
a race helps to clarify some things. In a race
7.&d6t there is a finish line. We have that: it’s the
7.&f6t!? works in the same way, just lst/8th rank. Forgetting for a moment that
mirrored. There are exactly enough ranks to the race is an obstacle race, let’s notice that the
make it work. two sides often start at different distances from
the finish line, or race in groups that move at
7.. .6d4 8.S£7 &e4 different speeds. While this can lead to a lot of
8...&e3 9.Hxf5 d4 loses to: 10.Se5t!, complications, there are also some things that
winning an important tempo, as shown above. can be simplified. The key point is that it is in
10.. .6f3 1 l.Sd5 &e3 12.&c5 the interest of each side to be as close as possible,
and to keep the opponent as far as possible
9.&c5 f4 from the finish line. While this may sound
abstract, it has a simple strategy that should be
86 Conceptual Rook Endgames

implemented as often as possible, which is to 46.§b8 was the most natural way to make the
delay the moment when the defending side is draw. 46...Sxf4f 47. <&xg6 §b4 48.§c8f &d4
losing the rook for a passed pawn as much as 49.Scl and White will have enough time to
possible. What we would seek to achieve is to save the draw with the g-pawn against a rook.
get the opponent’s king to be as far away from Notice that with the rook on b l , the black
the finish line as possible, which is frequently king will have to come to the 2nd rank for the
at the other end of the board. pawn to draw blood. This is far away from the
8th rank.
The players in the next example may not be
world class, but they are still strong amateurs 46...Bxb3 47.&xg6 BfB 48.f5
and as we shall see below, the mistakes they are
making may be easy to explain and understand,
but stronger players make them as well with
great certainty.

Tamas Tancsa - Zala Urh

Terme Catez 2022

abcdefgh
48...&d5??
At the Academy, the first fifteen suggested
the right move, and only those overthinking
(when asked not to), ignored their instincts
and suggested something else.

abcdefgh 48...&d6! 49.&f6 (49.f6 loses to 49...Sg3!,

46.Bxb3?? which can be missed) 49... <&d7! would win the
As explained, White needs to delay the race to the 8th rank. The king comes in front
capture of the black pawn as long as possible. of the pawns and the win is easy.
10 out of 19 of the students at Killer Chess
Training nevertheless suggested this move. One 49.&fi6!
IM being “100% certain” that it was the right 49.f6 leads to the lesser draw, relying on
way to play. White draws with §b7, Sb8 and defending knight against rook. The move
§e4. All delaying the moment when the rook played in the game claims the rook by
is lost. White only needs one pawn to draw. advancing the pawn furthest away from the
As we discussed above, the extra pawn rarely black king.
helps, as only one can be successfully pushed
anyway. But the black king being close matters
greatly.
 Chapter 4 - Rook vs Pawns 87

49...&e4 5O.g6 Bxf5t 51.&e6 Ba5 52.g7

Sa8 53.&f7 54.g8=® Sxg8 55.&xg8
Vi-Vi

The next example is thematically not too

different. There are a few extra pieces, the race
has more possible outcomes and a theoretical
fortress is suddenly possible. But at its core,
it is the same. Black will lose the rook, but
should make sure he does so at the latest
possible moment.

Viktor Haarmark Nielsen - Casper Yukon Liu abcdefgh

Svendborg 2019 51.Sf7?

51.&e3! would prevent Black from creating
counterplay, while at the same time also
making it impossible for him to set up the
fortress missed on the next move. Also, as this
is still to some extent a race, it not only arrives
at the finish line first, but it also prevents Black
from approaching it with his king. 5 l.-.&gb
52.<±>e4! h4 53.Se6!

51...g4?
A mistake committed due to lack of
knowledge. I cannot suggest frequently enough
that you read Theoretical Rook Endgames by
abcdefgh Sam Shankland and memorise some basic
positions. You will still make mistakes, but
48.Se7!
fewer.
The exclamation mark is added due to the
practical strength. This move won the game.
51...h4 52.&e3 <±>g6!53.Sf8 £g7 54.Sa8
f5 followed by ...g4 and ...h3 leads to a
48...Sxc5?
theoretically drawn position, which we will see
48...§xe7? 49. <&xe7 g4 is also a bad race for
later in this chapter too. (But you should still
Black. White queens first and gives a check
read Sam’s book!)
on g8, picking up the newly-crowned black
queen.
52.&e3 &g5 53.Sg7t £h4 54.£f4
White wins.
49.&xc5 h5 50.&d4 &f5
Black is a tempo short on setting up a
54.. .6113 55.Sh7 h4 56.Sh6 g 3 57.&13 5
fortress. After 50...h4 51.Se8 g4 52.<&e4 f5t
58.SE7 f4 59.Sh8 g2 6O.Sg8 &h2 61.Sxg2f
53. &f4 h3 White wins with 54.§f8I.
&hl 62.Sa2 h3 63.&xf4
1-0
88 Conceptual Rook Endgames

Black would have drawn easily, as well as given 54.. .h5 55.Se8!
his opponent the chance to go overboard, if he We shall see this idea repeated endlessly
had made the most of his rook for as long as below.
possible. 55.c8=®? transposes to the previous note.

48...SB! 49.c6 Sfl 55...H4 56.c8=® Sxc8 57.Sxc8 g 4 58.&e6

Heading for the natural place behind the &g5 59.Sg8f £f4 6O.£xf6 h3 61.Sh8!
passed pawn.

50.c7 S d l f !
Tempting White to get the king to the 8th
rank. Going behind the c-pawn also draws,
but is practically a good deal worse.

51.£c6 Self 52.£d7 S d l f 53. e8 Scl

Only now, when the checks have run out.

abcdefgh
White holds the draw, not least because of:

61...g3 62.Sh4f
<
Although 62.Sxh3 g2 63.Sh4f if3
64.Sh3t also holds.

One more!
abcdefgh
54.&d7!J Jesus Galvan - Marc Narciso Dublan
One exclamation mark for the strength of the Barbera del Valles 2 0 1 7
move, another for the realism. White already
could have aimed for a draw with 52.<&d6!,
inviting a perpetual, but here it is so tempting 8
to force the promotion of the c-pawn, that you 7
can forget that the road back is long and the
6
black connected pawns are strong.
5
54.&d8? loses to rolling the g- and h-pawns
4
up the board. For example: 54... h5 55.c8=®
Sxc8t 56.&xc8 g4 57.Se8 h4 58.Sh8 &g5 3
59. <&d7 g3 60.&e6 &g4 and it is all over. 2
1
abcdefgh
 Chapter 4 - Rook vs Pawns 89

Black lost after giving up the rook. The following endgame follows much of the
same narrative as the two previous games.
7O...Sxd7t? 71.&xd7 &b5 72.<&d6 b3 The players are much stronger, with the white
73.&d5 &b4 74.<&d4 a4 75.Sb8f &a3 player arguably world class. And although the
76.<&c3 &a2 77.Sb4! <&a3 78.Sh4 b2 mistakes are more complex, the guidelines for
79.&c2 how to play the position are entirely the same.
1-0
Jorden van Foreest - Jure Borisek
Again, delaying the moment where the rook is
Terme Catez 202 1
surrendered was the way to hold.

70.. .5h6t! 71.&e5

White cannot hide from the checks. He has
to approach the black rook.

71...5h5t 72.<&f4 Sh4t 73.&g5 Sd4

74.d8=®t Sxd8 75.Sxd8 a4!

abcdefgh
We enter the game at a moment of balance.
White is a pawn down, but his king is more
active and he is able to create a passed pawn far
from the black king’s sphere of influence.

32.b4 axb4f 33.&xb4 Sxg2 34.a5

abcdefgh It could be argued that 34.Sa3 was more
accurate, but compared to the game, Black is
Pushing the pawn furthest away from the
a tempo better off after 34...Sxh2 35.a5 Se2
white king.
36. a6 Se8 37.a7 Sa8 38.&c5 39.&xc6
&e6 when the draw is fine.
76.&f4 a3 77.Sa8 <&c5
Going to b5 also works, but this is the
principled move.

78.<&e3 &c4 79.&d2 <±>b3 80.<&cl <&a2

Black holds. For example:

81.Eb8 b3 82.Sb7 b2f 83.&c2 bl=®t

84.§xbl
Stalemate.
90 Conceptual Rook Endgames

abcdefgh abcdefgh
34...Sxh2? Black has two cards to play for: the race to the
Although this is not a losing move, it is 1st rank and the fortress with the pawns on f5,
certainly a big mistake. Black did not have to g4 and h3. The latter is hard to achieve and the
get into trouble. first fails to the white king running back.

34...Sg4f! 35.&c5 Sg5t 36. &b4 Sg4f would 43.&c5! &e5 44.&c4! &e4 45.&c3!
have secured the draw already (see the chapters Insisting on making it to the 1st rank, rather
on checks, starting on page 143). than taking out the weakest link in the black
In the above line, notice that White would pawn chain.
lose the race after 36. &b6? S b 5 t 37. &a6 Sb2
and when White eventually gets his pieces After 45.Sa6? h4 46.Sxf6 h3, the h-pawn
untangled, Black will have advanced his army would secure the draw.
of passed pawns far up the board.
45...g5!
35.a6 Sa2 36.Sa3! Sb2f 37.&c5 Sb8 38.a7 Giving up on the idea of shouldering the
Sa8 39.&xc6 white king, as he is too late anyway.
We will return to this moment below.
Despite the earlier mistake, Black could still 45... e3?l 46.Sa6! would have made the
make the draw. conversion much easier for White.

39.. .h5? 40.<&b7 Sxa7t 41.Sxa7! 46.&d2! &B!

Because this is a race for the f3/g2-squares, The best practical chance, as the game
and because the rook is often best placed documents.
behind the passed pawns, the automatic
capture of the rook in these races, is with the Playing for the fortress does not work. After
rook. Chess remains complex, and automatic 46...h4 47.&e2 g4 48.Sh7 h3 49.Sg7 f5
is not always best. The gut feeling stats give it 50.&f2 &f4 51.Sg6!, Black is lost.
as 90% of the time as the right capture.
47.<&el!
41.. .6f7 42.<&c6t £e6 47.SI7? would again erroneously go for
winning a pawn, rather than winning the race.
 Chapter 4 - Rook vs Pawns 91

47.. -h4 48.Sxf6t &g2! 49.Sg6 h3 and Black f3-square from the king. 50.&e3 h4 5 1 .&f4 f5 !
makes the draw. This is the key move. To take the pawns, White
has to run in the wrong direction. 52.&xf5
47...&g2 48.Sa2f! &gl (52.Sa5 g3 53.&f3 f4 is quite similar) 52...g3
53.&g4 g2 54.<&h3 & h l ! 55.Sxg2 Stalemate.

50.<&e3t?
The wrong direction.

5O.Sa4! g4 5 1 .Sf4! was the winning idea again.

50.. .<&g3 51.Sal?

White was still winning after 51.&e2 &g2
52.Sa4!, but it is rare that a player spots an
idea like this after having missed it only a few
moves ago.

abcdefgh 51...6.2?
White did not win the race for the 1st rank, Black had the chance to draw once again:
but only one player from the black team made 51.. .g4! 52.Sglf &h2 53.&f2 f5! and the
it across the finish line. As the purpose is for a pawns are rolling.
pawn to reach the 1st rank, this does not quite
work. But surprisingly, brute force does not 52.Sa2f! &g3 53.&e2! <&g2
win for White. 53...h4 would give up on the race, and
after 54.<±>fl £f4 55.&g2 &f5 56.&h3 £g6
49. <&e2? 57.Sa6! Black would not be able to set up the
Black would also make a draw after: 49.Sf2? fortress.
h4! 50.&e2 g4 51.Sf4 f5! and the important
pawns will stay alive for long enough. 54.Sa4!
Van Foreest finally finds the right idea.
The remarkable 49.Sa4!! was the winning
idea. The key idea is zugzwang. After 49...&g2 54...B 55.Sa5! h4 56.Sxf5! g4
50. e2, Black loses because of zugzwang only,
something he would never be able to avoid, 8
as White can pass the move by shuffling the
rook on the 4th rank (not making it mutual
7
zugzwang). 49...g4 loses to concrete play. 6
5O.Sf4! and for example: 50...g3 51.Sh4 f5
5
52.Sxh5 f4 53.&e2 &g2 54.Sf5, winning.
4
49...£g2? 3
Black is given a chance to bring the pawns
forward and misses it. 49...g4! was the drawing 2
move. It both advances the pawn and takes the 1
abcdefgh
92 Conceptual Rook Endgames

57.S£2f! 43.&c6 &e5 44.<&c5

A cleaner win than 57.Sf4 directly.

57...&gl
57...ll?g3 58.&fl! is the key idea. The race
is over.

58.Sf4 g3 59.Sxh4! g2 6O.Sg4

1-0

Black’s way to draw would rely on accurate

play and a single tempo. Obviously, it was
crazy for Black to ever arrive in this position,
but although difficult to hold, it is still within
the drawing margin. abcdefgh
44...B!
The most difficult move. The key to
understanding it is to notice that the only
pawn hanging is the h7-pawn, which is the
backward pawn.

44...&e4? would lose to 45.Sa4f!!, when

White either wins a tempo, or is able to go for
the black pawns after 45...&e3 46. &d5.

45.Se8f
45.&c4 <&e4 46.&c3 &e3 47.Se8t &f2
48.&d2 f4 49.Sh8 h5 5O.Sh6 f3 51.Sxg6
abcdefgh h4 52.Sh6 &g2 53.Sxh4 f2 also draws, after
39...&r! 40.&b7 S£8! many only moves.
The rook can also go to d8 and g8, both
squares where a check on the 7th rank exists 45...&F4 46.&d4
afterwards, and where the rook will not get
skewered after ...&e6 is met by Se3t-

40...5xa7t? 41.Sxa7! e6 42. c6 would

transpose to the game.

41.a8=® Sxa8 42.Sxa8

(Almost) always with the rook!

42.. .<&e6!
Shouldering and the key difference from the
game.

abcdefgh
 Chapter 4 - Rook vs Pawns 93

46...&B!! 50...h3? 51.Sxh3t &g2 52.Sh8 and 53.Sg8f

46. ..h5? would lose to 47.&d3 f3, when wins.
White has 48.Sg8, picking up a vital pawn,
and 46...g5? 47.Sf8! followed by d3-e2, Also 50...&g3? 51.&d3 f3 52.&e3 and 51...h3
which Black will be unable to deal with, now 52.&e2 win for White.
the pawn on f5 is under fire.
51.Sxh4
47.Sh8 51.&e4 f3 also holds.
Black is holding a draw in all lines:
51...B 52.Se4f &d2!
47. e5 h5 and ...h4 should draw without Black holds.
effort.
As talked about in the Introduction on
47.Se3t &f2! 48.Sa3 is also not dangerous. Schematic Thinking, exchanges of pawns are
For example: 48...f4 49.&e4 g5 50.&f5 f3 often in favour of the defender. The following
51.&g4 &g2 and White should look for the example is one of these.
draw now.
Stefan Mazur - Vladimir Malaniuk
47.&d3 f4 48.&d2 gives Black a number of
Koszalin 2014
ways to draw. For example: 48...g5 49. e l
&g2 5O.Se2f <&gl! and the black king
continues to keep the white king at bay.

47...h5 48.Sh6
White cannot allow ...h5-h4.

48...f4 49.Sxg6 h4 5O.Sh6

abcdef gh
53...b2? 54.Sb7 &c2
54... h4 comes too late. After 55. f4! Sh2
56.&g5 White is in time to take the h4-pawn
and secure the draw.

55.Sc7t
abcdefgh This check is usually good, as it forces the
king a bit further away. However, here it does
50...<&e2!
not make too great a difference.
Shouldering.
94 Conceptual Rook Endgames

55.. .6dl 56.Sb7 hxg4t In the next example White is obviously

Or 56. „±cl 57.gxh5, with a draw. 57...Sd4 completely lost, but still she managed to hold
58.&g3 b l = ® 5 9 . S x b l t £xbl 6O.h6 Sd6 the draw, since Black struggled with the basics
61.&g4 for example. as explained above.

57.hxg4 £cl 58.&e4 Julianna Terbe - Elena-Luminita Cosma

This shouldering move is not needed, but a
Bucharest 202 1
good habit.

58.. .bl=®t 59.Sxblf £xbl 6O.g5

A draw was agreed.
l l
/2- /2

Black should not have allowed the exchange of

pawns and instead played:

53...h4! 54.g5
There is not time to play 54.&f4 at this
point, as Black has 54...Sh2! and the h3-pawn
falls.
abcdefgh
54...b2 55.Sb7 &c2 56.Sc7t <£><11 57.Sb7
£cl 58.<&g4 47.Sh8 b3 48.Sd8t &c4 49.Sdl b2 50.8bl
&c3 51.&B Eb4?!
Aiming to keep the king cut off on the 4th
rank, but simply a lost tempo, as the rook
cannot stay on the 4th rank.

The easiest win would have been 51...&c2

52.S e i b l = ® 53.Sxbl Sxbl!.

52.h4 <&c2 53.Shl b l = ® 54.Exbl <&xbl?!

Black is continuing to try to cut off the white
king, but the h-pawn will draw the rook away.

54...Sxbl! won without effort.

abcdefgh
55.h5 &c2 56.h6
58...Sd4t!
The big difference. After:

59. <&h5 b l = ® 6O.Sxblf &xbl 61.g6 &c2

62.g7 Sd8 63.&xh4 Sg8 64. <&g5 Sxg7t
65.&f5 Sh7 66.&g4 &d3
Black is clearly in time to win.
 Chapter 4 - Rook vs Pawns 95

In the next example, the race to the 1st rank is

8
decided by White keeping the king as far away
7 from the finish line as possible. Dhopade had
slowly seen his advantage evaporate, before
6
missing the final (and as so often difficult) win.
5
4 Swapnil Dhopade - Romain Edouard

3 Gibraltar 2018

2
1
abcdefgh
56...Sb6?
An understandable mistake. Black is trying
to stop the h-pawn as quickly as possible, but
this allows White to exchange the h-pawn for
the black pawn, giving her a basic draw.

Black was winning after 56...&d3 57. h7 Sb8

58.<&f4 Sf8!, where White is way short of
time. If the f5- and g3-pawns were not there, abcdefgh
the draw would be easy. But as long as the f5-
81.g7?
pawn survives, Black wins by taking the g3-
Probably the last move Swapnil suspected to
pawn with the king. And after 59.g4 fxg4f
have cost him half a point, before checking the
60.<&xg4 Sh8 wins easily. But it is a pattern
game with an engine afterwards.
I have seen frequently over the years. The first
win or even the first few wins, are easy, but
81...5g3!
if missed, each win will be more demanding.
White has to make concessions to make
Thus, it is valuable to play with concentration
progress, so all Black has to do is wait.
and take the chances when they are given, not
hope for an easier one down the line. They
81 ...&d4? loses to 82. 17 Sf2f 83. &e6 Sg2
rarely come.
84.&d6 and White wins on the queenside.
84.. .5g6t 85.&C7 &d5 86.£b6 and we are
57.&f4! near the end.
Terbe does not miss her chance.
82.<&f7
57...SJ16 58.&xf5 Sh5t 59. >f6! &d3
82.Sfl &xc5 83.&F7 &d4 84.Ef6 5xg7t
6O.g4 Sa5 61.g5 &e4 62. g6 a6f 63.&f7
85.&xg7 c5 and Black draws.
$6 64.g7 Sa7t 65.&g8 &f6 66.&h8 Sxg7
Stalemate.
82...Sf3t 83.&g8 Sg3 84.Shl &xc5 85.Sh6
Vi- 1/?
&d5 86.&f7 Sxg7t 87.&xg7 c5
Black draws in well-known fashion.
96 Conceptual Rook Endgames

88.<&fi6 c4 89.&B c3 9O.Sh8 <&d4 91.&f4 82.Sfl!

c2 92.Sc8 &d3 93.&B <&d2 94.Sd8f &c3 Sooner or later White will have to find this
Vl- 1/! idea.

The win is based on keeping the white king 82...6xc5 83.&fi6

as close to the 1st rank as possible, which is This is the key concept. It can be reached
the 6th rank. To do so, it is a mistake to push by many paths, but they all end here. White
the g-pawn prematurely; which is far from is winning.
intuitive!
83...<&d4
81.<&g7I! 83...Sg2 84.Sf5t Now the only way for
Preparing to assist the rook to get to g5, Black to avoid Sf5-g5 is 84...<&d6, but the
where it serves multiple functions. king is now cut off and White wins with 85.g7.

81.&I7 Sf2f 82.&g7 transposes.

81...Sg3
Black has nothing better than waiting. But
as this also fails, the alternatives cannot be said
to be inferior:

81...6d4 82.Sfl &xc5 83.&f6 £d4 loses to

the same point as the main line: 84.Sf4f! &d3
85.Sf5! and the prophylaxis against ...c5 and
the threat of Sf5-g5 wins.

81...5f2 82.Sgl &xc5 83.&h6 Sh2f 84.&g5 abcdefgh

Sh8 85.g7 Sg8 86.&f6 &d4 87.Sg4f! wins in 84.gf4f!
the same way as the main line. Check the checks! The black king is pushed
away from the c-pawn, thus winning White a
tempo.

84...&d3 85.SB!
White wins. For example:

85...5e3
Giving up the rook feels premature, as Black
cannot easily push the pawn up the board.
85...5xg6t 86.&xg6 &d4 87.&g5 c5 88.&f4
c4 89.Sf8 White wins by a tempo as seen
several places already.

abcdefgh 86.g7 Se8 87.Sc5 Sc8 88.<&e6 <&d4 89.Sg5

Sg8
 Chapter 4 — Rook vs Pawns 97

Pavel Eljanov - Arkady Naiditsch

8
Dubai 2014
7
6
5
4
3
2
1
abcdefgh
9O.Sg4f!
Winning an important tempo.
abcdefgh
90...&e3 Naiditsch lost the following endgame
The king has to seek to harass the white rook.
trivially.
90...&d3 91. d6 and White wins.
75...&xB 76.Sc3t &xg4? 77.Sa3 Sxa7
91.<&f7
78.Sxa7 g5 79.&c2 <&g3 80.<&d2 6 81.<&e3
91. d6? &f3 92.Sgl &e4 would allow
g 4 82.Sg7
Black to draw, as the white king cannot both
1-0
hold the pawn back and prevent the black king
from coming to assist on the kingside.
The draw was unsurprising. The point of
having a passed pawn is to queen it.
91...Sc8 92.<±»fiSI? c5 93.Sh4 Sg8!
The move that offers most resistance. 76...6g2! 77.Sa3 Sxa7 78.Sxa7 f3 79.&c3
£2 8O.Sal fl=® 81.Sxfl &xfl 82.<&d4 <&g2
94.§c4 Sc8 95.Sxc5
83.&e5 &B
White has to win queen vs rook, which not
This forces White to make a draw with
all GMs have succeeded in, but hopefully we
84.&f6, as 84.g5? would even lose after
will...
84...6g4 85.&f6 &h5.

The following endgame between two talented

young players is full of instructive moments
and finesses, and thus a brilliant example of
the richness of ideas in what appears to be a
basic race.
98 Conceptual Rook Endgames

Kulkami Bhakti - Gabriel Flom 42.. .b3! 43.Sxft>t &c7!?

This is not yet the mistake, but as all White
Gibraltar 2020
can do is give checks that give Black more time
to think, it is hard to criticise this decision.

43.. .6c5! 44.Sf8 &b4! gives the same win

as on move 47 for Black. The difference is
the king’s placement; but luckily this matters
little, as f8 or g4 are equally far away from the
b4- and b5-squares.

44.gf7t &c6 45.Hfi6t &b7 46.Sf7t &a6

47.Sftt

abcdefgh
42.Sf7?
We have seen this mistake above as well.
White spends too much time winning the
f6-pawn and gets herself into difficulties.

42.Sh3! to put the rook behind a passed pawn,

and to block the black passer, was a solid
drawing move.

After 42...<&d5 White plays 43.Sf3!?, followed abcdefgh

by g4-g5 transforming the extra pawn to
47.. .6a5?
something tangible, a passed pawn.
The winning path was:
47.. . 6 b 5 ! 48.Sd6
And after 42...Se4 43.Sg3 & d 5 44. h4 White
48.Sf8 <&b4! will lead to the same lines.
clearly has enough counterplay to draw too.
48...b2 49.Sdl Sc5
A top-20 student of mine was incredibly
42...SB5 feels like the principled move, forcing
pleased to have found an alternative win
the white rook to be passive. 43.Sb3 & d 5
here: 49...Se2!? 5O.Sbl (50.f6 Sc2 51.f7
44.&f2 &c4 45.Sbl b3 46.&g3 &c3 47.&h4
S c l 52.f8=® S x d l f 53.&g2 b l = ® is also
White will lose the rook, but she is doing
winning for Black, and far easier than could
just fine in the race and will draw without
be believed at first glance. The king takes the
difficulties.
g4-pawn and then skates to c l , where it will
no longer be effectively checked.) 5O...Sc2
51.Sxb2f Sxb2 52.f6 &c5 and Black wins
by a tempo.
50.<±>f2
 Chapter 4 - Rook vs Pawns 99

51... b l = ® 52.Sa8t &b6 53.Sb8t &c7

54.Sxbl Sxbl 55.ft> sl?d6 56.h4?
56. &g2! or 56.g5 &e6 57.&g2! would draw.

56...sl?e6 57.g5

abcdef g h
50.. .5c2t’!
Forcing the king to the 3rd rank, which
might look like an improvement for White.
But it also allows a lethal check, winning the
most important tempo imaginable.
The most natural sequence in the world
fails: 5O...Scl?! 51.Sd8 &c4? (Repeating
with 51...Sc6 52.Sdl and this time finding abcdefgh
52...5c2f! would still work.) 52.Sb8 We have reached the fortress pawn setup.
b l = ® (52...Sc2t 53.<±>gl &d5 54.f6 <£e6 But one thing is missing: the king needs to
55.g5) 53.Sxbl S x b l 54.&g2! followed by prevent the black rook from coming to h l . If
f5-f6, g4-g5 and h2-h4 would set up the the rook gets there, Black wins - otherwise it
impenetrable fortress. is a draw, as you can read about in Theoretical
5 1 .&f3 S c l 52.Sd8 &c4 53.Sb8 Rook Endgames.
53.Sc8f &d5 54.Sb8 b l = ® and Black wins.
53.. .5c3t 54.<&f2 Sb3 57...&B?
Black wins. White will be able to give up Allowing White to set up the fortress after
the rook for the black pawn, but only after all.
checking the black king until it reaches e5.
Black wins after 57...§hl!, which is only
48.Sd6 b2 possible because the king is not on g2.
48. . . 5 e l f 49.&f2 b2 5O.Sd8! is not different After 58. &g3 &f7 59.&g4 &g6 White is in
from the game, although it is worth noticing zugzwang, making the win elementary for
that 50.&xel b l = ® t 5 1 .&f2 loses, but only Black. 60. &g3 & h 5 61.&f3 Sh3t! Below we
to 5 1 ...Wb2f!, when White has no convenient shall see what would happen if it was Black to
moves. play.

49.Sdl Sc5 5O.&£2 Scl 51.Sd8 58.&g3 Sb3t 59.&g2

White is lucky to be able to lose the rook. The players shuffled around for quite a while,
before Black allowed the pawns to advance,
After 51.Sd5t? &a6 52.Sd6f &b7 the pawn only to give up the rook for them. Draw in
would be unstoppable. 8 3 moves.
100 Conceptual Rook Endgames

8
Unoriginal 2021
7
6
5
4
3
2
1
abcdefgh
66...£t?d7!!
But Black needs a second triangulation in
abcdefgh order to win.
This position arose in my analysis of the
66.. . <&e5?l is premature. White draws after
previous game and I realised that Black was
67. &e2!, if Black does not backtrack with the
winning because White was in zugzwang. But
king and get the triangulation right next time.
it was not immediately apparent to nie what
would happen if it was Black to move. There is
67.&f4 Shi 68.&g3
no move saying “pass” as in previous examples,
White is also lost after: 68. g6 Sxh4f 69.&g5
so I wondered if this was a case of mutual
S h i 7O.g7 S g l t 71.&h6 <±>e8!
zugzwang. It is not. But the way Black needs
to lose a tempo is sensational.
68.. .6d6 69.&g4 &e6
This is the key position. White is in
By the way, I intentionally broke the study
zugzwang.
convention of “White to play”, because this
position almost happened in the game and we
7O.sl?h5 &f7 71.&g4 sl?g6
care more about this than anything else.

6O...Sglt 61.&f4 &f7 62.&B 63.&f4

&e8 64.&F5 &f7 65.&f4 &e6
This is now possible.

66.&f3

abcdefgh
 Chapter 4 - Rook vs Pawns 101

We have reached the position Flom could After showing this “study” to World Champion
have achieved in the game, which was also our Steffen Nielsen, he told me that Kopaev had
starting point. But this time with White to done a similar study, but with the pawns in
play and thus in a horrible zugzwang. the centre. The winning method should thus
be considered entirely human, as the composer
72.£g3 £h5 73.£f3 Sh3t! found it without the help of an engine in 1966.
73...£xh4? 74.£g2! would allow White to It turns out that this double triangulation
make a surprise draw. works in the centre and all other positions too.
I should note that Danish GM Mads Andersen
74.£g2 Sxh4 found this manoeuvre too. This is impressive,
Black wins. but having seen move 56 of his game with
Korley (see page 168), you will notice that this
is not the first time he found an impressive
king triangulation in a rook endgame.
Chapter 5

Passed Pawns

As mentioned in the Introduction on Schematic Thinking, success in the endgame comes from
creating and exploiting passed pawns. This is a deeper truth about the endgame than anything
else, still I have seen world-class grandmasters fail to create or push passed pawns so many times
you begin to wonder if they all have that basic truth engraved in the way it should be.

The first example is a basic example of this.

Victor Bologan - Alexandr Triapishko

Sochi 2018

abcdefgh
63.16!
Passive defence has no future.

63...gxft> 64.&15 Sa6 65.&xfi6??

It is hard to imagine what Bologan expected would happen after this. Most likely he had too
little time to think - or just forgot to.

65.'&g6! and taking the h-pawn would create counterplay and draw without great effort.
104 Conceptual Rook Endgames

65...&c5t 66.&g7 &b5 67.Sf4 a4 68.Sf5t

sl?c4 69.Sf4f sl?d5 7O.Sxa4 Sxa4 71. xh6
sl?e6 72.sl?g7 Sa7t 73. g6 Sa3 74.h6 Sg3t
75.£h7 &£7 76.£h8 Sg8t 77.&h7 Sg6
0-1

In the next example, White has the advantage,

but his reluctance to use it gives Black a chance
to equalise.

Rudik Makarian - David Gevorgyan abcdef g h

Chelyabinsk 2022 47...f5! Pushing the passed pawn! Not with

the hope that it will queen, but with the belief
that its potential to queen will at some point
create counter-chances. 48. h6 f4 49.Sb3 Sf8!
50.h7 (5O.Sf3 &d6!) 5O...Sh8 51.&g6 <±>d6
Black will draw in the rook vs pawn race by
a tempo. It is highly convenient to have two
passed pawns to push!

44...ft> 45.&e4 Se8f 46.&d4 &c7 47.Sbl

£5 48.gxf5 gxf5 49.Shl Se4f 5O.sl?d3 sl?d6
51.h5 &xd5 52.h6
Vi— ¥2

abcdefgh White would have won if he had pushed the

passed pawn. For a grandmaster, this should be
43.&e4? Sb8!?
the first choice. But it is really hard to see what
In principle I like the idea of activating
he disliked about it.
the rook and using the king to defend the
b6-pawn. It’s textbook after all.
43. h5! gxh5
43. ..f5 44.hxg6 fxg4 45.&e4 g3 46.Sb3!
However, at this point Black could also have
would soon leave White a pawn up.
played prophylactically with 43... f6!, when
after 44. h5 gxh5 45.gxh5, he would be able to
43...Sb8 would be met with 44. h6.
force a draw with 45...Sh7!, with an easy draw.
Threatening g4-g5, cementing the pawn and
clearing a route for the king to g7. 44...f6
44.&f4
45.Sbl Sh8 (45...Sb7 46. h7! is a neat little
44. h51? feels more challenging. But Black is
tactic) 46.Sxb6f &c7 47.Sxf6 Sxh6 48. g5
in time. After 44...gxh5 45.gxh5 &c7 46. &f5
and White wins.
Sh8 47.&g5 he will only have to find these
thematic moves:
44. gxh5 ft>!?
An attempt to avoid zugzwang.
 Chapter 5 - Passed Pawns 105

45.h6 5 46.SB1! Krzysztof Jakubowski - Michal Koziorowicz

46.Sb3? would allow Black to escape after
Poland 2022
46. . .f4 47.Sh3 Sh7 48. &e4 f3!, and we can
see that there was no advantage to putting the
rook on the 3rd row.

46.. .B5

abcdefgh
34...Sxf4? 35.b4! Se4f
It is too late to push the h-pawn. Black
would have to bring the rook to the queenside
abcdefgh to stop the b-pawn, after which the white king
can come over to stop the h-pawn.
47.Sgl!
White wins. For example:
36.&d3!
36. &f3? would be a mistake, as the king will
47...Sh7 48.Sg6t &e7 49.Sg7t
fail to block the h-pawn’s advance. 36...Sh4
White’s passers are further up the pitch,
37. b5 &xg7 38. b6 Sh3t 39.&g2 Sb3 4O.a4
ready to queen.
Sb2f! Black will keep checking until the white
king reaches the d-file, when the h-pawn will
Hie quality of passed pawns
start its surprisingly short sprint to the 1st
rank.
We have previously discussed how the creation,
pushing and promotion of passed pawns is the
36...Sel 37.b5 S d l f 38.&c2 Sd8
core purpose of most endgames (we will talk
more about promotion of pawns in the next
chapter). From this flows naturally that we
should be looking to push the most valuable
pawns. Chasing material is a middlegame
strategy in many circumstances, which we
should focus less on in the endgame. The
following endgame sees White winning, when
Black wastes time, rather than getting his
strongest pawn up the road.

abcdefgh
106 Conceptual Rook Endgames

39.Sc6 37.&B Sg6 38.Sc2

39.Sa4! is the principled move, but the
position is entirely winning no matter what.

39.. .6xg7 4O.b6 41.a4 &e7 42.a5

&d7 43.Sc7t &d6 44.a6 Hb8 45.a7! Hxb6
46.Hc8
1-0

Black could have secured the draw by playing

to the strengths of his position.

34.. .h5!

abcdefgh
38...Sgl!
The rook will go behind the white passed
pawns, securing an immediate draw.

Passed pawns must be pushed

The first of two mini-chapters on pushing the

pawn forward. The idea in the first of these
is as contentious as it will be in the second;
but here we go. Passed pawns must be pushed
abcdefgh - except when you lose control. It smells an
awful lot like Truman’s request for a one-
The key points of pushing the h-pawn are
armed economist, being tired of all these “on
that it is already passed and it will have to be
the other hand” reservations.
the white rook that comes home to stop it.

35.Sc6 So, let’s be clear. A pawn is generally better

35. b4 h4 36. &f3 Sg6! is quite similar. further up the board. But there can certainly be
situations where we lose control of the pawn,
35...h4 36.b4 important squares it was controlling, or even
36.Sc3 Sxf4 37. b4 is also too slow. Black simply lose it. Later we shall also see examples
will make the draw after 37...&xg7 38.Sb3 where the rook is sort of trapped in front of
Sf5!? for example, when 39. b5 h3! and 39.a4 a passed pawn on the 7th rank, leading to
Sf4 followed by ...h3 lead to immediate draws. zugzwang. The first example is a great example
of White missing a win by not pushing the
36...h3! pawn when he had the chance.
The simplest. Black also holds in the wild
variations where all the pawns are in play, but
the margin here is wider.
 Chapter 5 - Passed Pawns 107

Luka Budisavljevic - Alexander Donchenko

8
Biel 2 0 2 1
7
6
5
4
3
2
1
abcdefgh
35...&ft>?
The wrong direction with the king. It has to
abcdefgh be able to attack both white pawns.
32.&£2? 35...&h5 36. &c3 &g4! was the way to fight
for counterplay again, drawing as above (with
A natural-looking move, but as so often, the
minor variations).
focus is on the wrong pawn.

36.&c3?
32...Sa5 33.Sal Sa4!!
Budisavljevic missed a win which we all
The big, surprising difference.
would have missed...
34. sl?e2
36. a7! &e5 37.a8=® Sxa8 38.a4 Sa5
The key point of the black defence is
shown in the following variation: 34. a7 &g6
35. a8=® Sxa8 36. a4 Sa5! Blocking the pawn
immediately. 37.&e2 & h 5 38. &d3 g4
39. &c4 h5! 40.<&b4 Sa8 41.a5 &f3 42.Sa3
&g2 43. h4 &g3 44.e4f &xh4 45.exf5 &g5
Black makes the draw. The variation could
have been quite different, but the result would
be the same.

34...sl?g6 35.&d3
abcdefgh
So far it is all possible to find. But now...
39.&c3I!
39. c4? &d6 holds narrowly for Black.
39...&e4
39... d6 would allow 4O.Sdlf! when the
rook comes to d4 and gives White a chance
to keep control and regroup the king to f4.
108 Conceptual Rook Endgames

40.<&b4 Sa8 41.Sa3

White is winning, although precision is still
required.
4l...&f3 42.a5 h5 43.a6 &g2 44.Sa2f &f3

abcdefgh
37.a5
abcdefgh The pawn is one row higher than in the
game, which is a big enough difference.
45.&c5!!
Aiming for the kingside.
37...Sa6 38.sl?e2 &h3 39.&B &xh2
45...&xe3 46.&d5 f4 47.Sa3t &f2 48.&e4 f3
4O.Sa2f &h3 41.&f4
49.Sxf3t &g2 5O.Sa3 &xh2 51.a7 White wins.
White wins by first eliminating the h-pawn
and then exploiting that the black king is
There is a lot to be happy about with the white
entirely cut off by moving to b7 with the king
position below, but the moment the a-pawn
and queening the pawn. Individually, we could
was pushed too far, he entirely lost control.
have found all of these moves. But finding all
of them is close to impossible.
Alexander Krastev - Yannick Pelletier

36...sl?e5 37.sl?b3 Sxa6 38.a4 sl?e4 39.Sel Corti 2022

Sa8 40.&a3 h5 41.&b4
Having passed the time control, White
realised there was no chance of winning the
game and offered a draw.
Vi-Vi

White could have won by advancing the

a-pawn.

32.a4! Sa5 33.Sal &g6 34.&£2 &g5 35.a7

&g4 36.a8=® Hxa8

abcdefgh
48.a6? B 49.Sc2 <&d7 5O.Sh2 <&e7?
This move makes little sense.
 Chapter 5 - Passed Pawns 109

The black king’s job was to prevent the a-pawn 53.. .5b3t 54.&c7 Sc3t 55.&b8 Sb3t
from promoting (as it cannot help the f-pawn). 56.£c8 Sc3t 57.&b7 Sb3t 58.&c6 Sc3t?
See Chapter 13, page 189. 58...Sa3!, as above.

50.. .6c7 5 1.2h7t &d8! This was the difficult 59.&b5! Sa3 6O.a7?
move. Black is drawing. 6O.Sb2! was winning once again.
But not 51... b8? 52.Sf7 &a8 53.a7 and
White wins. Black cannot prevent the white 60.. .£2 61.Sx£2 Sxa7
king from going for the e-pawn. The black The game was drawn 50 moves later.
king is entirely sidelined.
51...&c8!? also works. The king is close White was winning in a lot of ways.
enough to make it to e8 in time from c8 too.
But it makes little sense. 52.a7 f2 53.SI7 Sxa7 48.Sc2
54.Sxf2 S a l Black is making a draw. If the 48.Sh7P to regroup the rook to f6 was
white king makes it to d6, Black has ...Sa6t, perhaps an even simpler win. Once the rook
and if the black king is on e7, Black can give is on f6, White plays &b5-b4, when the black
checks from behind. rook has to decide if it wants to defend the
f-pawn or leave the a-file.
51.&b6?
51.Sb2! was the winning idea. There are no
48.. .5b3t 49.&c4 Sa3 50.&b4 Sal 51.SE2
checks from the back. See more in the short
The f-pawn drops and White wins.
section on Rook on the Adjacent File, page 353.
I am not sure there is a lot of advice available
51...5b3t 52.&c6
in this chapter. I would suggest that you check
the games Laznicka - Naiditsch (page 381)
and Kravtsiv - Ding (page 372) in the final
chapter in this book to see more examples of
the importance of pushing pawns.

King in front of the Pawns

In his modern classic Small Steps to Giant

Improvement, Sam Shankland explained how
we frequently see the king being best placed in
front of the passed pawn. The reason for this
is simple: the goal is for the pawn to advance
abcdef gh safely. To do so, we need control of the squares
in front of the pawn. The king cannot do this
52...Sc3t? from behind the pawn, it has to do so from
52...Sa3 53. & b 5 &d8!! was the draw. See the side of the pawn, or in front of it. And
more on this type of manoeuvre in the chapter when it is in front of it, it is always only one
on The King’s Role. square away from where it wants to be, while
shouldering the opponent’s king away.
53.&b7?
53.&b5! Sa3 54.Sb2 was winning again.
110 Conceptual Rook Endgames

In the two examples we shall see here, advancing the pawns. 5O...Sa2 is probably best
connected passed pawns are involved. The met with 5 1 .Sxa2, but moving the rook along
king often takes a zigzag manoeuvre in front the 1st rank will also do it.
of the pawns for several reasons. One of them
is that the rook is behind one pawn and we 5O.Sgl <&£3 51.a7 Sa2 52.<&b7 &f2
therefore want to push the other pawn. The 53.Sxg2t &x g 2 54.a8=® Sxa8 55.&xa8 g5
king can guide, protect and assist them from Sometimes you wonder if people are
the space in front of the pawn. And usually it unfamiliar with the option of resigning. But
wins a tempo, as the rook will have to spend certainly, weird things have happened before
this tempo to get behind the other pawn. and today could be your lucky day!

Jan Sodoma - Erdene Baasansuren 56.b6 g4 57.b7 g3 58.b8=® <£hl 59.®h8f

<£gl 60.®d4f & h l 61.®h4f &g2 62.<&b7
Pardubice 2017
63.®f6t &e3 64.®g5t &f3 65.®Bt
&e3 66.® g4 &fl 67.®f4f
1-0

The next example is far more complex. It could

also have been placed in the later chapter on
drawing against connected pawns, as many
themes covered there are visible here.

The game is from the mid-1990s, at the

time when Topalov went from being another
promising teenager to the number three player
in the world.
abcdefgh
Veselin Topalov - Alexander Beliavsky
46.&B6!
46. b6? illustrates how slow it is to push the Linares 1995
pawn already under the supervision of the black
rook. 46...&f4 47.&a6 g3 48.a5 g2 49.Sal &f3
Black is in time to make a draw. After 50.&b7
Sa2! Black makes a draw immediately, and
50. b7 just delays things after: 50...d5 51.&a7
Sa2! 52.Sxa2 g l =®t 53.&a8 ®g6!! 54.a6 ®c6
White has no way to make progress. The most
obvious continuation is 55.a7 ®e8f 56.b8=®
®c6f with perpetual, although everything
ends in a draw around here.

46...<&f4 47.a5 g3 48.Sal g2 49.a6 S£2

After 49...&f3 5O.a7 we see the great
difference between the two methods of abcdefgh
 Chapter 5 - Passed Pawns 111

61...516? 62...Efl!
A passive move. There is not enough time for This is direct prophylaxis against b4 and
Black to manoeuvre the rook to e6, although &a3, as given above. White will therefore
it would be the perfect square for the rook. have to find a rather beautiful, but far from
Rarely do manoeuvres work in a race. simple win.

Passed pawns should be pushed. There is no

reason not to get going here. 61...e4! is the
principled move.

After 62.Ed8t &c5! Black manages to create

counterplay with all of his pieces. The king
against the white pawns, the e-pawn diverting
the white forces and the rook harassing the
king and attacking the pawns.

abcdefgh
And 62. a6 is rather harmless.
Black can bring the king back immediately, 63.Eb8!!
but simplest is a version of the game: 62...Ef7!, How easy would it be for Topalov to go to
when the a-pawn is far away from the white the b-file, the move after placing the rook
king. The black king will arrive at the queenside behind the passed e-pawn? Especially since
sooner. 63. a7 Ed7! The e-pawn secures a draw. it is based on a long variation that is not
The moment the white king comes to the obviously winning.
d-file, Black will be able to go to the c-file with 63. a6? looks natural, but after 63...Ef2f!
a discovered check and then the b-file, ready to 64. &a3 Ef6! 65.Ea8 &c5 we have a direct
eliminate the a7-pawn. draw.
And after 62... &c5, Black can draw with 63.Ed8f? &c5 64.Eb8 looks like it is
63.Eb8 Ef2f 64.&a3 E f l . winning a tempo, but the black king is well
placed, taking care of the white passers.
61...5f2t 62.&a3 e4 63. a6 was also 64... e4! There are two tries from this
holding. position, both are very instructive.
a) 65.b4t &c4 66.b5 e3 67.a6 e2 68.Ee8
62.Se8? e l = ® 69.Exel Exel 70. b6 Ee2f is a famous
62. b4 &c4 63. &a3 looks more natural to draw, first described by Horwitz and Kling
me. White wins rather mechanically, as far as in 1851. Checks and the threats of mate will
I can tell. rob White of the necessary time to advance
the otherwise unstoppable pawns. 71.&cl
The game still wins against best defence, but &c3 72.&dl &d3 73.b7 Eh2 74.&cl
would require at least one difficult decision. &c3 and so on. In those old studies, White
managed to get both of the pawns to the 7th
62.. .e4? rank, so in this discount version, Black is
A stronger defence was possible with: laughing.
b) Therefore, 65.a6!? seems more testing.
65.. .5f2t 66. &a3 Efl 67.&b2 Ef2t 68.&c3
112 Conceptual Rook Endgames

Ea2! 69.b4f &d5 7O.b5 e3 71.&b4 Eb2f Black draws after 65.&a4? &c4 66.Exe3 Ba I f
72. &a5 e2 73.Ee8 &c5! The counterplay 67.Ea3 Ebl!.
against the b5-pawn secures the draw. White
cannot make progress. 65. .. e4 66. a6
63. . .e4 64. a6 Bf7
The key idea behind 63.Eb8!! is revealed
after 64...Ef2f 65.&a3 Ef6, when 66.Eb4f!
and 67. a7 wins.
65.Eb7 Bf2f 66. &a3 e3 67.a7 Efl 68.&b4
Eal 69.Ed7t &e5 70.&b5 &e6 7 1.Ed3! Exa7
72.Exe3t &d7 73.&b6
White is in time for a theoretical win.

63. b4 e3
63...&c4 64.&a3! would also win for White.

64. <&b3 Sfl

With the rook ideally placed and Black abcdefgh
having no direct way to improve his position, 66.. .Eal?
Topalov now showed his poor endgame This looks like an only move, and it is easy to
intuition. This was probably one of the games believe that Beliavsky might have been making
that convinced Danailov to send Topalov on automatic moves by this point, not realising
a two-week trip to Moscow to learn to play that he had a surprising defence.
endings with Mark Dvoretsky. Once there,
Dvoretsky noticed that Topalov was struggling 66...e2! 67.Ee8t
with calculating variations and focused as
much on that as on endgames. The year after,
Topalov beat Kasparov and shared first place
with him in Dos Hermanas, and won the
tournament in Novgorod.

65.Sd8f?
A rather peculiar mistake. Since this was
the mid-90s, it would be fair to guess that
rampant time trouble was not the culprit. I
think Topalov calculated a clear line and went
for it, missing that Black has an escape hatch. abcdefgh
67...&d5!!
The most natural continuation is 65.a6! Eal Following our guidelines of letting the king
66. Ed8f only now, when the white pawns are deal with the passed pawns.
connected and the rook has lost its flexibility 67...&d3? loses to a familiar tactic: 68. a7
and already committed to the a l -square. After S a l 69.a8=® Sxa8 7O.Sxa8 e l = ® 71.Sd8f
66.. .6e4 67. b5 e2 68.Ee8f we have transposed
to the game.
 Chapter 5 - Passed Pawns 113

68.Sxe2 £c6 69.Se5

69.Sa2 Sb I f 70.&c2 S h i also holds, on
account of 71.a7 Sh2f.
69...&b6 70. b5 S h i
White has two extra pawns, but they are
entirely dominated. We shall look more at this
endgame on page 302, when everything that is
solid shall melt into air...

A disclaimer is in order: tablebases do not

distinguish between 66. ..e2 and 66.. . S a l . abcdefgh
Computers may analyse chess better than we If you ask computers, this endgame is a draw.
do, but they do not understand the concept My experience is somewhat different. My
of a game. friends Sam Shankland and Abhi Mishra have
both lost this endgame; the latter to another
67.b5 e2 68.Se8f <&d3 friend of mine, Boris Gelfand, who has won
68...&d5 69.Sxe2 &c5 7O.Sa2 and White the endgame a few times (we annotated the
wins. win against Jobava in Decisions in Major Piece
Endings, pages 213-222).
69.&B4 &d2 70.&c5
In theory it should be a draw. The black king
hides in the faraway corner and the queen
will be able to give a lot of checks. But the
difference between theory and practice has
always been none in theory, but immense in
practice.
After Abhi lost to Boris, Sam convinced me
that in practice the endgame is not holdable,
even in classical chess. I was all fired up to write
an article about it and got ahead of myself,
agreeing to write it for a magazine, before I did
the research... (sigh).
The reality is that it is holdable. It is quite
abcdefgh complex. And I did not fully understand
7O...el=®? why it is holdable in some situations, but
The final mistake, where the tablebases immensely difficult in others. It has a lot to
finally give up on Black’s chances. do with the correct waiting moves in situations
where checks are not working. But I did not
Black has a few versions of transposing into understand, in a way I could put into words,
the queen endgame with: 7O...Sa3! 71.&b6 how you can create this scenario. Simply,
(Black would be in time to make the draw it appeared that some held it seemingly
after 71.Sxe2f &xe2 72.&b6 <±>d3!) 71...Se3 effortlessly, while others lost without a real
72.Sxe3 &xe3 73.a7 e l = ® 74.a8=® chance.
 After 20 minutes I decided that if I needed
many hours to understand it, a magazine
article aimed at an under 2000 rated audience
would sink.

So this endgame is a theoretical draw. Yet

I believe the chances of Beliavsky holding it
would be less than 50% at the time of the
game. Topalov had youth, energy and cunning
on his side. And Beliavsky seemed spent,
failing to even reach it.

For this reason, I seriously believe that the abcdefgh

losing mistake was on move 66, while here he
This is the moment Dvoretsky comes in. He
could have retained a fighting chance. But if
points out that the clever-looking 72. b6? is a
you call this endgame a draw, you are banned
mistake because of 72...2a5t! and Black makes
from talking about advantage ever again. Only
a draw instantly. Instead, we use the good old
winning, drawn and lost can exist in your
zig-zag motion with the king.
vocabulary.
72.&B6! &d2 73.&a7
71.Sxel &xel
1-0
Chapter 6

Promoting Passed Pawns

In this chapter we shall investigate various challenges and themes relating to promoting pawns.
While the main focus is on the collaboration between the king, pawn and rook, we will start with
a few peculiar tactics.

While the key strategic aim in endgames is to promote pawns, the core goal remains checkmate.
Promoting a pawn can be the end of the struggle, but often it is not. At times both players get
a queen, at times a queen cannot be had, due to the threat of mate, and we have to promote to
a knight (sometimes losing; often not). Chess is a game of obvious surprises. A move can be a
surprise when played and obvious the second after.

The following game is hopefully a valuable lesson in remaining concentrated until the game is
over.

Peng Xiongjian - Shahin Lorparizangeneh

Orbis 2019

abcdefgh
Both players have played hard for a win at various stages of the game, but at this point it seems
that a draw is inevitable. After repeating moves, White tried a final trap.
116 Conceptual Rook Endgames

79.Sa2f &h3 8O.Sa3t <&g2 81.<&g4!? h l = ® 38.Sf4f!

Black promotes the pawn with the Ouch! Black has to take the rook, or the
expectation that White will take the queen and f3-pawn falls.
then the g7-pawn, resulting in a vacuum draw.
38...&xf4 39.b8=Bf &£5 4O.@c8f <&e5
82.Sa2f 41>c7t &e4 42>e7t &d4 43.Bb4f &d5
Leaving Black with two legal moves; one of 44>b7t
them losing. Black resigned.
1-0
82...<&gl??
Careless and unimaginative.
Black should have forced a perpetual.
82... fl was still a draw.
37...£2f! 38.<&e2 Sd2f
83.&g3! 38...Sdl? fails to 39.Sf4f! again.
Black is lost! This is possible due to the
g7-pawn. Without it Black would have ...®h8! 39.&xe3
defending the a l -square.
1-0

In situations like the next one, where both

sides are about to queen, tactics are often front
and centre.

Wei Yi - Zhou Jianchao

Xinghua 2015

abcdefgh
39...fl= t!
Underpromotion, here with perpetual check
as the saving grace. The unusual thing about
this perpetual is that the knight is on the 1st
rank and the king on the 3rd. Usually, it is the
other way around.

39. ..fl =®? would lose to 40.&xd2, although

abcdefgh there are more moves to be played before Black
37...Sd2? has to resign.
With the double intention of playing ...f2f
and . . . S d l f , and ...2xh2 and ...Shi#. But a 40.&B &xh2t
simple trick has been missed. White can only avoid the draw by allowing
37. ..e2? would fail to 38.Sbl. mate, which is not recommended.
 Chapter 6 — Promoting Passed Pawns 117

The following example is even more unusual. I feel great sympathy for White that he
Again, both players are about to queen and at misplayed the position. Not because he lost;
first it looks as if Black has the better hand. we have all done so and this is a part of the
Indeed, in the game, he won quickly. game. But because he missed the chance for
immortality with:
Mariano Ortega Amarelle - S. Yogit

Ciutat de Mont 2016

43.c8=®f! Sxc8

8
7
6
5
4
3
2
1
abcdefgh
abcdefgh
44.Sa7!!
43.Sh8t?
A strange-looking move. Certainly, the rook
43.Sg7? was also possible. Black would play
belongs behind the passed pawn? No! After
43. ..&f8!, when White should give up the
44.Sb7? b l =®! we will have a draw similar to
c-pawn to stop the b-pawn.
others seen above. Now, White wins.

43...±f7 44.c8=B?
44...bl=B 45.h7 ®e4f 46.&h2
A strange decision.
A remarkable position.
44.Sb8 would give good drawing chances.
The next example is a classic I remember seeing
44.. .Hxc8 45.Sxc8?
when I was a child. It made a great impression
There was still time for 45.Sh7t g6
on me, but as so often, what we learn in our
46.Sb7 xhb 47.Sxb2, when White should
childhood seems simple and convincing at the
hold, although Black will try.
time, but when we revisit it later in life, it is far
more complicated.
45...bl=®
We can now guess that White missed the
check on b7. He is just lost.

46.Sc7t &g8 47.h7t &h8 48.&h3 Bfl t

0-1
118 Conceptual Rook Endgames

Alekhine - NN I.a5!? Eh5 2.&c2!

With dual purpose.
Groningen (simul) 1933

2.Sh3? is a mistake. White will not win the

pawn ending with an h-pawn. The finish is
rather curious. 2...Sxh3 3.gxh3 &d7 4.b3
&c6 5.&e3 &b7 6.&f4 f6 7.h4 &a6 8.&g4
&xa5 9.h5 f5t! An important intermediate
move. 10.&g5 gxh5 l l . & x h 5 &b6 12.&g5
£c6 13.&f6 £d6 14.f4! &d7 15.&e5 <±>e7!
Stalemate.

2...Shl
2...Sh8 3.&b3 and White wins.

3.Sdl b3t
abcdefgh
The pawn ending after 3...Sxdl 4.&xdl
In this simul game, Alekhine found a nice &d6 is different. White has the g-pawn and
promotion tactic. wins. 5.b3 &c6 6.g4 f6 7.&e2 &b7 8.&e3
&a6 9.g5! and the king breaks through.
I.g4! Se4 2.a5 Sxg4 3.a6 Sh4
Other records indicate that the game went: 4.&d2 Sh8 5.&c3 Sa8
3.. . 5 g l 4.a7 S a l 5.Sa3! But it is possible
that the game was a construction altogether,
as Alekhine is suspected to be a little too
creative with the descriptions of his own
achievements...

4.Sd8! &xd8 5.a7

And Black resigned. But looking at this with
Sam Shankland, we both noticed that this was
as a minimum premature.

After:

5...5xc4 6.a8=®t &e7

abcdefgh
The rook will transfer to the 5th rank and
defend the c5-pawn. Whether or not Black’s 6.Sal
fortress can be breached takes some work with The tactics are over and White has a massive
a computer to determine. Karsten Muller advantage with his strong passed pawn and the
pushed me to do it, and I can confirm that it more active rook. The winning line is simple.
can be breached, slowly... and not easily.
6...&d7 7.&xb3 &c6 8.&c3 e5 9.b4 cxb4f
Looking deeper, Sam and I found another 10.<&xb4 B 11.a6 e4 12.c5 f4 13.&c4 e3
winning line for White. 14.fxe3 fxe3 15.Sa2 &c7 16.&d3 &c6
 Chapter 6 - Promoting Passed Pawns 119

17.Sa5 Ronan Le Goff - Abyl Kizatbay

And so on... White wins.
Helsingor 2018

Sam thus came up with an improved version

after l.g4! Se4:

2.£3!

abcdefgh
5O.Sg8?
50.c6! with the threat of 5 1 .Sfl , was winning
on the spot. After 50...§xc6 51.f7 Sc8 52.§g8
abcdefgh the pawn queens.
White is closing one path back for the rook.
5O...Sf7 51.c6 e5! 52.&g2 e4 53.h4! e3?
Another exists, but it is less effective.
Black could have drawn with 53...Sc7!
followed by further accurate moves.
2...Sxc4 3.a5 b3 4.Sxb3 Sa4 5.Sa3
White now wins the pawn ending with not
54.Se8 &g6
one, but two distant passed pawns on the a-file
54...5c7 would also have lost. White has
after 5...Sxa3 6.bxa3.
55.f7! 5xf7 56.Hxe3 Hc7 57.Se6f! followed by
bringing the king in.
Black can try 5...Sb4 and a desperate dash to
a8 is possible, but it does not hold. 6.&c3 &d6
7.a6 §b8 8.a7 §a8 9.g5 8
7
Two passed pawns working together
6
In the next example White has two 5
well-advanced passed pawns, while the black
king is brutally cut off. It should be possible 4
to queen one of the pawns without too much 3
of a fuss.
2
1
abcdefgh
120 Conceptual Rook Endgames

55.Se7? 7O...Sxe8t 71.&xe8 &g3

White was still winning after: 55.§xe3! Sc7 Black draws.
56. §e6! &f5 57.§d6 &e5, when he has the
clever intermediate move 58.F7I! Sxf7 59.§d7 72.Sg7t 73.Sh7 £g4 74.<£e7 h4
Sf4 60. c7 Sc4 61.§xh7 and the two passed 75.&e6 h3 76.&e5 &g3 77.&e4 h2 78.&e3
pawns win the game. &g2 79.&e2 hl=® 8O.Sxhl &xhl

55...Sxl6 56.c7 S£2f!

Perhaps White was hoping for 56...§c6 The principle of this endgame is that Black
57. §e6t!. draws if White wins the black rook by taking it
with the king on the 8th rank. The technique
57.&g3 Sc2 58.&B e2 to win time in the race is to make sure the
The game was drawn almost 20 moves later. black rook is captured with the white rook.
White also needs to keep the king cut off.
We shall see more than once in this chapter
that promoting passed pawns is difficult and 69.&e8! Sei
can be mishandled even by World Champions. There are other tries, but they are thematically
In the next example Magnus Carlsen has the similar. In some lines (as below) the white rook
excuse of limited time, as the game was a comes to g6, threatening Sg6-e6, forcing Black
rapid game, but he will still have been kicking to take the pawn on the 7th rank, thus saving
himself afterwards for messing this up. a tempo for White.

Magnus Carlsen - Wei Yi 7O.Sg8!

Internet 2022

abcdefgh
abcdefgh This is the idea.

69.Sf7?
7O...Se2
The wrong path.
A waiting move.
69...Sei 7O.e8=S?
After 70...&113 71. f7 h4 72.§g6! White also
70.§g7! still won.
wins.
 Chapter 6 — Promoting Passed Pawns 121

71.&f7 S£2t 72.&g6 Sg2f 73.&h7 Se2 the d-pawn will queen, but also where the black
74.e8=® Sxe8 75.Sxe8 &g3 76.&g6 rook has found itself in the most horrid place
h4 77.&g5 h3 78.Se3t &g2 79.&g4 h2 possible. After 57. &f3! Black would lose, were
8O.Se2t £gl 81.&g3 he to queen the pawn. So, the game should
A well-known theoretical position. The end in a draw after something like 57...Sb8
knight is trapped in the corner. 58.Sd6 c5 59.&e2 §e8f 60.&f2 §b8 61.&e2
with a repetition of moves.
81...hl=&t 82.&B £fl 83.Sg2
White wins. 54.axb5 &c3 55.Sc8f &b4 56.Sd8 &c4
57.Sd7 g6 58.Sd8 Sb6
When we have to queen a pawn, it is often
inevitable that we will have to hide the king in
front of the pawn when harassed by a perfectly
placed enemy rook. In this case, we should pay
attention to the details and make sure we do it
in the most efficient way possible. In this case,
this means making sure the pawn makes it to
the 2nd rank before getting blocked.

Max Blau - Vasily Smyslov

Moscow Olympiad 1956

abcdefgh
59.&B!
White escapes as in a miracle. Smyslov tried
a few things, but the deal is done.

59...&c3
59...Sxb5 60.&e4 is the key point.

60.&e3 Se6f 61.&E2 d2 62.Sc8f &d3

63.Sd8f &c2 64.Sc8f &dl 65.Sc6 Se2f
66.&fl Se3 67.&E2 Sb3 68.b6 Sb2 69.&fl
abcdefgh Sb3 70.&E2 S b l 71.Sd6 &c2 72.Sc6f &d3
51...&e3?! 52.Se8f &d2? 73.Sd6t &c3 74.Sc6t
Black could still return and go the other way Vi-Vi
around with the king.
The winning manoeuvre exploits the
53.B5! cxb5 temporarily half-closed c-file to make White
I presume that Smyslov was shocked he was advance the b-pawn. Then, by threatening to
not winning after 53...&c2 54. b6 d2 55.Sd8 take the pawn, Black forces the white rook to
Sf7 (55...Sf8 56.§xf8! d l = ® 57. b7 only gives the 5 th rank, where the checking distance is
Black opportunities to lose.) 56. a5 Sb7, where too short.
122 Conceptual Rook Endgames

51.. .6c3! 52.b5 52...Sb8 53.a7 Sa8 54.&a6 e5 55.fce5

52.a5 does not work. After 52...§f7! Black is &xe5 56.&B7 Sxa7t 57.&xa7 6 58.&B6 f4
ready to take the white queenside pawns and 59.Sc8 <±’d4!
win with the c-pawn. Not necessary, but good technique.

52...cxb5 53.axb5 d2 54.Sc8f &b2 55.Sd8 60.S18 &e3 61.&c5 f3 62.&c4 £2 63.Sx£2
&c2 56.Sc8f &dl 57.Sc7 Sb6 58.Sc5 &xf2
The black king is too close for White to */2-*/2
have any chances after taking the kingside
and giving up the rook for the d-pawn. Black’s The winning triangulation is highly instructive.
next few moves would be ...§xb5, ...§d5! and
...&e2, winning. But on the 5th rank, the rook 52.&c6! Sa3
will soon find that the checking distance is one Otherwise, White puts the rook behind the
row too short. pawn.

58...&e2 59.Se5t &d3 6O.Sd5t &c3 53.&b5! e5

61.Sc5t &d4 53. ..Sal 54.Sa4 g b i t 55.&c6 gb8 56.a7
The checks have dried up and White has no ga8 57.&b7 gf8 58.a8=B Sxa8 59.Sxa8!
defence against ...§e6-el. Almost always the correct capture. 59...'&xf4
60.&c6 White is in time.
The following missed win by my friend Sabino
is one we could all have replicated. 54.fice5 &xe5 55.Sa4 Sb3t 56.&C6 gb8
57.a7 Sa8 58.&b7 Sxa7t 59.£xa7! f5
Sabino Brunello - Erwin L’Ami 60.&c6 f4 61.£c5 £e4 62.&c4 &e3
63.&c3 8 64.ge7t
Reykjavik 2018
White wins.

The following example of conversion of an

advantage from Emil Sutovsky, before he
gave up playing for politics, is a great example
of how to use one passed pawn to remove
counterplay and win time to do what we really
want to do.

abcdefgh
52.&a5?
The king is now too far from a8.
 Chapter 6 — Promoting Passed Pawns 123

Dmitrij Kollars - Emil Sutovsky

White resigned in view of 60.§xfl §c8!.
Gibraltar 2018
0-1

The following game is from the 2021 World

Rapid Championship, where every move
is made with a ten-second increment. The
mistakes are easy to understand and forgive,
but also valuable as free lessons for the rest of
us. Black is in trouble. He is a pawn up, but the
f6-pawn is dangerously close to promotion and
Black has to find a way to get the rook back
quickly. If this was all that was happening,
this example would be in the chapter on Rook
behind passed pawns, but the miss by White is
what places this example here.
abcdefgh
43...f4! 44.Sxh5t &e4 45.Sh7 B! 46.Se7t Aditya Mittal - Andrew Tang
&d3 47.Sf7 &e3 48.Se7t &d3 49.Sf7 &e3
Warsaw (rapid Wch) 202 1
5O.Se7t &f4!
50...&d3 51.S17 would be three-time
repetition.

51.Sf7t &g3
Threatening to win the h-pawn, thus gaining
a tempo.

52.h5 £2! 53.h6 &g2 54.Sg7t 55.2f7f

&e3 56.Se7t &d3 57.Sf7 Se8 58.Sd7t
&e3 59.Sf7

abcdefgh
The unfathomably low rating of the Indian
Wizkid (1488) is his rapid rating. Don’t worry
about it.

abcdefgh
124 Conceptual Rook Endgames

66... 5 h i t ? Matthias Bluebaum - Shakhriyar Mamedyarov

An understandable check, but a losing one.
Warsaw (rapid Wch) 2021

66.. . Bf5? would be a total blunder. 67.B£2f

and wins.

67.&d2?
White missed the chance to play: 67. &f2l,
when the king is ready to help the rook get
to f2. 67. ..Sh5 68.(7 Sf5 69.&g2 and White
wins.

67.. . 5 A ! 68. Sei S£2f 69.Se2 S£3 7O.Se3

S£2t71.Se2

abcdefgh
White cannot be prevented from getting the
rook behind the passed pawn. But by playing 74.Sflt?
the moves in the right order, Black can use a It seems reasonable to believe that this check
check to prevent White from queening the makes things easier for White. But Black can
pawn. win on both sides of the board.

66.. .6g3 ! was the key move. After 67. S £2 74...<£e3! 75.Shl <&£2?
Se5t 68.&fl Se8 Black holds. 75... <&d4! would have won. Black is
threatening ... c3, winning a pawn, as well
The next example sees Black with a passed as ...Bg6t and ...Sh6. White cannot defend
pawn and the advantage. Still, White is within against both.
the narrow drawing range we so often see in
rook endings. Under time pressure, in a rapid 76.&c6 &g2
game, the players were not able to give their
best, but the game was fascinating all the same.

abcdefgh
 Chapter 6 - Promoting Passed Pawns 125

77.Sh4?
White should not have a chance with passive
defence in this endgame, but it turns out
it is more difficult to overcome than a first
impression reveals.

At this moment, when the black king is as far

away as it possibly can be, White can draw
with a continuation which is both simple
and logical, yet at the same time incredibly
surprising:
77.Scl! h4 78.§c5 Sxc5t 79.bxc5 h3 80.&b6!!
This is the truly surprising move. If the king abcdefgh
goes to b7, Black wins a tempo by moving
85.Sd2t?
the king with a check, which brings it closer White could draw with a theme we know
to the action in the queen ending, on the well from the chapter on Rook vs pawns.
queenside. This underlines how narrowly
White draws the queen ending. After 85.§h3 S h 5 86.&c6 White is just
80.. .h2 81.c6 hl=W 82.c7 waiting for Black to play 86...&g2, when the
The position is a tablebase draw, although in king is finally far enough away for White to
practice White might draw it half the time at play 87.§xh4!, and make a draw.
most.
85...&e3 86.Sh2 Sh5 87.Sh3t &e4
77.. .6g3 78.Sd4 h4 79.Sd3t &g4 8O.Sd4t The king is on the other side of the 3rd
&h5?! rank and closer to the 8 th rank, making it
Black needed to create stability. This could be
impossible for White to give up the rook for
done with: 80...&g3 81.Sd3t &g2 82.Sd2f
the h-pawn with the black king on g2 and the
&f3 83.Sd3t &e4 84.Sh3 S h 5 85.&b6 £f4 black rook having to do the recapture.
And Black wins.
88.&a5 &B 89.Shl <&g4 9O.Sglf &f3
81.Sd3 SB?
91.Shl h3 92.Sflt &g2 93.Scl h2 94.Sc2f
81...5g3?! would not make progress. After
&h3 95.Scl SB
82.§d8, Black would have to play 82...Sg5
0-1
83.Sd3 and then think again.
Returning to the starting position, White
81...6g4! would return to the right direction.
would have been able to hold with accurate
play.
82.&b6 &g4 83.Sd4f &g3 84.Sd3t &(2
84.. .*&g2 looks winning, as 85.§d2f &f3!
74.Shl!
returns to the winning ways.
Black is not allowed to advance his king or
But White can play 85.§c3! h3 86.Sc5
pawn without cost.
with a potential return to the miracle draw
on move 77, or another draw after 86... h2I?
87.Sxf5 h l = ® 88.&xb5 and Sf5-c5 with an
established fortress.
126 Conceptual Rook Endgames

74...-&B
The idea of ... e3-d4 is no longer in time.
White can move the king to c6 and then
defend against ...&C3 by placing the rook on
h3.

8
7
6
5
4 abcdefgh
3 8O.Sh5
The black king is too far away. White draws.
2
1 The following game was a rapid game, but the
ideas are useful to understand.
abcdefgh
75.£d7! Evgeny Postny - Romain Edouard
This was suggested by Sam Shankland. The
Oscaro 20 13
king cannot be better placed for a potential
queen ending.
8
75... g2 76.Scl!
7
The main point of this example. White needs
to queen a pawn to survive. He will never 6
be able to prevent the h-pawn from being 5
queened. So, the b4-pawn needs to get going.
4
76...h4 77.Sc5 Sg7t 3
The queen ending this time around is not
remotely dangerous. After 77...Sxc5 78.bxc5
2
h3 79. c6 h2 8O.c7 h l = B 81.c8=® Sh3t 1
82.&c7 @xb3 83. &b6 the draw is just around
abcdefgh
the corner.
6O.a6! Sxe4f 61.&b5 Sei
78.&c6 h3 79.&xb5 h2
 Chapter 6 — Promoting Passed Pawns 127

70.. .e4?
But the e-pawn has no value at all. Push the
pawn with potential!

70.. .6g4! drew.

71.&d4 £g4 72.&xe4?

White won on time, after having played a
bad move. The beauty of rapid chess...
1-0

The correct move was 72.&e3!, winning by

shouldering. 72...&g3 73.&xe4t and the extra
abcdefgh
tempo is everything.
62.a7?
This natural move is a mistake. The white Again, the side with an extra pawn
king comes too far away from the scene of unsurprisingly has the advantage. Giri took a
action. pawn in the opening, providing his opponent
with plenty of compensation, but also the
The winning idea is: 62.&b6! threatening responsibility to prove that the material was
to push the pawn at a moment when White invested soundly. Over time the compensation
is ready to put the rook behind the pawn. diminished and at the time where we come
62...§cl 63. a7 §c8 64.§a5 §a8 65.&b7 in, Black is down to only moves to stay in the
Sxa7t 66.§xa7! *&f5 67. &c6 and White wins. game. When he makes a small mistake, Giri s
iron technique takes him down.
62...§al 63.&b6 Sxa7 64.&xa7 &f5 65.&b6
&g4 66.Sd3 15 67.&c5 f4 68.gxf4 &xf4 Anish Giri - Maxime Vachier-Lagrave
Taking with the king or pawn does not
matter. The black king is well placed to help Baku 2015
the h-pawn either way.

69.&d5 h4 7O.Sa3

abcdefgh
48...&B! 49.&g3 &e4 5O.Sbl &d3?
The wrong plan.
abcdefgh
128 Conceptual Rook Endgames

Against 50...f5? 51.S e l f and 52.§e5! would The second important position arises after
win. this rather long analysis:
54...§h3 55.Sa2t £ d l 56.&f5 Sh4 57.&xf6
5O...h4t? 51.&g4 f5t 52.&g5 S b 8 is another Sxf4f 58.&g5 Sf8 59.£xh5 S h 8 t 60.&g5
unsuccessful try. The simplest way for White Sg8t 61.&f6 Sh8 62.&g7 Sh3 63.&g6
to play is: 53. b5 Sg8t 54.&f6! &xf4 55.Sb4f Played with the idea of &g5 on the next
and after more moves, White will win. move, when Black’s likely retreat of the rook
will allow White to advance the h-pawn.
Black could make a draw by restricting White’s Black has to retreat straight away.
ability to improve his position. 50...&e3! 63...5h8
51.Sb3t &e4 It is hard to see how White
would make progress.

51.&B &c2
An interesting position arises after:

51...f5 52.&g3 &c2 53.S a l Sxb4 54.Sa5 h4f

55.£xh4 Sxf4t 56.&g5 Sfl

abcdefgh
64.Sa4J!
The key idea of this endgame. Black cannot
take on h2, as the king is as poorly placed as
can be. White will advance the h-pawn soon:
64...Sh3 65.&g5 &e2 66.&g4 Sh8 67.h4
White wins.

abcdefgh 55.&B &d3 56.&xfi6 <&e4 57.B

Black looks close to making the draw. If White won.
White takes on f5, the black king will easily
join the defence after the tempo-winning 57...h3 58.Sa4t 59.&g5! Sb5 6O.Sa2
check on g l . Sb4 61.16 Sg4t 62.&B Sf4f 63.&e6 Se4f
57.Sa4!! 64.&d6 Sd4t 65.&e7 Se4t 66.&18 Sb4
White wins narrowly, but also in a beautiful 67.17 Sf4 68.Sb2 Sa4 69.Sb6
way. The f-pawn is restricted, while the 1-0
h-pawn is advanced. For example:
57...&d3 58.Sf4 S g l f 59.&xf5 &e3 6O.h4 When the opponent is pinned to the bottom
S a l 61.Se4t &f3 62.Se5! rank, it is far preferable to have a bishop’s
And so on. pawn, rather than a knight or rook pawn. Sam
Shankland has written about this in Theoretical
52.Sal Sxb4 53.Sa5 Sb3t 54.&e4 h4?! Rook Endgames, but here we shall see this
principle in a more dynamic setting.
 Chapter 6 - Promoting Passed Pawns 129

Koustav Chatterjee - Michal Krasenkow 54.&g5 hxg6

54...b4 55.Sa8f &g7 56.f6f would win the
Dhaka 2021
rook.

abcdefgh
This game is from a tournament in
Bangladesh, where White is playing strong abcdefgh
active moves, but with the intention of holding
At this moment White offered a draw.
the draw. At the same time Black is trying to
Krasenkow accepted.
win - a symptom of the rating difference more
Vl- 1/!
than the position on the board.
This is a tragedy, not only because the position
41.gg8t &h6 42.Sg5 Se4 43.Sxf5 Sb4
is winning for White, but also because no
44.&e6 Sxb3 45.g4!
difficult decisions remain. For example:
White has strong compensation for the
55...Sb8 56.Sxg6 b4 57.Hg7t &f8 58.Sh7
pawn.
and Black can only resign.

45...&g7 46.Sf7t &g8 47.Sa7 Se3t 48.&RS

Endgames are mainly about creating passed
Se8 49.6 Sf8f 50.&e6 Se8f Sb8
pawns and promoting them. In this chapter
Realism would be 51...Sf8f! with an
we have seen a lot of different angles of pawn
immediate draw.
promotion, yet we have clearly only scratched
the surface. Still, the isolated ideas given will,
52. g5 b5?
when combined, give a decent feeling for the
52...Hf8t was still a draw.
process of pawn promotion.
53. g6! Sb6t
Black would also be lost after both 53...b4
54.Hg7t! and 53...hxg6 54.&xg6!.
Chapter 7

The Umbrella
The umbrella is when you use one of the opponents pawns as a shield against checks from
behind. Although simple, it can be a bit counterintuitive at first, as we often decide not to take
the opponents pawn. But remember: the goal of the endgame is not to have more pawns, but to
promote pawns.

The first example sees White under some pressure. In principle the position is a draw, with equal
material and many ways for White to play. But most of them lose a pawn and the draw is not too
easy to hold thereafter. Especially since the pawn ending without the h6-pawn is lost.

Theo Gungl - Jakob Leon Pajeken

Magdeburg 2022

abcdefgh
84.Sa7t!?
White decided to give up the pawn immediately and to start to harass the black king.

84.<&g2 is also interesting. The idea is to transfer the king to f5, when it would be impossible
for Black to make serious progress. Black can also go after the g4-pawn, showing how marginal
the draw can be. 84...Hb3 85.Hf6 gg3t 86.&f2 Sxg4 87.&f3! Hh4 88.gf8! Hxh6 89.&g4 gg6
9O.Ha8 and Black cannot make progress against decent defence.
132 Conceptual Rook Endgames

84...&xh6 85.Sa6t &g7 86.Sa5 £(6 Xavier Vila Gazquez - Julio Granda Zuniga
87.Sa6| &e5 88.Sa5t £e4 89.Sxg5 f3
Andorra 2012

abcdefgh
abcdefgh
90.&g3?
66...E&2?
This natural move is a mistake, as the king
An understandable move, which turns out
gets cut off along the 2nd rank.
to be a narrow mistake.
White was still within the drawing margin, but
White also narrowly draws after 66...Bd6?
had to find 9O.Bf5! &e3 91.Se5t!> when after
67.&fl! &g3 68.Ba5 E d i t 69.&e2 E h l
91 92.g5 the draw should be easy.
70. h5, when the h-pawn offers counterplay.
90.. .Bglt 91.&£2 Sg2f 92.&A Ba2 93.SB
67.Sb5?
&e3!
White fails to exploit the chance given.
We now see that if the pawn on g4 was not
there, the position would be a draw. White can
67.Sg5? is also lost. After 67. ..g3 68.&fl
check the black king whenever he goes to the
&f3 69.Sf5t &g4 7O.Bg5t &h3, we have the
3rd rank and we do not end in the situation
thematic umbrella scenario.
with the queening f-pawn, a la the last example
in the previous chapter. But with the g4-pawn
67.. .6B!
there, the black king can hide from the checks
Preparing a check on the 1st rank, while
on g3, hence the name Umbrella.
aiming for the g2- and h2-squares.
94.<&gl
67.. .6g3? fails to 68.h51, when the h-pawn
94.Se5t £f4 95.Sf5t &g3! wins. (But not
needs to be stopped.
95... <&xg4?? which would lead to a basic draw.

67.. . g3? gives White a chance to defend along

94.. .galf 95.£h2 £2
the back rank with 68.& f l ! &g4 69.& g l &h3
White resigned.
70.Eb 1! and White draws, as this is a knights
0-1
pawn, not a bishops pawn.
 Chapter 7 - The Umbrella 133

68.Bb3t 67...6f5
68. h5 this time loses to 68...g3!. 67...&g3 also fails. After 68. h5 Ea6 69.&fl
&h2 7O.Hd2t! £ h l 71. f2 White draws.
68...&g2! 69.h5 g3 7O.h6 Sa6 71.h7
68.Bd3t &g2 69.Sd2f!
The white pawn is in time.

For this reason, Black should have started with:

66.. .Bh2!

abcdefgh
71...Se6t!
It is always useful to kick the opponent’s
king away.

72.&d2 Bh6 73.&e3

73.Sb7 &h2 and Black wins.
abcdefgh
67.Sa5
73...Sxh7 74.Bb2t &h3 75.Sb3 g2 67. &fl &g3! 68.& g l Ba2 also wins for
0-1 Black. After 69.Sf5 S a l t , Black wins the fight
for the pawns promotion squares: 7O.Bfl
White failed to see a very nice defence. Exfit 71.&xfl &h2!

67.Sd5I! 67...&fy.
Threatening the check on h l .

68.Ba3t &g2 69.Sa2t &h3

Exchanging the rooks does not work here, as
the white king is on the 1st rank.

7O.Ba3t g3 71.h5 ±g2 72.Ba5 £ h l

Black wins.

The umbrella is not too common a theme,

but it occurs often enough to deserve a few
examples. The following is fresh in my mind
as I write this.
abcdefgh
134 Conceptual Rook Endgames

Abdulla Gadimbayli - Wang Hao Instead, Black could have won instantly with:
Baku 2022
41...&c4!!
White will still take on b6, but now cannot
check the king from behind.

42.Sxb6 &d3
Black is simply winning. The king goes
to d2, when White cannot fight against the
advance of the e-pawn without ridiculously
passive manoeuvres.

abcdefgh
Wang Hao played the automatic:

41...&xd4?
And only won later (on move 78) because of
a blunder by his opponent.

abcdefgh
43.Sb3t &d2 44.Se3 Sg4 45.Se2f &d3
46.§e3t &xd4
Only now, when the white rook cannot give
checks from behind, the pawn is taken.

47.Sb3 f4
Black will soon have two passed pawns, and
is clearly winning.
Chapter 8
Breakthrough
The breakthrough is generally a theme from pawn endings, which at times spills over into rook
endings. It is another variation on a recurring theme: pawns are not important; rather, passed
pawns and promoting pawns are important.

In the first example, Lasker lost without a fight.

Emanuel Lasker - Grigory Levenfish

Moscow 1925

abcdefgh
49.&fi6? &b5 5O.Sal a4 51.5!?
Too late.

51...ex5 52.e6 fre6 53.&xg6 f4 54.h5 B 55.h6 e5 56.Sel a3 57.Sxe5t &c4 58.Sel a2
59.h7 Sa8 60. <&g7 £2 61.Sal &b3 62.Sfl al=W| 63.Sxal Sxal 64.h8=® Sglf
0-1

White could have drawn if he had urgently created a passed pawn.

136 Conceptual Rook Endgames

49.f5! 53...&B4 54.h6 a3 55.h7 Sa8 56.&g7 &b3

57.h8=® Sxh8 58.&xh8 a2

8
7
6
5
4
3
2
abcdefgh 1
49...exf5
abcdefgh
49...gxf5 5O.h5 makes things even easier for 59.Shl!
White. It is important to avoid the black pawns
taking squares away from the rook. Therefore,
5O.e6! fxe6f 51.&xg6 this is the only square. Luckily it is also the
White has counterplay. square most people would choose without
thinking.
51... b5 52.Sal a4
59.Sgl? loses to 59..T4 60.&g7 e5 61.&f6 f3
62.&xe5 <&b2 and there is no check on g2.
The same idea works with the e-pawn if the
rook goes to fl .

59...e5 60.&g7 e4 bl.&fiG e3 62.&x£5 e2

62...&b2 63.Sh2f is not a problem.

63.&e4 <&b2 64.&d3

White draws by a tempo.

The following example probably looks more

complicated than it is. In essence, Black is
abcdefgh converting a pawn majority to a passed pawn.
53.h5! It does not matter that White gets a passed
The point of everything is to get the passed pawn of his own, as it is a bit further from
pawn moving. promotion.

After 53.&f6? f4 54.&C5 f3 Black wins

similarly to the game above.
 Chapter 8 - Breakthrough 137

Yannick Pelletier - Vasyl Ivanchuk 45.&gl

45.& e l also does not work. 45...Sxg2
Internet 2013
46.Exc5 (46.&fl Ef2f wins slowly and
easily) 46...Eg I f 47. &e2 E b l The g-pawn is
8 queening. It is important that the b-pawn is
covering the c4-square.
7
6 45...Sb2!
5 The rook is better placed here, as it prevents
White from attacking the black b-pawn.
4
3 46. fl

2
8
1
7
abcdefgh
6
43...c5!!
The game went 43...&g7? 44.&fl and 5
White drew on move 72. 4
43. ..b4? would be a wrong move order, as 3
White can play 44.&fl Bf2t 45.& e l ! Bxg2 2
46.Bxc6, when after 46...Eglt 47.&e2 E b l he
has a miraculous saving move: 1
abcdefgh
46...c4!
The break! The goal, as always, is to create
a winning passed pawn. As the check on f2 is
waiting, it does not cost a tempo to have the
rook on b2.

46...Bxb3? is not the idea. Black needs to keep

the white rook passive. 47.Sxc5 b4 48.&e2
Eb2f 49.&e3 Exg2 5O.Ec4 b3 51.Hb4 and
abcdefgh White draws after either 51...Sb2 52.Eh4f, or
51.. .b2 52.&f4.
48.Bc4!! Black cannot win.

47.bxc4 b4 48.c5 b3 49.c6 Ec2! 5O.Ebl b2

44.&A E£2t
51.c7 B e l t 52.&e2 Exbl! 53.c8=®
The white king has to be pushed back into
the corner.

44...c4 would be rushed.

138 Conceptual Rook Endgames

Javier Gonzalez Sanchez - Christopher Chabris

Internet 2021

abcdefgh
53... S e l f !
The transition into the queen ending is
necessary - and winning. White should not abcdefgh
be allowed to play ®c8-f8 with a perpetual in
43...&e6!
sight.
43...&d6? would allow 44.Ef5! Ef2f
45.& e l a5 46.c5f!> when White holds. The
54.&xel b l = ® t 55.&d2 ®gl extra tempo is very useful.
The queen ending is winning.
44.SB S£2| 45.&el a5!
The following game is more advanced, but
Christopher was not able to find the winning
is at its core the same. A pawn is sacrificed
line and took the draw with: 45...He2f?
to create a passed pawn. It is a small price to
46.&fl Sf2f 47.' £ e l 5e2f 48.&fl
pay. Again, Black found the win only after
the game. This time with no excuses beyond
46.h5
“chess is difficult”, as the game was played with
a classical time control.

abcdefgh
 Chapter 8 - Breakthrough 139

46...a4! 49.Sxe3t &f7 5O.Sh3 Sb2

The correct timing. Not the only win, but by far the coolest.

It would be a disaster to go for the pawn 51.117 a2 52.h8=® al=S#

ending with 46...Be2f 47. &fl Sxa2? 48.Sxf3
Sf2t, as after 49.Sxf2 exf2 50.h6 &f7 51.c5 In the next example, White won without great
White wins by queening both his pawns with happenings.
check.
Lucas van Foreest - Andrey Orlov
47.h6
Germany 2022
But the pawn ending does work after
47.bxa4 Se2t! 48,<±>fl Sxa2 49.Sxf3 Sf2t’.

47.. .5xa2!
Taking directly is strongest.

47.. .5e2f 48.&fl Sxa2? looks clever, but is

anything but. The issue is that after 49.Sxf3
a3 5O.Sxe3t £f7 51.h7 Sh2 52.Se2 Black
cannot take on e2 with check, followed by
...&g7. And after 52...Sxh7 53. g2, White
will hold.

48.SxB
abcdefgh
41...&g6?
41...‘&e6? 42.& d4 5a4 43.&C5 also wins for
White.

42.&d4 &h5 43.h3 Sa3 44.Sg2 c3 45.§c2

&g6 46.&c4 h5 47.&B4 Sa8 48.a4 h4
49.Sxc3 Sb8f 50.&a3 Sa8 51.gxh4 &h5
52.&b4 Sb8t 53.&c4 Sc8f 54.&d4 Sa8
55.Sa3
1-0

Black had a chance to activate the king by

abcdefgh throwing the pawn at White first. This is not
quite a breakthrough to create a passed pawn,
48...a3
but it is still breaking open the kingside.
It is important to have a passed pawn.
Promoting it is what it is all about.
140 Conceptual Rook Endgames

41...h5!! The final example in this chapter is preceded by

two wonderful examples, yet is by far superior.
This time the breakthrough happens after a
transition into a pawn ending - but only by
tracking down and exchanging the opponent’s
rook by force, leading to a finish of study-like
beauty.

Evgeny Levin - Daniil Golikov

St. Petersburg 2021

abcdefgh
42.&d4
42.h4 Sa3t is also fine for Black.

42...h4 43.&xc4 &g6!

White cannot keep the king out. Black
holds. For example:

44.gxh4 &h5 45.SE2 &xh4 46.&b3

abcdefgh
51.&B!!
The game was drawn almost immediately,
with 51.&d2? Sf5 52.Sb7 Sf6

51...&c3
Black does not have other reasonable ideas.
Trying to activate the rook will lose all three
pawns before anything gets going on the
queenside.

52.g4 &xb3 53.g5 SB

White also arrives first after 53...Sc6 54.Sxf7
abcdefgh
<&xa4 55.Sxg7 b5 56. f5!> when his strongly
46...Bb8f! placed rook helps a lot.
White cannot easily make progress. The same
happens if White goes for the f-pawn, when
Black can keep the king on g4 and defend with
...Sa5t. Black draws.
 Chapter 8 - Breakthrough

abcdefgh
54.Be5!! Bxe5
If White is allowed to take on f5, the h-pawn
queens.

55.1xe5
There is no defence against the white break:

55...6xa4 56.h5!
Threatening h5-h6.

56.. .gxh5

abcdefgh
57.g6 lxg6 58.e6
White wins.
Chapter 9
Check the Checks!
This is the simple mantra worth remembering. In rook endings, often the pieces are perfectly
coordinated, but with a check, it is all ruined! Also, there are many times where checks can be
used to reposition the rook to a better square. I was coaching a 2600+ grandmaster and had given
him roughly 50 rook endings to solve. He had written “check!” on a piece of paper next to the
board. Still, he missed a lot of checks! Of the concepts in this book, this is probably the most
important new one.

All of this sounds terribly simple, but as with so many ideas in this book, it is powerful. A good
example of this is the first decisive game of the 2018 World Championship match, which was also
the first game of the rapid play-off.

Magnus Carlsen - Fabiano Caruana

London 2018

Lc4 e5 2.£ic3 f6 3.g3 JLb4 4.e4 0-0 5.£}ge2 c6 6.JLg2 a6 7.0-0 b5 8.d4 d6 9.a3 JLxc3
10.£ixc3 bxc4 ll.dxe5 dxe5 12.£ia4 JLe6 13.®xd8 Sxd8 14.JLe3 £ibd7 15.f3 Sab8 16.Sacl
Sb3 17.Sfel £ie8 18.JU1 £id6 19.Scdl £ib5 2O.£ic5 Bxb2 21.£ixe6 £xe6 22.JLxc4 £id4
23.JLxd4 exd4

abcdefgh
144 Conceptual Rook Endgames

The opening had gone well for the defending

champion, but from here on it started to go
awry.

24.JLxe6f?
Carlsen overlooked a direct win with:
24.Sxd4 &f7 25. hl!, when Black is denied
the trick ... e5xf3t.

24.. .6A8 25.Sxd4 &e7 26.Bxd7t

White would have liked not to enter the
rook ending, but it’s all there was on offer.
abcdefgh
26. JLg4? ? e5! would have allowed Black to The critical moment of the game, and by
solve all of his problems. extension the match. If Caruana had held this
game, he may still have lost the match, but in
26.. .Bxd7 27.txd7 &xd7 28.Sdlf &e6 these short play-off matches, losing a game
29.f4 c5 often brings about a level of desperation that
does not make us play better.

37...&xe4?
Missing the check! After 37...Sa2f! 38. h3
<&xe4 Black draws. Either 39.Sxg7 Sf2, or
39.Se7t &f3 4O.Sxg7 S a l 41.&h2 Sa2f with
a draw.

38.Se7t!
And Carlsen answers with a decisive check.
The black king is sorely missing the f3-square.

38.Sxg7? would allow Black to draw with

abcdefgh 38.. .5a2t 39. &h3 Sf2 again.
3O.Bd5?
3O.Scl ! was much stronger, but deep analysis 38...6xf5 39.Sxg7 &K 4O.Bg5 a5 41.Bxh5
suggests that Black might be able to hold with a4 42.Sa5 Sal 43.&B a3 44.Ba6f &g7
deep and accurate defensive play. 45.&g2 Sa2f 46.£h3 Sal 47.h5 &h7
48.g4 &g7 49.£h4 a2 50.&g5 51.h6
3O...Bc2 3Lh4 c4 32.5 1 33.Bc5 h5 B b l 52.Ba7t &g8 53.Bxa2 Bb5t 54.&g6
34.&fl Bc3I? Bb6f 55.£h5
34...c3 feels more natural to me, but the text 1-0
move was good enough.
Carlsen successfully defended his title, winning
35.&g2 Bxa3 36.Bxc4 &e5 37.Sc7 the play-off 3-0.
 Chapter 9 - Check the Checks! 145

Alireza Firouzja - Richard Rapport the start of the ending, Black is under slight
pressure that quickly escalates to serious
Paris 2021
problems.

Magnus Carlsen - Santosh Gujrathi Vidit

Wijk aan Zee 2022

8
7
6
5
4

abcdefgh 3
The next game is a random blitz game, but 2
illustrates the point made excellently (as do 1
plenty of other examples from other chapters).
White is in danger. His king is shouldered, and
abcdefgh
if Black is able to play ...h4, the pawns will be 28...Se8?I
impossible to stop. Better was 28... f6! with the idea of entering
the rook ending without giving White a passer,
82.Sb3t? &£2! or defending the b-pawn with ...£>d7.
Black could still have blundered with
82...&f4 83.Sb4f &f5?> when 84.Sb8 29.cxd5 Sxe4f 30.&d3 Bxa4?
narrowly saves the day. Only this is a serious mistake according to
the computer, but in reality, Vidit had already
83.Sb2f &g3 84.Sb3t &h2 85.§b2 &hl made this decision a while ago.
86.&e4 h4 87.Sb3 gl=® 88.Sb2 ®elf
89.&B ®g3t 90.&e4 ®g2f Stockfish tells us that 3O...Sf4!, with the idea
0-1 31.Sxb6 Sxa4, would have held after a long
series of unlikely moves. Instead 31.g3! is
White could have held the draw with: more dangerous. The main line goes 31...Sf6
32.&e4 £f7 33.Scl Sg6! 34.g4 Sg5! and so
82.Bftf! &g3 83.Sg8f, when the black king on. Black is walking on a tightrope.
has to hide somewhere, either on g l or h2.
If on g l , White can go for the h-pawn. And 31.d6!
83...&h2 is met with 84.§h8! with a direct Carlsen knows the value of a passed pawn.
draw. (84. e4!? also holds.)
31.Sfl? was weaker. Black can use checks to
Another game from Magnus Carlsen, showing get the rook behind the d-pawn: 31...Sa3f!
the difficulty of rook endings and the power 32. <&d4 Sa2 and the lines are longer, but Black
of the simple mantra “check the checks”! At holds.
146 Conceptual Rook Endgames

31...6f7 32.Bel The winning line was powered by checks!

Cutting the king off at a more sustainable 36.Se7t’!
time. It is important that the d-pawn is close This intermediate move makes all the
to the finish line before White spends time on difference. The black king is well placed
other things. on £7, but now has to find a square that is
inconvenient for one reason or another.
32.. .5a5 33.&d4 Sa2 34.&d5 Sd2t 36. ..<&f6
34...§xg2 loses to 35.d7 followed by Sel-e8 36. ..<&f8 loses directly to 37.Sc7 b4 38.Sc8f!
or <&xb6. and the d-pawn queens.
37.Se4!
35.&c6 b5! Stopping the b-pawn and threatening to
The only chance. Black needs to have advance the d-pawn.
something to play for. 37...Sc2t 38. £b6 Sd2
38... b4 no longer works, on account
39.Sd4!, when the pawn would queen with
check.
39. &c7 Sc2f 40.<&d8 Sd2

abcdefgh
We have arrived at the critical moment of abcdefgh
the game. Carlsen had plenty of time and did
not fully understand this, as he only spent 3 41.d7
minutes on this decision, compared to 10 on Threatening <&d8-e8.
move 40, which suggests that he was looking 41...6f7 42.Se7t!
for the win there, but at this point was suffering Not the only win, but the simplest.
a bit from tunnel vision and focused entirely 42...6f8 43.Se8t &f7 44.&c7
on pushing the pawn. White wins.

36.d7? 36.. .5c2f 37.&b6 Sd2 38.&c7 Sc2f

To understand how White could have won, 39.&d8 b4 4O.Be7t
we first need to know that 36.Se4? fails to 4O.Sflf <&g6 41.Sf8 is now too slow. After
36...Bc2f 37. b6 b4!, when the b-pawn 41...b3 42.<&e7 b2 43.Sb8 Se2f White has no
supplies counterplay. advantage.
 Chapter 9 — Check the Checks! 147

Lh4!
A typical prelude to the core idea of the
study, which is about shouldering made
possible by checks.

l...Sxh4 2dUlf 3.Sb3f!

An important intermediate move that
removes the king from f3, the best square in
what is to come.

3.Jbdi4? gxh4 4.Sb3t &g2 5.Sb2f <±>g3!

would allow Black to draw. But for now, this
abcdefgh square is not available.
40...&A8! 41.Se3 &f7 42.Se7t
3...&g2
42.Hf3t &g6! and the white king cannot
hide from the checks.

42...&f8 43.Se5 g6 44.Se6 b3 45.Sf6f

<
’g7 46.&e7 He2f 47.Se6 Exe6f 48.&xe6
b2 49.d8=® b l = ® 50>c7t &g8 51.®d8f
&g 7
Vl-Vl

Next up we have a brilliant study from the

greatest Danish composer, who at the time of
writing has just won the World Championship
in study composition and will thus be the
World Champion in study composition from
2022 to 2025.
4.Sb2f!!
Steffen Nielsen Getting the rook to the 2nd rank, where it
can temporarily block the pawn.
Original 2020
This draw is a common theoretical position:
4.jbdi4? gxh4 5. e4 (5.Sb2f g3 and again
Black is in time.) 5...h3 6. f4 h2 7.Sb2f <&h3!

4...&B 5.Jbdi4 gxh4 6.&e5 h3 7.&B &g3

8.Bb3t &g2 9.&g4
White wins.

Checks are a useful defensive resource in many

positions. This is one of the reasons why both
queen and rook endings have higher drawing
percentages than other endings.

abcdefgh
148 Conceptual Rook Endgames

Praggnanandhaa Rameshbabu - Anish Giri White could have made the draw by checking
from the side.
Internet 2022
61.Ba5! e3 62.Ba6f &e5

abcdefgh
In the following online rapid game, White abcdefgh
had a surprising drawing resource, but instead
The king cannot hide on the g-file, as White
played:
will be able to get the rook behind the e-pawn
and snatch it up. 63.Sa5t &e4 64.Ba4f &d3
61.Se8? &f5 62.Sf8t &e5 63.Sa8 Sf3f!
65.Sa3t &d2 66.Ba2f Eventually Black will
Dislocating the white king from g3, where
have to do something to avoid the checks, but
it controls both f2 and f4, important squares
after 66.. J&cl, White draws with 67.Se2 Be4
for Black.
68.&B and 69.Sxe3.

64.&g4 Sd3?
I do not want to give the wrong impression.
64...Sf7! was winning.
The old saying goes: “Patzer sees a check,
patzer plays a check.” The idea is that we need
65.&xg5?
to check the checks, not replace thinking with
A sloppy mistake, which is what happens in
checking. Take the following example as a
rapid games.
chance to get scared straight.

White had a surprising draw in 65.Sg8! < ?d4

66.Sxg5 S d l 67.Sa5! and the checks from the
side will hold.

65...Sg3t 66.&h4 Sg7

White could have resigned here, but this also
sometimes happens later than needed in rapid.
...0-1
 Chapter 9 - Check the Checks! 149

Lukas Vlasak - Tigran Petrosyan Aagaard&Yoo

Czech Republic 202 1 Original 2019

abcdefgh abcdefgh
59.Sh7t? First of all, let’s write off any ideas of building
The idea of the strong checks is to force the a fortress:
king to go to a worse square. In this case the
black king is chased in exactly the direction he l.Sb6f? &a5!
wants to go. Winning a new queen is the way to go.

59.Eh6! cutting off the black king was the The hope for White is to force the black king
drawing move. 59...Exb4 The critical move. If further away, so he can draw with: l... a3?
Black does not take on b4, the rook can block 2.Ea6f & b 3 3.d5 E b l 4.Exa2 &xa2 5.&d4!
everything from d6. 6O.Exc6 Ee4 61.Exa6 <&b3 6.<&c5 with typical shouldering tactics.
Exe5 62.Eb6 The draw is near.
2.Exb2 al=W 3.Ec2 Welf 4.&d3 Wdlf
59...&e6 6O.Sh6f &xe5 61.Sxc6 Exb4 5.<&c3 <±>b5
62.Sxa6Sa4 The black king comes and the d4-pawn
The black pawns are strong, but the decisive eventually goes. It is fiddly to win these things,
factor is the king. but winning it is.

63.&B d4 64.Sa8 d3 65.a6 d4 66.a7 When giving this exercise to students, almost
Ea2 everyone comes up with the same move, which
0-1 is what makes the position interesting.

The following position arose first as a little I.£e4?

exercise I gave to (now) GM Christopher Yoo, What looks more natural than this? Yet it is
which became far more interesting from his a losing move.
suggestions. In this one, we shall see checks
used as a way to transfer the rook across the
board in nuanced and surprising ways.
150 Conceptual Rook Endgames

3.&e5 Sh5t 4.&e6 Sa5

Black wins.

For this reason, through elimination and

despair, we can find the only move:

l.&f4!! &c4!
The most testing move, suggested by
Christopher.

1... b3 2. e5 is an elementary draw.

The point to putting the king on f4 is that after

abcdefgh 1...5h2 2.Sa8 Sh4f, the black rook is hanging
l...Sh2!! after 3. g3, and the white king is too close for
The refutation. Black to give any checks.

1...5e2f? looks strong, but White can play

8
2.<&f5!Sh2 3.d5! and he manages to escape.
7
2.Sa8 6
White has no moves.
5
2.d5 loses to 2... b5! followed by ...Sh4f and 4
...Sa4.
3
2.&f4 <&b3 is also lost. 3.d5 loses to 3...Sh4f; 2
or if 3. e3 S h i there is no shouldering.
1
2...5h4f! abcdefgh
2.<&e3!!
Again the only move is putting the king on
the less anticipated square.

2.<&e4? would lose to: 2...Se2f! 3. f3 Sh2!

and there is nothing White can do. Ironically
4.<&e4 is met with 4...<&b4! returning to
the l . e4? annotations above by a bizarre
transposition.

2...Sh2 3.d5!
3.Sa8 <±>b3! 4.Sb8f &c2 5.Sa8
5.Sc8f & b l 6.Sa8 a l = ® 7-Sxalf xal wins
abcdefgh for Black through traditional shouldering.
Forcing the king to a worse rank. White cannot play 8.d5, as 8...Sh4! cuts ofF
 Chapter 9 - Check the Checks! 151

the king at the 4th rank. So White has to 3.. .<&b5 4.Sa8 Sh3t
try 8.<&e4 <&b2 9. d5!?> but it is narrowly
too slow: 9...&c3 10.<&c5 Sd2 11.d5 d3!
8
And the king appears on the other side of the
pawn, winning. 7
6
5
4
3
2
1
abcdefgh
abcdef gh 5.&d4!
5. d2 loses to 5...Sh 1!, a well-known trick.
It looks as if White is making the draw, but
Black has yet one more trump up his sleeve.
5...5h4f 6.&c3! Sa4 7.Sxa4 &xa4 8.&b2
5...5h3t’’
White draws.
The right moment to give the check, so that
the rook can go to b3, shielding the king
Played in the early stages of the 2023 Polish
when it goes to b2.
Knockout Championship and thus a rapid
6.&e4 <&b2 7.Sxa2f &xa2 8.&d5’?
play-off game, I asked people on Twitter to
The only try, but still not enough.
guess which pawn White should advance. The
8...<&b3 9.&c5 &a4!
wisdom of crowds (often highly unreliable)
was strongly in favour of the move chosen by
Licznerski.

Lukasz Licznerski - Grzegorz Gajewski

abcdefgh
I could perhaps give the check on c3 and
return to d3 with the rook, but in reality it is
a transposition. The king comes around the
white king, breaking the shouldering by a
single tempo.
10.d5 &a5 l l . £ c 6 &a6 12.d6 Sc3t
Black wins.
abcdefgh
152 Conceptual Rook Endgames

56.g5? &d5? 57.h4 c3 58.&f4 Sflf 59.&e3 S h i 6O.Sg2!

Black missed the chance to draw the game Sh3t
with: 6O...Sxh4 61.g6 Sh8 62.g7 Sg8 63. d3
56...B3! 57.cxb3 cxb3 58.axb3 axb3 also wins for White.

61.&e2 Sxh4 62. g6 Se4f 63.&£3 Se8 64.g7

Sg8 65.&e3 &c4 66.Sg4f <&c5 67.&d3
Sd8f 68.<&e4 Sg8 69.<&f5 b3 7O.cxb3 axb3
71.axb3 <&d5
1-0

When we know the defensive idea, it is easy to

understand why the h-pawn needs to be the
spear of the campaign.

abcdefgh
56.h4! b3
59.Sb2I? Black has no other ideas, as we saw in the
The trickier try. game.
59.h4 Sd3t! 6O.£g4 Sd4f 6 1 . £ h 5 Sb4
62.Sb2 d7 63. g6 Sb7!! holds in the same 57. cxb3 cxb3 58.axb3 axb3 59.h5!
way, although Black has a slightly easier ride.
59...Sd3t 60.<&g4 <±>d7! 61. h4 Sd4f 62.<±>F5!
Sb4! 63.&g6
63. h5 e8! draws easily.

abcdefgh
abcdefgh White wins. Black does not have the ability
to quickly regroup the rook to b4.
6 3 . . . W I ! 64.<±>F7 &d6t 65.&f6 &d7 66.g6
&e8
The following is the end of a bad game. My
The king comes in to block the pawns in
opponent told me that I “was lucky to find
time, with checks from the side keeping the
counterplay every time I had blundered.” At
white king away.
this moment, I have recovered from a few
blunders to regain the advantage. However,
the conversion was anything but convincing.
 Chapter 9 — Check the Checks! 153

Jacob Aagaard - Tim Jaksland

8
Odense 2009
7
6
5
4
3
2
1
abcdefgh
66.Sf7t!
Not a difficult move, but overthinking was
abcdefgh possible.
64.&b3? Sa5 65.Sa4 Sb5t??
Back in 2009 I believed that 66. a7? was a
Black, short on time, falsely believes.
winning line. The idea is 66...Sxa7 67.Sf7t
<&b8 68. (&b5! and White wins. But Black is
65...&c7 was a simple draw.
not forced to take. 66. ..<&b7!!67.Sf7t a8
68.Sxg7 Sg5! It is difficult for White to make
66.&a2 Sf5 67.a7 SfB 68.a8=W Sxa8
progress, but Black will still have to find one
69.Sxa8 <&e6 7O.Sf8
accurate move to make the draw. 69. c4 S c 5 t
1-0
70.&d4 Sg5 71.&e4 b5 72.&f4
It was only when I got home and analysed the
game that I realised that I had a nice winning
line.

64.Sf4!! Sa5 65. W !

A wonderful tempo-gain.

65...&c7
65...gxf6!? 66. g7 Sxa6 67.g8=® Sa5 would
be a decent practical chance, as breaking down
that fortress is not at all easy. Although if you abcdefgh
have read Theoretical Rook Endgames, you
72...Sg2!! This is the key move. Black makes
will be able to cause your opponent a lot of
a draw on account of: 73. &f5 b4 74.Se7 b3
difficulties in the Queen vs Rook ending.
75.Se8f xa7 76.Se3 b2, when the pawn
is defended. See more in Chapter 20 on
Anticipation (page 273).
154 Conceptual Rook Endgames

66.. .6b8 67.Hxg7 Black could have ruined the white drawing
Threatening 68.Sf7. setup with:

67.. .5g5 68.Sb7t 57...Sa4f!! 58.&xg5 Sal!, when the white

White wins. king is misplaced. The check on g l will force
it to the h-file and keep it cut off. Black is
The next example looks deceptively easy for winning.
Black, but is anything but.
In the next example, the usage of the check is
Adrian Gschnitzer - Bence Korpa advanced and surprising, taking a big page out
of the chapter on anticipation also.
Biel 2021
Javokhir Sindarov - Jakhongir Vakhidov

Uzbekistan 202 1

abcdefgh
Black gave his opponent a chance to save the
game after: abcdefgh
57...Se8
57...Sal? 58.&B? 57...Se3t 58. <&b4 is lost for Black. The
White could have held after both 58.Sxg5f!? white king will zigzag up the board, via
<&e4 59.<±>h31! and 60.&g2, or 58.Sd8!?
a5-b6-a7.
<&e4 59.&xg5 e5 6O.Se8! and White has a
theoretical draw. 58.a5
Otherwise Black will play ...Sb8xb7.
58...Sflf 59.<&e3 SB 6O.Sg6 <&d6 61.<&e4
<&e7 62.<&e3 <±>£7 63.Sh6 64.Sh8 58.. .fl=W?
&g7 65.Sh5 &g6 66.Sh8 g4 67.&e4 &g5 Call it desperation. O r hope. Whatever you
68.<&e5 Sfl 69.Sg8f <&h4 70.<&xe6 g3 want; it does not work. Connected passed
71.&e5 &h3 72.Sh8f <&g2 73.Sg8 <&B pawns, this far up the board, will win.
74.2ffit £e2 75.Sg8 SB 76.<&e4 &fl
0-1 58...<&d6? looks close, but after 59.a6 <&d7
6O.Sfl! Black is lost, just like in the game.
(60.a7? is a curious mistake, allowing Black to
draw. 6O...Se3t! ft does not take long to work
 Chapter 9 - Check the Checks! 155

out that if the white king goes to the c-file or 59...Se2? makes no sense. 60. a6 and White
the 1st rank, Black will play ...Sei, securing a wins.
drawn queen ending.)
60.&b5 S e i 61.Sxel!?
59.Sxfl c6 60. a6 b6 61.Sft>t a7 61.Sc5t e6 would put White in a little
62.Sc6 Sb8 63.Sc8 h5 64.<&b4 h4 65.<&b5 danger. He would have to play accurately to
h3 66.Sxb8 &xb8 67.&b6 h2 68.a7# draw by stalemating himself with: 62.Sc6f!
1-0 <&d7 63.Sf6! fl=®t 64.Sxfl Sxfl 65.&b6
S b I t 66. a7 &c7 67. a6 h5 68.b8=®t Sxb8
Black could have held the draw by giving the Stalemate.
check now:

58...Se3f!!

abcdefgh
We have arrived at a trap I placed for the
students of Killer Chess Training. I did a poll
abcdefgh and fourteen of them wanted to promote to
The reason this works is that the white king a rook, while the remaining three wanted a
can no longer hide on a5. It can go to b5, but queen. All three admitted to not having seen
then queening the pawn on fl would come White’s response to this superior promotion,
with check, changing the dynamics of this while the fourteen others fell hard for my trick.
highly volatile 4th phase ending.
61...fxel=S? looks clever, but after 62. b6!
59.&b4 the pawns win! 62...Sb I f 63. c7 S e l f
There is no other path. 64.&d7 S b l 65.a6 h5 66.<&c7 S e l f 67.<£d8
S b l 68. a7 and Black can only resign. A truly
Black is also fine after 59. b2 Se2f 60.<&c3 evil trap, accompanied by a truly evil laugh
Sei!, and 59. a4 Se4f 60. <&b5 Sei!, which from yours truly. We put this small part of the
transposes to our main line. lesson on YouTube and people have confided
to me that they all fell for it, hard and fast.
59...Se4f
With the idea that the king cannot avoid the 61.. .fxel=®! 62.b8=®! ® b l f 63.&a6®xb8
checks without going to a l , b l , c2, c3 or b5. Stalemate.
All of which can be answered with ...Sei!.
156 Conceptual Rook Endgames

The following example of the effectiveness of 56.<&c5?

checks is simple, yet brutal. It was played in a If you combine the knowledge of the
closed GM tournament and supplied one of chapters on Checking Distance (page 163) and
the IM norms to one of the young members of Defence from the Front (page 171), you will
Killer Chess Training. see that White could have made a draw with:
56. S h i ! <&f6 (It relies on tactics. After 56...g5
Suat Atalik - Shahil Dey 57.Sh7t! the pawn ending is a draw.) 57-Sflf
Belgrade 202 1 &e5 58.Sgl £f4 59.Sfl t &g3 6O.Sglf &f2
61.Sg6 &f3 62.Sgl! and Black cannot make
progress.

56...&ft> 57.S£2f <&e5 58.<&c6 Sd6f 59.&c5

Sd8 6O.Sg2 Sc8f 61.&b6 Sg8 62.<&c7 g5
63.&d7 Sg7t 64.<&e8 <&f4 65.£f8 Sa7
0-1

Black could have won a vital tempo with:

48...Sc8f!!

abcdefgh
48...h5? 49.b5 h4 5O.b6 h3
White draws without effort after: 50...g5
5 E b 7 S b 8 52.Sb6 <&g7 53.&d3

51.Sa3
5 E b 7 S b 8 52.Sa3 Sxb7 53.Exh3 Ed7
would transpose to the game, more or less.

51...Sh8 52.b7 h2 5 3 . S a l ! Sb8 54.Shl

Sxb7 55.Sxh2 Sd7!
abcdefgh
49.&d5 Sb8 50.&c5, which would be useful
before pushing the pawns. 50. ..h5 51.b5 h4
52.b6 g5 53.&c6 h3 Black wins.

We enter the final example of this chapter just

after the time control; the moment where the
white advantage is at its peak and the position
can be won with the most natural moves, with
just a little sprinkle of imagination. But as so
often happens when we have regained the right
to think, nothing seems as simple as it does for

abcdefgh
 Chapter 9 - Check the Checks! 157

the guy with the computer and Bu managed

to make things much harder for himself right
away.

Bu Xiangzhi - Vasyl Ivanchuk

Sofia 2008

abcdefgh
The critical moment where White has to
find something extra. There are not one, but
two ways to bring it home.
45.Sa7t
The alternative is 45.Sb6 5xf2 46.Sb5!!. The
rook transfers to f5 and the passed pawns
will almost promote by themselves. Rook
behind the passed pawn, dude.
abcdefgh 45... g6 46.Sc7!
A clever manoeuvre.
41.g 5? 46.. .5xf2 47.Scl
Overthinking. It is clear that there was The rook comes to the 1st rank to defend
something in the chain of natural moves against the advance of the f-pawn and to
that Bu did not like. There is no purpose in support the white passed pawns. In both lines
speculating what. What we can see is that extra details can be added, but the outcome
the winning line did require finding a strong should be clear enough.
continuation a bit down the line, just outside
the natural horizon of a human. The same can 41...&g6
be said of the game, but apparently Bu felt it White’s idea was to play 41...fxg5 42.&f3
was a better chance. <&e7 43.Sb6, winning.

The winning continuation goes like this: 42.gxfi6 gxfifi 43. 6 g5

41.£f3 g5 42.<&e4 Se8f 43.<£d3! f3!
Black has to try something active. Passive
play will see White convert his extra pawn
on the queenside.
44.Sxa6 Se2
158 Conceptual Rook Endgames

5O.gg8t’!
This check is the only winning idea. But it
is there!
5O.Sh8 Sa5! and Black holds.
50...<&f5 51.Sh8
White wins; first the h2-pawn, then the
game, as the b5-pawn is now indirectly
protected through the X-ray check on h5.

White had another win after natural play,

again with a sprinkle of brilliance.
44.&e4 Se8t 45.&d3 f3 46.b5!I
abcdefgh Creating a passed pawn urgently, as we have
covered in Chapter 5.
It turns out that leaving Black with isolated
46.Sxa6? fails to 46...Se2 47.Sa8 Exf2,
doubled pawns was not an improvement for
planning ...&f4 to support the pawn.
White. The counterplay from the active king
48. <&e3 Sb2 Black holds.
is annoying and the win is harder to visualise
now than it would have been, had White 46...axb5 47.Scl! £f4 48.a6 Sa8 49.Sal f5
50.<&d4
played natural moves all along.
White wins. Black will not be able to give up
the rook and take the f2-pawn quickly enough
44.Sc5t?
to make it.
44.Sb6 f5 transposes to the position before
White’s 47th move. As Bu did not find the win
there, we can imagine that this in itself was not 44...B 45.Sc6
appealing to him. 45. b5 axb5 46.Sxb5 Sa6 also makes it hard
for White to make progress. The white king
Black could also meet 44.Sb6 with:
can attack the a6-square as much as the black
44...5c8!?
king can attack the f2-pawn. But no more.
Although this narrowly loses too. The
winning line goes:
45.Sxa6 Sc3t 46.<£e2! 8
Keeping the king close to the action.
7
46. . .5xh3 47.Sa8 Sa3 48.b5 h3 49.a6 h2
6
5
4
3
2
1
abcdefgh
Black has been given a chance to take
abcdefgh the initiative and orchestrate counterplay.
 Chapter 9 - Check the Checks! 159

When we get this type of chance, we have to the second phase starts. In the second phase
grab it with both hands, as it is not likely to we see a race of some sort. Both players are
come again. executing their ideas and plans. When we are
the side in control and we have the chance to
45...Sb8! 46.Sb6 Sa8? improve our position more than the opponent
An awful passive move, which deserved can his, we should do so. This is a slightly
to lose on the spot (plus 35 moves for the abstract version of this principle.
execution of the winning idea).
The winning move was:
It seems that Ivanchuk has not read Chapter 1 . 47.&g2!!
If anything in this book is true, the logical play The white king cannot be checked here,
for Black would be to give up the a-pawn and allowing White to play f2-f3 before going
get the rook in behind the white queenside for the race.
pawns, slowing them down, from where it 47...5a7 48.f3 &h5 49.&f2 &g5 5O.Se6
can attack the white kingside pawns too. Threatening to bring the king out, covered
Well, would I lie to you? Black draws after: by the rook. Black has to enter the race.
46.. .5c8! 47.Sxa6 Sc3t 4 8 . £ e 2 f3t 49.£d2 50...5b7 51.Sxa6 Sxb4 52.Sa8 &f6 53. a6
Sb3 5O.Sb6 Sb2t 51.&e3 Sb3t 52.£d4 Sb2 Sb2f 54.<±>el Sa2 55.a7!
53.Sc6 The last trick. 53...Sxf2 (53...Sxb4f?
loses to 54.Sc4 and 55.Sa4.) 54.Scl Sa2
55.&e3 (55.&c5? even loses to 55... f4.)
55.. T2 A draw can be agreed soon.

abcdefgh
White wins. The winning plan is simple,
although it will take a lot of moves to execute.
Using zugzwang, White will take all the black
abcdefgh pawns with the king and advance the f-pawn.
47.£e2? Arising on f6 or f7, it will force the black king
The most natural move, but this time around to either go to the 6th rank, when the rook
this was the wrong choice. Black is so tied down will come out of hiding with a check, or to f7,
that White had a chance to fix the structure when Sa8-h8-h7t wins.
before bringing out the king. This is a common
theme in the endgame. Often it has two phases. 47...§e8t 48.&d3 Se4!?
In the first, the players are improving their Quite playable, but it was more natural to
position, often with one side in charge of when seek counterplay with:
160 Conceptual Rook Endgames

48...f3! 49.Exa6 Be2 5O.Bc6 Bxf2 51.E c l ,

when Black has 51...Ea2 when the draw
should not be so easy to ruin.

49.±c3 Ee2?!
This is good enough to draw, but it is a tempo
slower than the alternative and is therefore part
of the story to how Black came to be lost a few
moves later.

49...f3! is the move Black wants to play, but it

has to be backed up with tactics:
abcdefgh
Black has fought hard and is within a
whisker of saving the game. He just has to find
one more brilliant idea.

52.. .±f4?
Logical - and losing.

52.. .f4? also fails, this time to brute force:

53. b5 Ea2 54. b6 Bxa5 55.b7 Eb5 56.Bd5t!
abcdefgh
52...5.2? obviously fails to 53. &b3!, but this
5O.Bxa6 Be3t!! The key idea. 51.&c4 Be4f!
does not mean that the rook does not belong
The simplest. (51...Be2 draws narrowly, but
on a2!
could go wrong.) 52.&c5 E e 5 t Keep checking
till the king is far away. 53. d6 Only now it
52...Be2! (52...Bh2 with the same idea, as
is time to go for the f2-pawn. 53...Ee2 54.Bc6
well as some differences, also works) with the
Bxf2 5 5.E c l Bb2 Black is by no means behind
intention of meeting 53. b5 with 53...Ea2!!
in the race and in many lines it is White who
would have held. There are many complex lines
will have to take care to secure the draw.
that can be created from here, but essentially
Black is no worse already. White has other
5O.Exa6 Ex£2 51.Ed6
ideas, but none better.
When in doubt, push the passed pawn.

53.&b3?
51...Efl? looks logical, but remember it is
53. b5? and 53. a6? both fail to 53...Ba2,
usually better to push the pawn and keep
when the black rook is behind the pawns and
flexibility for the rook. 52. c2!! f3 53.Edl! is
the counterplay on the kingside is coming on
the refutation.
quickly.

52.Edl
The winning idea is a stunning two-mover.
53.Bal!Be2
 Chapter 9 — Check the Checks! 161

55...£g3
Black has to spend a tempo on this - a tempo
he will sorely miss.
55...5a2 loses to 56.b6! Sxa5 57.b7 S b 5
58.Sxf2t and 59.Sb2.
56.b6 S e i
56...6g2 loses to 57.Sxf2f followed by
58.a7.
57.Sxf2 d?xf2 58. b7 S b l 59.a6 f4 60.a7 Sxb7
61.a8=® Sg7
abcdefgh
54.B5!!
The surprising move. The idea of pushing
the b-pawn is not too difficult, as it would
promote with check. It is the combination
with putting the rook behind the a-pawn,
which is surprising. It is not too hard to
understand. Black should not be allowed
to play ...Sa2xa5-b5, which is the drawing
idea. So, he is prevented from playing ...Sa2
for just a moment.
abcdefgh
54.a6 f2 55.a7 S e i 56.a8=W fl=® does
not give White an advantage and is the core There is still some play left. Given a tempo
problem of most of the variations looking more, Black would make the draw. So, let’s
for one. take this one till the end.
54..T2 62.®e4 Sg3t 63.&d2 f3 64.®xh4 &g2
The only move that makes sense. 65.®e4 Sxh3 66.&e3 Sg3 67.&f4 Sh3
68.®g6t &h2 69.&e3 Sg3 70.®c2t &h3
71.&f4
White wins.

53...Se2 54.a6 Se3t 55.&c4

This effectively forces a draw.

55.&a4 f2 56.Sfl ends in a draw too, after

some complications: 56. ..&g3 57.Sxf2 &xf2
58. b5 S e i 59. b6 f4 6O.b7 f3 61.a7 <£>gl
62.b8=® S a l t 63.&b3 Sb I t 64.&c2 Sxb8
abcdefgh 65.axb8=W f2 With a well-known fortress.
55.Sfl!
It turns out that advancing the f-pawn is 55...£2 56.SA ±g3 57.b5 Sa3 58.±b4 Sa2
not without difficulties. The pawn is now 59.±b3 Sa5 6O.±b4 Sa2 6 1 . ± b 3 Sa5
hanging. Vi-Vi
162 Conceptual Rook Endgames

Check the checks! The advice is as simple as that. In endgames in general, checks are a useful
way to harass the opponent, but especially in rook (and queen) endings, they are effective and
can be an extra weapon. We tend to give checks without mentally absorbing their dominating
feature. Truly grasping the power of checks can provide the extra 10% that will elevate your rook
endgame play to the next level.
Chapter 10

Checking Distance

When talking about checks, one subject cannot be avoided. Checking distance. The general idea
is this: the greater the checking distance, the better for the side giving the checks. So, if you do
not know what to do, increasing the checking distance is almost always a good idea.

The first example is from the 2020 US Championship, which was played online in a rapid format,
which is probably why the game was unnecessarily long.

Hikaru Nakamura - Aleksandr Lenderman

US Championship 2020

abcdefgh
51...gg7t?
An understandable mistake, but also an instructive one.

51...Sh8!
This was the only drawing move. You can explain this with variations - and I shall - but don’t
lose sight of the reason why this is working: It extends the checking distance.
52.<±>f5 Sf8t 53.&g6 <±>d5I!
Also an important move. The key idea is to avoid the Se6f followed by Sg6 idea.
164 Conceptual Rook Endgames

54, £g7
54.Sa2 would allow the black king to come
closer. After 54...&e5! Black simply draws.
54...Sf4!
Preventing the g-pawn from advancing.
55.<±>g6

8
7
6
abcdefgh
5
4 59.&f5! Sf8t 60.&e6
Black again has to choose between allowing
3
the g-pawn to advance, or being checked
2 away. Since allowing the g-pawn to advance
1 loses without a fight, it is not much of a
abcdefgh choice.
60...5g8 61.Sc2f &b6 62.<±>f6!?
55...Sf8!
A tablebase move which is interesting to
And Black holds.
Notice that 55...Sg4f? would lose to 56. &f5 understand. Black’s rook is stuck and White
puts Black in a sort of mini-zugzwang, where
Sg8 57.Sd2f.
he has to make an uncomfortable move with
52.±f5 Sg3?I the king.
Not the strongest defence, even if it held 62...6b7
the game. White would have to show great 62...&b5 is inconvenient, as after 63. &f5
technique and the game would potentially Sf8t 64.£g6 Sg8t 65.&h7 Sg3 66.&h6
have found its way into the “Slow Play” Sg8 67.Sf2, the check on f5 with the idea
chapter, had Black given another check: Sg5 is useful.
63.&f5 Sf8t 64.&g5 Sg8f 65.&f4 Sf8t
52...5f7t 66.&e3 Sg8 67.&f2 Sf8f 68.&gl Sg8 69.Sc4
This still loses, but is tougher to grind down. Sa8!?
53.&g5! This fails to 70.&£2. But White should avoid
White can zigzag between threats of 70.g4?, when 7O...Sa2! would hold.
advancing the g-pawn and checking the
black king away to the far queenside. The
technique is important to understand, so
you can use it in your own games.
53.. .5f8 54.Sa2! Sg8f
54...&e6 55.Sa6f &e7 56.Sa7t &e6
57.Sg7! is the key idea behind 52.Sa2!.
55.&f6 &d7
A waiting move is necessary.
After 55...Sf8t 56.&g7 Sf4 57.&g6 Black
cannot prevent the g-pawn from advancing.
56.<±>f7 Sg5 57.Sd2f &c6 58.&f6 Sg8

abcdefgh
 Chapter 1 0 — Checking Distance 165

53.&f4? The next example led to a loss by one of

Missing the easy one! 53.Se6f! &d7 54.Sg6 Denmark’s promising younger players.
Sa3 55.&f6! and the pawn will advance.
Martin Percivaldi - Oleg Korneev
53...Sg8 54.g4 Sffif
Koge 2015
Black has achieved a theoretically drawn
position. Nakamura was hungry and tried
for another 90 moves to win. Let’s be kind
and say that the last 6 5 of them were totally
unnecessary...

The next example is covered at great depth in

Theoretical Rook Endgames by Sam Shankland.
We had both individually selected the game
to show important points. As mine is simple,
and different to his, we will have this one
overlapping example.

Andrey Stukopin - Lazaro Bruzon Batista

abcdefgh
San Francisco 2019 Despite being a rook up, White is obviously
in trouble. Both the black pawns are close to
promotion. Looking at lines such as 72.Sc8?
h2, when the h-pawn queens, or 72.Sa6
&c2 73.Sc6t &d3 74.Sd6f &c3 75.Sc6f
&d4 76.Sd6t &c4! (76...&c5? 77.Sdl and
White would hold by a tempo) 77.Sdl &b3
Percivaldi perceivably got desperate. Here
Black makes it in time to the 3rd rank, before
White makes it to h l with the rook. Unable to
take on h3 with check, White now loses.

For this reason, he decided to take his chances

with rook vs queen, asking his opponent to
abcdefgh show his technique. Sadly, for him, Korneev
Black played the horrible: has plenty of technique.

54...S£S? 72.SE6? &c2 73.Sxh3 b l = B 74.Sh6 &d3

54...SH! was the winning move. White is no 75.Se6 d476.d?e7®f5 77.2fi6®g5 78.&f7
longer able to get enough checking distance, &e5 79.Hg6 ®h5 80.&g7 &f5 &g5
when checking from the front (see Chapter 1 1 82.Sa6 ®e8 83.Sa5t &f4 84.Sa6 Bd8
for more on this topic). Black simply wins. 85.Sfi6t &g5 86.SF7 Wd4t 87.&h7 ®e5
88.Sg7t 89.&g8 ®h5 9O.Sgl @d5t
55.Sel 56.Sdlf 91.&117 ®d3t 92.&h8 @d8t 93.Sg8 @c7
White made the draw by move 8 1 . 0-1
166 Conceptual Rook Endgames

This is a perfect example of paying attention Matthew Perrine - Jude Shearsby

to the checking distance. The tempo White
Internet (rapid) 2023
is lacking in the critical variation, could have
been won by increasing the checking distance.
8
72.Sc7!! 7
6
8
5
7
4
6
3
5
2
4
1
3
abcdefgh
2
59.Sd5?
1 White does not have the time to waste.
abcdefgh
He could have made a draw with:
72...±a2
59.a5! f2 6O.Sg6!!
72...h2 leads to a draw after 73.Sh7 al
A nice combination of checking distance and
74.Sa7t with a repetition.
anticipation. White is ready to defend the
a-pawn on a6, which means it would be close
73.Sa7t &b3 74.Sb7t &c3 75.Sc7t &d4
enough to the finish line to secure the draw.
76.Sd7t &c5 77.Sdl &c4 78.Shl 6O.a6? Sd8f 61.&C1 &e2 62.Se5t <&f3
White makes the draw.
63.Sf5t e3 Black is threatening ...Sa8
and the elimination of the a-pawn. At the
The next game was played in the club
moment the white rook is too close to the
championship at Killer Chess Training, in
f-pawn to give checks. 64.Sf6 (If 64.Sf7,
the third heat, which was absolutely brutal.
Black has 64...Sd7! 65.Sf8 Sa7! and the
Michael Adams was third with 4/6...
a-pawn is lost. White cannot give checks and
return the rook to the 6th rank, as with the
king on f l , Black would have ...Sf7! and the
f-pawn cannot be controlled.) 64...Sd6! Now
64...Sa8? does not work, as White would
give checks and answer ...&fl with Sf6! to
keep the f-pawn under control. 65.Sxd6
fl =WT 66. &b2 Wf8!! Black wins. The threat
is ...Wh8t and ...&d2-cl, with checkmate.
If there was no pawn on b3, White would
be able to create a fortress with Sd3-b3, but
without the b3-square, he would end up
 Chapter 10 - Checking Distance 167

in zugzwang. 67.Sd3t e2 68.Sd7 ®f6f Before we move to the final advanced example
69. & b l ®xa6 This fortress is by no means in the chapter, let’s take another clear-cut one.
easy to break down, but a key reason why it
fails is that 7O.Sd3? loses to 70...®xd3!. Quinten Ducarmon - Emil Powierski
60.. .5.8t 61.&C1 &e2
Germany 2015
61...Sf8 62.& d l would not allow the black
king to exit its prison, with the rook on f8.
62.Se6f
Black cannot make progress.

59.. .f2 6O.a5 ±g2 61.Sg5t 62.Sf5t

±g3 63.c3!
Matthew plays his chance.

63.. .bxc3 64.&c2

abcdefgh
White lost after:

54.Sd3?
54.h4? &e5 and ...f4f is also completely
winning.

54...h4f 55.&xh4 Bxg2 56.h3 Bh2 57.&g3

The pawn ending after 57.Sa3 Sxh3t
abcdefgh 58.Sxh3 gxh3 59.&xh3 &e5 loses elementarily.

64...Se5?
57.. .5.h3t
Missing 64...2e2f! with the idea 65. xc3
Black won on move 69. A key point is
Se3t and 66...Sf3, winning. that 58.&f4 does not lead to stalemate after
58... e6.
65.Sx£2
By far the simplest, but 65.Sxe5!? fl=®
White draws marginally by extending the
66.Se3t &f4 67.Sxc3 would also hold. checking distance to the max.

65...&x£2 66. b4 e3 67. xc3 e4 68. c4

54.Sb4!!
Se8
54.Sc4? is not far enough. After 54...h4f
0-1
55.&xh4 Sxg2 56. h3 Sh2, the king is too
close to the rook: 57. Sc6f e5 58.Sc5t e4
Matthew lost on time, in a clearly drawn
59.Sc4f &d3 and Black wins.
position. 69. a6 Sc8f 70. b5 &d5 71.a7 and
the draw is near.
168 Conceptual Rook Endgames

54...h4f
Still the critical try.

White also survives after 54...&g5 55.h4f

and 54...&e5 55.Sb5t <&e4 56.Sb4f &e3
57.Sb3t &e2 58.Sb5!.

55.±xh4 Sxg2 56.Sb6f ±e5 57.Sb5t &e4

58.Sb4f ±d3
The black king has been kicked away from
the black pawns, meaning that White can
now draw with both 59.h3 Sh2 6O. g5,
and 59.&g5 Sxh2 6O. xf5, based on 60...g3 abcdefgh
61.Sb3t. 47.Sc7?
An unnecessary pawn sacrifice. It is easy
We shall start the final example just after the enough to follow Korleys thinking. By giving up
time control, where White is slightly worse, the pawn, he activates his rook and leaves the
but well within the drawing margin. black king passive, defending the f6-pawn. The
problem with this thinking is that it loses a pawn
Kassa Korley - Mads Andersen
- and the examples Korley will have seen where
Helsingor 2019 giving up a pawn to activate the rook is strong,
are all examples where the rook was terrible. The
white rook was already active, so in this case,
all White achieves is dropping a pawn, while
staying marginally within the drawing margin.

47.&c3 S g l
It is also hard for Black to make progress
after: 47...f5 48.gxf5t &xf5 49.Sg7! g4
5O.Sg8 The white king will come to e2 and
defend against the g-pawn if it advances to
the 3rd rank. Black can take the b4-pawn,
but would lose the g-pawn.
48.Sb7
abcdefgh White can hold in many ways, but as a
principle, I prefer to do it actively in rook
41.axb4 Saal 42.Sc2 Sxc2 43. xc2 Sfl
endings.
44.Sa3 Sxf3 45.&d3 Sfl 46.Sxa7 S b l
48...Sxg4 49.Sxb6t &f5 5O.Sd6 S g l
51.Sxd5t &e4 52.Sd6 f5 53.Se6t &f3 54.d5
S d l 55.Se5
Having over-pushed (forced to do so by the
author, admittedly), Black will now have to be
a bit careful to avoid losing the game, as well as
the illusion of an advantage.
 Chapter 1 0 - Checking Distance 169

47...Sb3t 48.&e2 Bxb4 49.Ec6f &e7

50.&B Ebl 51.Sc7t &d6 52.E£7 &e6

abcdefgh
58.&e2!
It is important to bring the king over to stop
the b-pawn. There is no better time than
now, when the black king is cut ofF on the
abcdefgh 8 th rank.
53.Sb7? 58. Ea6? does not work. After 58...b4!
The moment we have all been waiting for. 59.Exf6? E f l f Black wins.
White puts the rook in the seemingly natural 58.Ef7? is also poor. After 58. ..b4 59.&g2 b3
spot behind the passed pawn, and loses. The 60.Exf6 &c7 61.Ef7t &c6 62.Ef8 & b 5 the
drawing method relied on maximising the black king comes into the game, winning it.
checking distance with a peculiar-looking Black now cannot do anything without
move: dropping the d-pawn.
58... 6 b 8
53.Ea7!! 58. ..b4 59.Ea5! is similar.
Checking distance! 59.Ed7
53.. .b5 54.Ea6f &e7 55.Ea7t &e8 56.Ea8f White will not lose when he has a passed
&d7 57.Ea7t pawn of his own.
This is the key point. The king cannot
approach the white rook quick enough. 53...B5 54.§b6t &e7!!
57.. .6c8 Mads is a great endgame player.
57.. .6c6 allows White to draw with:
58.Ea6f! If Black plays 58...&b7, White 54...&f7? 55.Ed6 would allow White to make
has 59.Ed6!, which is the core advantage the draw.
of the greater checking distance. 58... c7
59.&e2! b4 6O.Exf6 b3 61. d3 and White 55.&e2 b4 56.&d3
is in time to stop the b-pawn. Black can try: 56.Eb7t &e6 57.Eb6t £f7 58.Eb7t &e8!
61.. .b2 62.&c2 Bgl 63.&xb2 Exg4 But his Eventually the king needs to go here. The
king remains cut ofF, and White can draw in king needs to go to e8/d7, not to d8, as
many different ways. 64.&b3 Bgl 65.&b4 White would have Eb5, taking the d-pawn
being the most logical to me. with check.
58...&g6 59.e4! draws. White will be push
the d-pawn and play Exb4 and Exe4, holding
the two vs one endgame easily, as Black has
170 Conceptual Rook Endgames

no passed pawn and the white king is ideally 67.. . 5 d l ! 68.Sxb2 xe3 69.&xg3 Sxd4
placed. 7O.Sb5 S d l 71.Sb3t Sd3 72.Sb5 Sa3!
59.SB5 Black wins, due to:
The checks are running dry. 73.Sxd5 &e4f
59.. .b3 60.&d2
60.§xd5 b2 61.§b5 S h i and Black wins. 56...b3 57.Sb7t &e6 58.§b6f &e7 59.Sb7t
60...b2 61.&c3 &f7! &e6 6O.Sb6t &f7! 61.Eb7t &g6!
Black wins. White cannot take on b2, as Again, this is where the king has to go.
Black exchanges the rooks and breaks with
...f5!, queening the g-pawn. 62.§e7
62.Sxd5 62.&c3 b2 63.Sb3!?
This also does not work. If White takes on b2, Black will win with the
62... 5 . i 63. xb2 Sxe3 ...f5! break again.
Black will grab the g-pawn and push the
connected passed pawns, winning easily.

White can also try passive defence, but will

get stuck and Black will find it easy to make
progress.
56.&f2 b3 57.&g2 b2 58.&h2 f5 59.gxf5 g4
60.&g2 &e8!
60...g3? would ironically throw away the
advantage after 61.Se6f! &f7 62.Sb6 and
the players are caught in mutual zugzwang.
abcdefgh
61.&h2 g3t 62.&g2 £f7
White is in zugzwang. Black will take the Black now wins by cutting ofF the king and
f5-pawn and enter the white position. putting the rook on the optimal square.
63.Sb7t £f6 64.Sb5 &g5 65.Sb6 &xf5 63. ..Self! 64.&xb2 §c6! 65.Sb5 Se6 66.§b3
66.Sb5 &e4 67.Sb3 f5!
Black creates a passed pawn as far away as
possible.
66...§e4? 67.&c2 Hxg4 68.§b5 would allow
White to draw.
67.gxf5t &xf5 68.&c2 g4 69.&d3 g3 70.§bl
§a6!71.Sgl &g4
Black wins.

62...b2 63.&c2 Sgl 64.&xb2 Sxg4 65.&c2

Ee4
0-1
abcdefgh
Black uses his advantage to transpose into a The advice is simple. If you have nothing
perfect version of the single pawn endgame. better to do, increase the checking distance. It
could save your game.
Chapter 11

Defence from the Front

As a quick intermezzo chapter, let’s have a look at a few scenarios where the best defence comes
from the front. It is rarely something we would go for, but at times it is appropriate, if nothing
else, then because everything else fails.

I want to underline that there are specific reasons why the defence from the front is appropriate
in each of these games.
First off, the black king is cut off and there is still time to prevent the g-pawn from advancing
to the 5th rank, where it would be unstoppable.

Magnus Carlsen - Maxime Vachier-Lagrave

London 2015

abcdefgh
MVL lost this rapid game mainly due to time trouble. But it would have been good to have
these sort of decisions on auto.

51...Sfl? 52.Sxh3 &e7

52...§f8 loses most simply to 53.§g3 and the g-pawn advances.
172 Conceptual Rook Endgames

53. g6 Sf6t 54.&g7 Sf7t 55. g6 Effrf 52.§dl? would also allow Black to successfully
56.&g5 Sa6 defend from the front. After 52...2e8f 53.&f4
56...&f7 57.Sh7t &e6 58.Sh6 and White Sf8 1 54.£e3 Se8f 55.&f2 Sf8, White cannot
wins. make progress. For example: 56.§d4 &c5!

57.EB 52...§e8t 53.&d4

1-0 53.Se5 d7! also works. If nothing else, the
king has come one file closer.
Black could swing the rook at once, or after
pushing the h-pawn first (which is more 53...5e6 54.5 Ed6!
natural to me). The simplest.

51...h2 52.§e2 Sal! 53.Sxh2 Ea8! 55.Sxd6f &xd6 56.F6 &e6 57.&e4 &f7
White cannot make progress. The checks 58.&d4 &xfi6
from the front secure the draw.

Generally, this topic belongs in Theoretical 52.§d3!

Rook Endgames. But as checking is a big topic This was the winning move. White defends
here, I wanted to at least show a few examples. the pawn and is now free to advance the king.
The next one is not too dissimilar from the If Black does not give checks, the pawn will
previous example, as it is all about advancing/ start running.
stopping the f-pawn, in a situation where the
black king is dangerously cut off. 52...Se8f 53.&B Sffif

Richard Palliser - David Howell

Torquay 2019

abcdefgh
54.&e6!
Threatening to use a few checks on the black
abcdefgh king to bring the rook to f5 or f7, winning.
52.f4?
54...§e8f 55.&£7 Se5 56.&g6!
Sloppy.
White is winning. The f-pawn will advance.
For example:
 Chapter 1 1 - Defence from the Front 173

56...§e6t 57.&B 52.&e3?

White wins. The black king is too far away. Koykka missed a chance to make the
conversion incredibly difficult for Black with:
White lost the following game after playing a
move he most likely did not even suspect was a 52.<±>e4! a3 53.&e5 a2 54.§fl
mistake until the computer told him so. The winning idea is just incredible.
54...§e8t!
Pekka Koykka - Elshan Moradiabadi The common idea of forcing the white king
to a worse square.
Stockholm 2019
55.&d6 §d8f! 56.&e7
The white king has to run forward. It cannot
remain on e5, as Black will manoeuvre the
king to c2 and defend from the checks with
the rook, after which the a-pawn will queen
with check.
56...§a8 57.f7 &c4
Freeing up the b-pawn.
58.§f4t & d 5 59.Sf5t &c6 60.§fl
60.§f6f &c7 61.Sfl a l = ® and Black wins.

abcdefgh
49.5? a5! 5O.F6!
Pushing the pawn is likely to give the best
chances. Slow play rarely works and indeed,
with a stronger 52nd move, White could
have made the win difficult to achieve for his
opponent.
abcdefgh
5O...Effi 51.Sf4 a4! 60.. .5.7tI! 61.&e6
61.&e8 al=W wins.
61... a l = ® 62.f8=®!
62.§xal? §xal 63.f8=® S e l f and White
has to resign.
62...®a2f! 63.&f6 ®b2f 64.£g6 ®g2f
65.&h6 ®h3t 66.&g6 ®h7t 67.&f6
Black is winning, but there is still a lot of
work remaining. For example: 67...&b7!?
intending to play on slowly.

52...a3 53.&d3 a2 54.§fl

Black wins uneventfully.

abcdefgh
174 Conceptual Rook Endgames

55.Sal SBt 56.&c2 &b4 57.Sxa2 g£2f 5 1 . 5 b5 52.&e4 a4 53.A6 &a5 54.&d3 b4
58.<&bl Sxa2 59.&xa2 &c3 60.&a3 b5
61.&a2 b4 62.&bl &b3
0-1

A solid win for Elshan, who later went on to

win the tournament.

White needed to buy time to organise the sort

of setup that would make a draw (see more
about this in the chapter on Connected Passed
Pawns). This can only happen with a defence
from the front.

49.EdH! abcdefgh
55.EA! Effi 56.f7 &b5 57.EBt
Black will not be able to make progress. The
key reason why these lines work is that the
black king is unable to come in front of the
pawns. When the king is behind the pawns,
the white king is able to come in and control
the pawns, making a draw.

I am sure that there are plenty of examples

where defending from the front is better than
defending from the back, although they are
not found in this chapter. It is the poor man’s
defence; the last straw to cling to, when all
abcdefgh other options have vanished. But sometimes it
49...a5 5O.Sblt &a6 is enough.
Against 50... &c5, White has to play 51.S a l !
to slow Black down. (51.S e l f ? would be a
mistake, as 51... d4 would bring the king
closer for no purpose.) 51...Ea8 52.&e4 a4
53.&d3 b5 54.f5 b4 5 5.f6 a3 56.f7 Ef8 57.Efl
And White makes the draw.
Chapter 12

Cutting off the King

In rook endings, each side has only two pieces left. One of them is the rook, which is generally
difficult to dominate, as it is the strongest remaining piece on the board. We have seen situations
where passed pawns and other responsibilities can force a rook into passivity, but it always takes
an army to hold it down. It is different with the king, which can be cut off completely by the
rook. Sometimes it is worth using this higher-value piece to dominate the king, while at other
times the rook can multitask.

The examples in this chapter will generally not be too difficult (the exception being the last),
as the theme is quite basic. The first deals with a common theme of “the king cut off along the
5th rank”, which is based on the idea that the black king will not be able to come to the aid of
the g-pawn when it reaches the 3rd rank and can be scooped up by the rook. (The same method
would not work further up the board, as on the 2nd rank the pawn would threaten to promote).

Teimour Radjabov - Lazaro Bruzon Batista

Tromso Olympiad 2014

abcdefgh
51.gb5!
Another useful variation is: 5 1 ... g6 52. b6! (52. b7 h5 53. a7 Sa3 and Black draws)
52,..§a3 53. a7! White is threatening 5=£b5-a5 and Black has to take. 53...Sxa7 54.<&xa7 &h5
176 Conceptual Rook Endgames

Trying to get around the 5th rank, but because 81...Sa7t? 82.&c6! also wins for White.
of the shouldering principle, this is hopeless.
55.&b6 &g4 56. &c5 White is in very good But Black had the surprising pinning idea of
time. 8 1 ...Sa5!I, when she will make the draw after
82.&c6 h4, or 82.Sc6f &g5 83.b6 Sb5, as the
52.&b6 g4 53.a7 Sh8 54.<&b7 Sh7t king is no longer cut off.
54...g3 55.a8=® Exa8 56.&xa8 g2 57.Ebl
and the king is too far away to help the pawn. 82.b6 h3 83.b7 h2I?
Black has to try something.
55.&a6 Sh8 56.Sb8 S h i 57.Sb6f!
Cutting the king ofF on the 6th rank, which 83.. . 5 b I 84.b8=® Sxb8 85.&xb8 leaves Black
is almost cruel. If the king goes to the 5 th rank, cut off along the 5 th rank with nowhere to go.
White will play 58.Sb5t and be a queen up. 85...h2 86. Sc I and White picks up the pawn.
White is totally winning and Black resigned.
1-0 84.b8=B hl=® 85.®g8t &h6 86.Sc6f
<±>h5 87.®g6f &h4 88.Sc4f &h3 89>g4f
The next example is seemingly thematically
1-0
identical, except there is a small difference.
Ernesto J. Fernandez Guillen - Leo Crevatin
Marie Sebag - Natalia Pogonina
Sitges 2021
Sochi 2015

abcdefgh
abcdefgh
The king can also be cut off from going
8O...h5 81.&c7 backwards. The following is a good example.
White cannot afford to spare a tempo, as After 65...Sg5t« 66.&e4, White drew
after 8 1 . £ c 6 h4 82.b6 S b l 83. b7 h3 84.Sc3 effortlessly.
h2 85.Sh3 Black will play 8 5...Sc I f and make
the draw. Instead, Black could have won with 65. ..Sg4!,
when the white king is cut off and Black can
81...h4? improve his king before advancing the pawn.
Black only had a few seconds left and missed Black wins in 39 moves according to the
a chance to save the game. tablebase.
 Chapter 1 2 - Cutting off the King 177

The next example is a good deal more fiddly. 7O...Sc5t 71.&a6

White has cut off the black king, but also needs
to make space for his own king to advance.
This turns out to be a bit more complicated
than it looks at first glance.

Francisco Vallejo Pons - Falko Bindrich

Germany 2018

abcdefgh
71...&c4!
The white rook cannot keep the black king
cut off.

72.§h6
72.Sbl S h 5 and Black is ready to give lots of
abcdefgh checks from the side.
67.Sb6?
72...gg5 73.&b6 Eb5t 74.&a6 Eg5 75.Shl
67.Sb8? is also wrong. Black can play
Sg6| 76.&b7 &b5 77.Sh5t &a4 78.a6
67.. .6d4 68. &b4 <±>d5 69.&b5 &d6 and
Sg7t 79.&b8 Sg8t 80.&c7 Sg7t 81.&c6
70.. . c7 with a draw.
Sg6t 82.&c5 Sxa6 83.Shl Eg6
y2 -%
After 67.Sh4?! Sb7t, White is still winning,
but has to come up with something else after
White would have been winning with simple
68.Sb4 Sa7.
play, keeping control.
67...&d4 68.&b4 &d5 69.a5 Sc7
67.Sb5! &d4
The rook could also have gone round the
After 67...Sh7 White can win with 68.a5, or
back via other files. This one gives Black the
the more human 68.Sd5t e4 69. c4 S h i
additional possibility of a check on c5, not that
70 .Sg5 when the black king is cut off.
it is needed.

70.&b5
White can also make no progress after 70.a6
Scl 71.Sh6 S a l , due to 72.&b5 S b l f 73.&a5
&c5! and the draw is near.
178 Conceptual Rook Endgames

Ivan Popov - Evgeniy Solozhenkin

Sochi 2015

abcdefgh
68.a5!
68. &b4? Sh7 69.a5 S h i 70.a6 S a l and
Black draws. 71.Sb6 S b l f 72.&a5 S a l t abcdefgh
73. & b 5 & d 5 is a thematic line. 74.Sb7 &d6
56...Sal?
White cannot make progress.
56...Sfl ! was the drawing move, cutting off
the white king. If White pushes the pawn to
68.. .§a8
the 7th rank, he will have to let the black king
68...Sh7 69.a6 S h i 70.a7 wins, because
re-join the fight, when it is easy to see that it
70.. . 5 a l runs into 71.Sb4f and 72.Sa4.
will take the g4-pawn and the h-pawn will
make the draw. But there are no other ways for
69.&b4 Sa7 7O.Eb8!
White to improve his position.
This is the key difference. In the game the
rook was in the way on b6.
57.&f4 Sa5
70.. .6d5 71.&b5
White wins. The king will advance and a 8
check on d8 will drive the black king away. 7
The next example takes a big step up in 6
complexity. White has the advantage. He is a 5
pawn up and his rook is well placed, defending
the pawn and cutting off the black king. But 4
still it was possible for Black to hold the draw, 3
as his rook is the best piece on the board.
2
1
abcdefgh
58.Sc6?
The white rook is perfectly placed. It is time
to bring the king.
 Chapter 1 2 - Cutting off the King 179

58. <±>e4! I would assume that Popov believed that he

Black will have to do something, as White was winning and that everything was going
will otherwise bring the king to b4-b5-c6 well.
anyway.
58...h5 Instead, 66.Sc7I! followed by a check on c5
After 58...Sa4f 59.<±>d5 Sxg4 60.&c6 and either c6 or c4, transferring the rook to
White wins, as the black king is cut off far the a-file, would have won.
back.
59.gxh5 Sxh5 60.< ?d4 66.. .Eb3t 67.&a5 h4?
White is totally winning. The rook is ideally Black misses a similarly simple, yet difficult
placed on b6. Almost any method will work. to spot possibility.
For example:
60.. .617 6 1 . £ c 4 Sa5 62.&b4 S a l 63.Sh6I? 67.. .5a3t 68. &b4 Sa6!! Delaying giving up
&e7 64.a7! &d7 65.Sh8 Sxa7 66.Sh7t the rook, as taught in the chapter on Rook vs
Pawns. 69.Sd5t &g4 7O.Sa5 Sxa7 and Black
58.. .5g5? draws.
58...h5! held easily. The black king will race
to the queenside after 59.gxh5 Sxh5. 68.Sd4
1-0
59.Sd6 &h7 60.&e4! Sxg4f 61.&d5 Sa4
The next example is a surprising twist on the
62.&c5 h5 63.&b5?!
theme.
Although this wins a tempo on the rook to
get to the b-file, it makes little sense.
Rauf Mamedov - Hrant Melkumyan
The aim should be to advance the a-pawn, Heraklion 2017
which is best done with: 63. b6! h4 64.a7 h3
and now 65.Sd3, or 65.Sd5 and White wins.

63...Sa3 64.Sd7t &g6 65.a7 $6

abcdefgh
The game is terribly close to a theoretical
draw, but White still has a chance.

abcdefgh
60.<$g3?
66.&b6? Not the right choice.
180 Conceptual Rook Endgames

6O...ghl 61.Sxa5 After 61. e3? Sh2! the king would be cut off
This setup should be a draw, but it is still from stopping the pawn and the rook would
possible for Black to make a mistake. have to deal with it, allowing a draw.

61...Sglt 62.&£2 Sg4 63.&B Sgl 64.8e5 61...a4 62.&dl a3 63.&c2

Shi 65.&e4 Sdl 66.Sd5 Self 67.&d4 63. £ c l ? ? would be a blunder on account of
Scl 68.Se5 Sdlt 69.&c5 Sd8 70.&c6 63.. .S h i fl, when Black even wins.
Sa8 71.&d7 Sa7t 72.&e8 Sa8t 73.&e7
Sh8 74.8d5 Sh7t 75.&d8 Ea7 76.&c8 63.. .5.2t 64.£bl
Ea8t 77.&b7 SfB 78.&c7 Sfl 79.Sb5 White will be able to promote the pawns
Sf7t 80.&c6 Eft 81.£d7 Ea8 82.§d5 Eft slowly.
83.&e7 Sf7t 84.&e8 Sa7 85.&d8 Sh7
86.&c8 The next game is a bit more complex. This
was the long game in the first round that kept
me in the playing hall of the 2018 Olympiad,
8
which made it possible to spectate the game
7 Aithmidou - Li Chao (see page 304). Earlier
6 in the game Tania got into horrible difficulties,
but had managed to fight her way back and
5 was now dreaming of getting the most out of
4 the game.

3 Tania Sachdev - Vyanla Punsalan

2
Batumi Olympiad 2018
1
abcdefgh
86...Sa7?
A blunder in perpetual time trouble.

After 86...Sh8t 87.Sd8 Sh7 Black is holding.

87.Sd7 Sa8f 88.&b7 Sg8 89.&c7 Sa8

9O.gg7t $6 91.&b7 Se8 92.h7
1-0

The winning option was all about ensuring

that the white king is able to get to the a-pawn. abcdefgh
41.Sxh6?
6O.Sg7t &f5 61.&e2!!
A bit careless, giving her opponent a final
We will discuss this in more detail later.
chance to win the game.
The white king belongs in front of the passed
pawn, so it should seek backwards urgently.
 Chapter 1 2 - Cutting off the King 181

41.Sf6f e3 42.d6! would have held the Exploiting that the white rook cannot be on
draw, but in all fairness to Tania, the game had f6 and check the black king on the d-file at
turned so much that she was trying to win it the same time.
by this point. 49...Sxd6? 50.< ?g2 and White holds.
50.&gl &d3 5ESf5 Sxd6 52.£fl &d2
41.. .6e3? Black wins.
Black had an elegant win:
41...6f3! 42.d6 42.d6 &d2
42.Sf6f e2 and ...Sg7t is not an
improvement.
42...e3
The pawn is close to promotion. On top, the
d6-pawn is in the way of the white rook.
43.Sf6t &e2

abcdefgh
43.Sh5!
The rook belongs behind the passed pawn.
abcdefgh
43...Sd7?I
44.Sf5
43.. .5g7t 44.s£?fl Sf7t would have forced
The best attempt to activate the rook, but it a draw at once. But Punsalan was also hoping
does not work. to win.
44...Sg7t 45.&h2 Sd7 46.Sd5 &f3 47.Sf5t
&e4 48.Sf6 &e5 49.§f8
44.Sd5t &e2 45.&g2 e3 46.&g3 £el
47.&B e2 48.&e3 £fl?!
48...5f7! with the threat of ...<&fl, and
the idea 49.Sd2 Sf2! would have drawn
effortlessly. 50.d7 &fl 51.Sxe2 Sxe2f 52.&d4
Sd2f 53. c5 Sxd7 54.a4 is a plausible end of
the game.

49.SBt & e l !
Black would already be lost after 49...&g2?
5O. xe2 Sxd6 51.a4!.
abcdefgh
5O.Bh5!
49...&d4!
182 Conceptual Rook Endgames

54.Sh2?
After the game I assured Tania that she had
played the endgame brilliantly.

But it turns out that she was right and that she
here should have played: 54.Sf2f! and 55.Sf4,
advancing the a-pawn. The black king is too
far away.

54.. .5a6 55.&d3 £ e l ?

Although there were a few moves that drew
for Black, it is safe to say that we have already
abcdefgh encountered the correct way to play in such a
The critical moment of the game. The black position: 55...Sa8! and the checking distance is
king is in great danger of being cut off, leaving optimal, making all of White’s tries fall short...
the three white chess men alone to decide the
game. 56.&C4!?
This wins, but delays the moment where
50...&dl? good moves will have to be played.
The wrong choice.
The strongest was 56.< ?c3! Sa8 57.Sh4!!,
50... fl ! 5 1 . S h i t (5ESf5t & e l 52.Sf2 Sxd6 when the pawn advances and the black king is
53.Sxe2f d l is a key point. The king is close hopelessly cut off.
to the pawn, which it will quickly eliminate.)
51...6g2 52.Sel Sxd6 53.Sxe2f &g3 The On the other hand, 56.Sh4? would fail to
black king has escaped the back rank, giving 56...5a3f! with an immediate draw.
Black a draw. 54.a4 Sa6 55.Sa2 Sa5! 56.< ?d4
&f4 The black king is in time, heading for c7
with a draw.

51.Sd5t &el 52.Sd2! <£fl 53.Sxe2 Sxd6

abcdefgh
56...&dl?!
A sign of hopelessness.

abcdefgh
 Chapter 1 2 - Cutting off the King 183

56...Sa8! would have made the win difficult Anton Demchenko - Antoni Kozak
for White. After 57.&b5 Sb8f 58.&c6 Sa8
Piestany 2022
59. b7 Sa3 60.s£?b6 Sa8, Tania would have
had to find:

abcdefgh
abcdefgh
61.Sh4!!, exploiting that the black king is cut
off for tactical purposes. 61...Sa3 62.< ?b5 75.Se8?
£ d l 63.Sa4 Sh3 64.Sd4t &c2 65.a4 And White could have drawn in a number of
White wins. ways. I like the draw with checks from the
front, so for this reason I would have chosen
57.&b5 Sa3 58.&b4 Sa8 59.a4 Sb8f 75.Sg2 Sxh4 76. xb3 Sxh5 77. c4!. It is
60.&c5 Sa8 61.Sa2 £ c l 62.a5 £ b l 63.Sa4 important to prevent ...Sd5, cutting off the
&b2 64.a6 &b3 65.Sal Sh8 66.a7 Sh5t king. It now comes to d4 and is close enough
67.&d6 to make the draw.
1-0
75...Sxh4!
Starting out, the next endgame should be The game continued: 75...Sb5? 76.Se4 < ?h6
harmless for White, but by misplacing the 77.Sg4 with a draw after so many extra moves,
rook (which is not at all easy to see), he gave it might have been finished the next day...
Black a big chance to win the game.
76.&xb3 Bxh5
76. ..< ?h6? would be a mistake. 77.Sh8f
&g5 78.Sh7 with a draw.

77.&c4
77.&c3 loses to 77...Sd5! 78.Sel, where
Black can improve the position slowly and
win. All he has to avoid is 78. ..g5? 79.Se6!,
when the king is cut off and it is not possible
to make progress.
184 Conceptual Rook Endgames

83...Sg3
8
Black is winning after either 84.<&e5 Se3t,
7 or 84.Sg8 < ?h4.
6
The next example is rather deceptive. But
5 whenever we see a race, we should look for
4 ways to cut the king off. By now I hope you
have realised that this can happen in all four
3 directions.
2
Tomas Sosa - Dhulipalla Bala Chandra Prasa
1
Gibraltar 2019
abcdefgh
77...SE3!!
The key move. The white king is cut off and
will not be able to help in the defence.

78.&d4 &g6! 79.Sg8

79. <&e4 < ’g5! is important. The white king
is shouldered.

79.. .617!
This triangulation with the king is important.
It wins an important tempo to advance the
g-pawn.
abcdefgh
8O.Sd8 81.Sg8
81.&e4 &g5! is an important detail. The Black won this game after mistakes from
black king is shouldering the white king — thus both sides:
assisting in cutting it of.
54.. .6b4?
81...g6 82.&e4 &g5! 83.Sg7 This was the most logical move in the world,
but it allows White to get counterplay.

55.e5!
55.&g4? would allow Black to cut off the
king after 55...Sa5!.

55.. .6b3 56.§al Sa4

abcdefgh
 Chapter 1 2 - Cutting off the King 185

Alexander Donchenko - Ezra Kirk

France 2022

abcdefgh
57.g3?
This allows the king to be cut off and frankly
loses a tempo for absolutely nothing. abcdefgh
50...g5? 51.c4 £f5 52.&b4 <$f4 53.&b5
White needed counterplay and could have
&e3 54.Sg2 Se5t 55.c5 &d4 56.&b6
drawn with: 57.e6! Sf4f (57...a2 58.e7 Sa8
And resistance is futile.
59. S e i is a simple draw. The white rook will
1-0
give lots of checks.) 58.< ?g3 Se4 59.S b I f c2
6O.Sb6 And the rook will give checks from the
Black would have been able to draw with an
back.
unbelievable combination of giving checks
from the front, as well as cutting off the white
57...g4f!? 58.hxg4 a2 59.e6 Sa6 6O.Sel
king.
al=® 61.Sxal Sxal 62.&e4 &c4 63.&e5
&c5 64.&fi6 &d6 65.g5 Sa7 66.g4 Se7
50...SE4!! 51.c4 Sh8
67.&xg6 &xe6 68.&h6 &£7
Advancing with the king allows a frontal
0-1
defence, with a few details, which we shall skip.

Instead, Black could have won the game

52.c5
with 54...Sa5!I, which is threatening ...&b4,
The frontal defence is no longer possible
winning. The key point is that after 55.e5,
when the pawn has crossed the middle of the
Black has 55...Sa4!, keeping the white king
board.
cut off, preparing to take the e-pawn with the
king.

The final game is fantastic and no one would

ever have noticed anything odd about the
game, had computers not pointed out that
Black missed a chance to hold. The finish was
uneventful.
186 Conceptual Rook Endgames

The next game shows that the author is

8
better at writing about than playing chess...
7 Despite looking deeply into the ideas of rook
6 endings the last few years, I still struggle with
them in practice. In equal parts, overthinking
5 and blundering were too strong for me to
4 overcome. First off, I was confused if I should
take on c4, or delay it one move. I got entirely
3 confused and decided that I should play the
2 simplest and take the pawn. But in reality,
I should have played simply by removing
1
options for the opponent.
abcdefgh
Viktor Haarmark Nielsen - Jacob Aagaard
52...SE4!!
The surprising defence. The white king is cut Svendborg 2023
off.

53.c6 &e7
Defending against c6-c7 followed by Sd8, as
well as threatening ...Sh6 in some situations.

54.Sd7t
Advancing the pawn would see it lost.

54...&e8!
The king is ready to block the pawn.

abcdefgh
44...Bxc4?
44...&f7! was the right move, transposing to
the game.

Viktor now had a chance, but was thinking in

the same way as me.

45.Sa6? &f7 46.&g3

abcdefgh
55.Sxg7 &d8 56.Sg2 &c7 57. Sc2 SE6
The simplest. Black draws.
 Chapter 1 2 - Cutting off the King 187

5O.Se7t!
5O.Sg7 Se4 51.Sxg4 Sxg4t 52.&xg4 <±>e4!
was my intention. But on move 47, not 46.
From there I have nothing intelligent to share.

50...&d5 51.Sd7t &c6 52.Sg7 &b5

53.Sb7t &a4 54.Ec7 &b4 55.Sb7t &c3

abcdefgh
The new moment of overthinking. I wanted
to play 46...Sd4, the most natural move in
the world, but I was confused about the idea
of <&h4-g5-f5 and so on, with counterplay. I
decided that I needed for the king to help the
c-pawn first.

abcdefgh
46...&e7?
I had this idea that I was being really clever. 56.Sc7!
In reality my brain had taken leave of its senses. Black cannot make progress. As I refused to
“take a draw”, I constructed a nice way for the
46...Sd4 would win easily. The king comes game to end.
up and the pawn advances. I understood this,
but being really rusty, I struggled with general 56...&d3 57.Sc8 &e3 58.Sc7 Sc2 59.&xg4
decision making in the tournament. Sg2t
He had a sudden jump in the chair, when he
47.Sxa7t &e6 48.Sg7 $6 49.Sf7t <&e5 realised my trick.

abcdefgh abcdefgh
188 Conceptual Rook Endgames

60.&h3! Let’s return to the thematic moment.

60.&f5? Sxh2 and Black wins. 61.Sxc5

abcdefgh
Stalemate.
Chapter 13
The King’s Role
I have been working on this book for the last five years (with greater and greater intensity, which
is why most of the examples are recent). Once the idea of it became clear in my mind, I talked
to my colleague, G M John Shaw, to explain the ideas. One idea, which I thought was greater
than all the others, was the answer to “what is the role of the king?” I had asked the great rook
endgame specialist and enthusiast Boris Gelfand this question, and he said he did not know,
and had never given any thought to this question. I said to John I had all these great ideas and
concepts, and the book would be different to all previous books about rook endgames. “I have a
useful observation too,” John said. “The role of the king is to deal with passed pawns.” He must
have realised the horror in my face, as he fell silent. “Indeed,” I managed to push past my gritted
teeth. “Indeed so...”

Joking aside, it is interesting that John and I, two people with wildly different thought processes,
came up with the same key point. We all know what the role of the rook is in rook endgames.
But with only two pieces each, it is rather careless to forget about the other piece entirely! When
we say that the king has to deal with passed pawns, it should be understood like this. If we have
the advantage, we should aim to have the king helping the pawns promote. This is the winning
plan after all. If we are defending, the role of the king is most often to block or control the pawns.
Obviously, there are scenarios where we have the advantage, but the king’s role is defensive. As
there are also positions where we are trying to hold, but the king’s role is offensive. Below we will
see a number of interesting examples that will hopefully make this far more tangible.
190 Conceptual Rook Endgames

Viktor Haarmark Nielsen - Martin Haubro Evgeniya Doluhanova - Vitaliy Bemadskiy

Taastrup 2018 Sitges 2023

abcdefgh abcdefgh
53...&c3?? Not knowing what to do, White missed the
A move made out of pattern recognition. chance to improve the position of her king.
This seems to be the way the king is going in
these positions. But that’s because the rook 39.E3? a4
usually is in time to block the b-pawn. Here, Black will win the f6-pawn anyway, as the
it is not, so the king belongs on the 8th rank, rook is soon diverted.
not the first.
4O.Sc2 Sb5 41. b3 42.axb3 axb3
53. .. c5! was winning easily. 43.Sb2
White resigned.
54.Sdl &c2 0-1
After 54...bl=® 55.Sxbl Sxbl 56.h5 we
can spot that Black would win with king on White could have drawn with: 39. f4!!
c5 (and even c4), as it can race back. Instead,
it will now get shouldered and White holds.

55.Sfl b l = ® 56.Bxbl &xbl

You cannot get farther away from the
action... The game ended a draw on move 67.

abcdefgh
An easy move to miss for sure.
 Chapter 1 3 - The King’s Role 191

39... a4 40.<&e4 S b 5 41.<&d4 and the draw is Jaime Santos Latasa - Lev Yankelevich
not far away.
Sitges 2022

And after 39... xf6, 40.<&e4t <&e6 41.Sf6f!

is an essential tactic. Both the a-pawn and the
h-pawn will queen.

Improving on the next game should be

automatic by the time you reach the end of
this chapter, if not long before.

Austin Tang - Leo Lin

Internet (classical) 2 0 2 1

abcdefgh
48...<&d5? 49.Sf4 h4 5O.f7 Sb6f
5O...Sb8 could be tried, with the point that
51.f8=®? Sxf8 52.Sxf8 <&c4! is a draw. But
White wins after 5 1 .Sf5t 1, when the rook will
capture whichever pawn the black king tries to
go to defend.

51.£15 Sb8 52.Sa4 h3 53.Sxa5t d6

54.&ft> Sh8 55.Sh5 h2 56.Sxh8 hl=£)
1-0
abcdefgh
41...Se2? 42.d7t &d8 43.&d6 Se8 Black could have made the draw with
44.dxe8=®t xe8 45.Hb7 h5 46.Hb5 &£8 48...<&d7!. White cannot play 49.f7 without
47.Sxa5 b3 48.Sb5 being harassed by endless checks. And on
1-0 49.Sf4, the king can run to f8 with 49... e8!.
Put the king in front of the passed pawn if you
Instead, Black could have held the game with can - it is that easy.
41...< ?d8!, when 42.d7 could be answered
with 42...<&c7 and White would even have to The next game ended in a draw after mutual
be a bit careful. mistakes.

The “simple but difficult” examples will keep

coming till you find them easy... and perhaps
way beyond that.
192 Conceptual Rook Endgames

Justin Tan - Nikola Nestorovic 55...&d7! 56.&d5 &c7

Black is preparing 57. c5 Sc6fl, where
Germany 2022
White cannot go for the pawn ending, so the
king comes to b6. And if instead:

57.E5
Black would have to find:

57...Sb8!
57...Sd6t 58.<&e5 Sb6 59.&d5 only leads
to repetition and the requirement to find the
win again.

58.&c5 Se8!

abcdefgh
54.Sb4? <&xe7 55.£e5 Sb8? 56. ?d5!
57.Sftt! &g6 58.&c6! h5
58...b4 59.£c7 2b5 60.&c6 is a common
pattern, well worth paying attention to.

59.g5 Sc8f 60.&xb5 Sc3 61.Sd4 Sg3

62.&c5 Hg4 63.Sd6f
Not a difficult move. White draws.

63...&f5 64.Bd7 Bxh4 65.Sxg7 Bg4 66. g6

E4 67.&d5 Sg5 68.Bg8 Bxg6 abcdefgh
l
/2-y 2
Black is winning. The rook will give a check
on the 5 th rank, then the king comes to defend
Black overlooked a win with:
the b5-pawn and finally the rook will be free
to manoeuvre, for example to a4, which will
be deeply annoying for White, or g3, which is
just mean.

White’s defensive method is not surprising,

even if his margin is slim.

54.&e3!l Sb7

abcdefgh
 Chapter 1 3 - The King’s Role 193

58.&B!
The simplest.

58...&e7 59.&e5 Sb6!? 60.&d5

abcdefgh
55.&d3!
An important detail.

55.&d2? would lose to 55...Sxe7 56.Sb4 Se5,

abcdefgh
where the rook is allowed to drop its anchor
on e5 with enough time for the black king to 6O...Sb8!
provide support. 60...&d7 61.<&c5 &c7 62.Sf4 is an
uneventful draw. For instance, 62...Sf6 63. Se4
55...Sxe7 56.Bb4 Eb7 and Black gets nowhere.
Going to e5 no longer makes sense. White
has 57.<&d4. 61.Se4f! &d7 62.Sb4 Sb7 63.Sf4 b4
64. >c4 &e7 65.&b3
57. e4 <±>£71? White makes the draw, but not without a bit
Trying to do something clever, as after of discomfort...
57...<&d7 58. <&d5! White makes the draw.
Black is a tempo short and not able to reroute The following position is clearly dangerous for
the rook to e8 as above. White, but by following the program, he could
have saved it.

abcdefgh
194 Conceptual Rook Endgames

Sergey Fedorchuk - Ivan Saric Shakhriyar Mamedyarov - Levon Aronian

Corsican 2013 St. Louis (rapid) 2019

abcdefgh abcdefgh
White lost after: In this time scramble, Black gave White a
chance to save the game.
54.Bg5t? &xh4 55.Sg8 Be7!
The key move. The white king is cut off and 43...&e6? 44.Se2f?
Black is free to exploit his pawn advantage. A foolish check.

56.b5 £2 57.218 &g3 58.Sg8t &h2 59.218 Instead, White could have played: 44. d7! The
<&gl 60.2g8f fl 61.b6 2b7 62.Sg6 game will end in a draw after either 44...Sc7
h4 63.2h6 h3 64.<$e3 &el 65.2c6 2e7t 45.Se2t &f7 46.Sc2, or 44...Sxd7 45.Se2f.
66.<&d4fl=®
0-1 44...±d7 45.Sdl &c6
0-1
Hopefully it is not a great surprise that White
had to play: You will not be surprised that the winning
move was: 43...£e8!!, when Black wants to
54.&el! To draw this endgame a pawn down play ...&d7 and ...Sdc8, and just wins.
relies partly on the common occurrence of
draws with f- and h-pawn. 54...Bc7 55.Sg5t The way for White to hold in the next example
&xh4 56.Bg8 Bc2 Trying to cut off the king, is not too complicated and should look
but the black king is also in trouble. 57.&fl familiar. But the winning lines after ineffective
Black cannot make serious progress. When he defence are spectacular.
takes the b-pawn, the king comes to f2.

This theme is also decisive at the highest level.

 Chapter 1 3 - The King’s Role 195

Damian Lewtak - Ihor Samunenkov

Piestany 2022

abcdefgh
79...Sa3!!
The key move. In line with our theme, the
white king is unable to participate in the
fight and Black wins - if not easily, then at
abcdefgh least slowly...
8O.Sg8t &f3 81.Sf8t <±>g2 82.Sg8t Sg3
70.&el? f4 71.a7?l
83.Sh8 Sg4t! 84.<&e5 h4 85.Sa8 h3 86.<&f5
A hopeless defence. White could still make
Sb4 87.Sa2t &g3 88.Sa3t
the win difficult for Black.

71.&fl!
This would lose to the idea we saw above,
but only after a lot of complications and
accuracy.
71... g3!
71.. .6f3? 72.& g l h4 73. a7 would allow
White to escape with a draw. The rook is
stuck on a8, but so is the king on f3. Black
cannot make progress.
72. a7 abcdefgh
The check is useless, as Black would play 88...&h4! 89.8a2 Bg4 90.8f2 8g2
72... <&f3! threatening mate, thus winning And things are going in the right direction...
the a-pawn.
72. . . 5 a l t 73.<&e2 f3t 74.<&e3 Sa3t 71...&S!
The f-pawn decides the game. White has to This too should be familiar too.
give checks all the way.
75.£d4 £2 76.Sg8t &h2 77.Sf8 §xa7 72.&dl h4 73.&el h3 74.gh8 8xa7
78.§xf2t &g3 79.Sf8
75.8xh3t g2 76.gh8 8e7f!
And it is time to resign...

77.&d2 f3 78.&d3 £2 79.Sg8f £fl 80.8b8

8f7 81.&e3 &g2
0-1
196 Conceptual Rook Endgames

White needed to seek the right place to block 76..T3

the pawns.

70J&fl! f4 71.£gl!
The only move. The king is perfectly placed
on g l . Any ...&f3 move would come without
the threat of mate, which is the key point, but
the king is also closer to the h-pawn, which is
not insignificant.

71.2g8f? &f3 we already know.

But we should take pleasure from the key abcdefgh

points arising after:
The astonishing thing is that this position
71.a7? <£f3 7 2 . £ g l
is mutual zugzwang. Black is only winning
because White cannot play 77. h3 due to
mate.
Now he loses after:
77.<&xh5 Sa4! 78.<&h6 Sa5!
When the king either has to go to g6, when
Black gets a free square for his king, or to
h7, when the a-pawn would fall with check.
The variation from here is longer than you
would suspect. Black will have time to play
...&fl without losing the f-pawn. Then ...f2
abcdefgh and soon after be able to dominate the white
king in the same way, as the king cannot
72...Sa4!
The rook can go anywhere on the a-file, with occupy the e-, f- or g-files without providing
cover for the black king. And on h7 there is
the exception of the corner:
the check on a7 if the rook moves.
72.. . S a l t ? 73. &h2 is a position of mutual
zugzwang. 73. ..h4 (It is too late for waiting 79 .<&h7 <£fl 80.<&g7 f2 81.&h8
moves. For example: 73...Sa5 74.<±>h3! This was all I could come up with. But Black
Sa2 75.Sh8 Sxa7 76.Sxh5 is a theoretical can lose a move, whereas White cannot.
draw. But simpler is 75.Sb8! followed by a 81...Sal
Zugzwang and wins.
lot of checks from a distance.) 74. Sb8 Sxa7
75.<&h3!! We now see the big difference
from the main line. The h-pawn is hanging! 71...&B
After 71... <&g3 72.Sg8t &f3, White would
Was it still on h5, Black would be winning.
draw with both 73.Sg6 and 73.Sa8.
73. <£hl &f2 74.<£h2 Sa2! 75.<&h3 S a l
76,&h4
76. <&h2 &f3! would put White in zugzwang.
The a-pawn is lost and the endgame with
f- and h-pawn is lost too, as the king is
misplaced on h2.
 Chapter 1 3 - The King’s Role 197

Jose Francisco Veiga - Dominik Horvath

8
Chennai Olympiad 2022
7
6
5
4
3
2
1
abcdefgh
72.Ba7!!
It is important for White that he will not
get into zugzwang. This specific move we will abcdefgh
see often in the chapter on Anticipation (see
44...Bxb3?
page 273).
44...<&g3? 45.&fl &f3 46.Sf6t! <&e3 47.h4
White would have enough counterplay to
72...E4 73.&E1
make the draw.
73.Sa81? also holds.
44.. .<&e3? 45.Sh5 <&d3 46.Sd5t! also holds.
73...&E2 74.H17 B 75.a7
Either the black king has to block the pawn, or
White makes a draw, as the black rook is tied
the white king will.
to preventing white from promoting the pawn.

45.gh5 £e3 46.8d5 Sd3 47.8xa5 £<12

The next example shall see the black king serve
48.Se5 e3 49.h4 8d4 5O.h5 e2 51.&E2 8f4f
a double role of assisting the passer and doing
52.&g3 SB 53.2xe2f!
a bit of shouldering. I don’t think Black ever
The usual escape route.
suspected that he was winning at the point
where we join the next example, even if he was
53.. .6xe2 54.&g4 Ba5 55.E6 &e3 56.E7
trying to create chances.
Sxa4f 57.&B Sa8 58.&g6 &f4 59.&g7
Vi-Vi

The winning idea is often seen, but can be

easy to miss (see also Sachdev - Punsalan, page
180). The idea is to get the king to where it
longs to be (e2), while trying to win a tempo
by pushing the opponent’s king away.

44...&B!
Threatening ...Self.
198 Conceptual Rook Endgames

45.W The game was drawn after:

No other moves make sense.
32...Sxb2? 33.Sxa5t &f4 34.Sa4f &e5
45...&e2 35.SE4 E5 36.g4 hxg4 37.hxg4 f5 38.gxf5

8 Instead, Black could have played 32...<&f5!!

33.Sxa5t &e4, followed by eliminating the
7 b2-pawn and winning the game. White does
6 not have to take on a5, but Black will find
it easier to improve his position until the
5
moment when ...Sxb2 is stronger, or White
4 finds it impossible not to take on a5.
3
First principles. Use the lower value piece to
2 block the pawn. Don’t use the higher value
1 piece.

abcdefgh Sandipan Chanda - Le Quang Liem

46. E4 e3
Xingtai 2019
46. ..Sc51? would also win.

47. h5 &d2 48.Sd6t Sd3

Black wins, for instance:

49.Se6 Bd5! 50.E6 e2 51.E7 Sd8

And it is all over.

The next example is surprisingly simple.

Hannes Stefansson - Ante Brkic

abcdefgh
4O.Sc2?
A passive move. The rook is a poor blockader
and White loses the game without any chance.

40...&g6 41.&e4 g5 42.E3

Sandipan is trying “something”, but there is
nothing.
42.f3 is what White would have liked to
work, but after 42... h4! 43. &d4 hxg3 44.hxg3
f5 45.Sxc4 Sxc4t 46.<&xc4 f4 Black wins.

abcdefgh
 Chapter 1 3 — The King’s Role 199

42...c3 43. < ’e5 ft»t 44.&e6 h4 45.f4f l&g6

46.g4 Se8f 47.&d5 Se3 48.Sh2 B 49.&d4
gg3 5O.g5 S£3 51.&e5 Se3t 52.<&d5 Sd3t
53.&e5 gd2 54.Shl ge2f 55.i?d6 Ee4
56.SF1 c2 57.Scl Exf4 58.Exc2 gf3 59.&e5
&xg5 6O.Sh2 Ee3t 61. d4 &f4
0-1

40.&e2! was the principled move. If the Black

king comes to g6, White would give a check
and put the king on d2. Otherwise, the white
king will block the c-pawn, as well as attack
it, trying down the black rook. 40...c3 is thus abcdefgh
the critical try. But after 41.Sc2 &g6 42.&d3 57.g4??
$6 43.B, the arising pawn ending is a draw. Splitting pawns is as bad as it can get.

The next example can appear a bit more 57.Hf4! He5t 58.&f3
abstract at first glance, but the principle is the This was the most principled way to play.
same. Aronian clearly seemed adrift and soon White is threatening Hd4f followed by He4,
ended up in a lost position, showing the value when the pawn ending is winning, as well as
of the idea presented in this chapter. Hf4-f6, attacking the g-pawn. Black has to
allow something bad.
Anton Demchenko - Levon Aronian
58...Hd5
Riga 202 1 58...g5 59.hxg5 Hxg5 60.&g2! wins without
too many difficulties.
59.Hf6 g5
White is winning in more ways than one.

abcdefgh
abcdefgh
53...S16? 54.i?e2 <&d8 55.Sxe4 <&d7
The most convincing is brutal:
56.&e3 Sf5
60.&e4 Ha5 61.hxg5 Hxg5 62.Hf5 Hg4f
63.&f3
The h-pawn falls and White wins.
200 Conceptual Rook Endgames

57.. .hxg4 58.Bxg4 Shant Sargsyan - Rinat Jumabayev

Black would even be drawing in the f- and
Moscow 2019
h-pawn endgame, but Black does not have to
let go of the g-pawn.

58...5E5 59.Se4 &d6 60.B g5 61.hxg5

Sxg5 62.&f4 Sg8 63.&P5 64.<&g4
<&d5 65.Se7 Sg8f 66.&f4 Sffif 67.i?e3
&d6 68.Se4 Ef6 69.Se8 i?d7 7O.Se5 &d6
71.Se4 Sf7 72.Se8 Ee7t 73.Sxe7 &xe7
74.&f4
Vi-Vi

The drawing method is unsurprising.

53.. .6d8! 54.Sxe4

abcdefgh
Sooner or later White has to take the pawn.
Black lost without any complications after:
54.. .6d7 55.i?e3 Sa5
31...a4? 32.Bxc6 a3 33.Sc7! Ba8 34.d7
Black resigned.
1-0

We should also cover:

31...c5?
This loses in longer and more complicated
lines. The following variation is rather fun,
though not entirely forced.

32.Sc6 Bd7 33.Sxc5

33. < ’g2!? also wins.
abcdefgh
33...5.d6 34.Bxa5 B d l f 35.&g2
White cannot improve his position without
Black does not have time for ...Bel, which
allowing ...&e7, when Black equalises.
would hold if there was not a check on g5.
56.B Bb5 57.g4 Ba5 58.g5 Bb5
35...h6I? 36.Ed5 Bel 37.Bd4 h5
This would not make serious progress for
White either, but it is all I can think of.

The next example is equally formulaic.

 Chapter 1 3 — The King’s Role 201

abcdefgh abcdefgh
38.Sd5! Sxc4 39.Sxh5 &g7 40.&g3 &g6 35.Sbl i?c6
41.Eh4 6 42.h3! ftf 43.exf4 Black draws.
White is winning on account of:
The next example is more complicated, but at
43.. .gxh3 44.f5t! its core, it is no different.

The principled move is the same as always. Dion Krivenko - Lars Johan Brodtkorb

Fagernes 2020
31.. .6B!

8
7
6
5
4
3
2
1
abcdefgh
abcdefgh
Black lost after a continuation that seemed
32,Sxc6
quite natural, but was simply bad.
Black draws easily after 32.Hb8f 'ig? 33.Sc8
and now either 33. ..a4, or 33. 34.Hxc6
41...&xe6? 42.g4! hxg4 43.h5 15 44.h6
&e6 and only ten advancing the a-pawn.
The king is not good at controlling pawns
from the side. This is not exception.
32...<&e8! 33.SB6 a4 34.c5 &d7!
34.. .a3? would lose to 35. c6.
202 Conceptual Rook Endgames

45.g7 Sd8 46.Sxf5f! Sam Shankland - John Burke

Black resigned.
St. Louis 2021
1-0

The drawing variation is nothing short of 8

magical. 7
41...Sxe6! 42.g4!? 6
42.Ef5 forced Black to find 42...&f8! 5
43. Exh 5 <±>g7! 44,Eh7t £xg6 45.Sxb7 S e i ,
with a counterattack on the g-pawn, when 4
Black draws. 3
2
42...hxg4 43.E5
1
abcdefgh
51...Sflt? 52.'&e3 S e l f 53.£d3 Sal
54.&c4 &g7 55.g5!
Controlling the black king.

55...f5 56.Se7t £18 57.Se6 Sa4f 58.<&c5

&g7 59.d5!
The h-pawn is not that important.

59.. .5xh4 60.d6 &f7 61.d7 Sh8 62.&d6

1-0
abcdefgh
As said, Black needed to improve the king.
The white pawns are menacing. The solution
is as beautiful as can be.
51...6g7 52.d5
52.g5 &f8 53.d5 Hdl would also be O K
43... >fi8! 44.E6 &g8! 45.Eh2 &h8! 46.E7
for Black. White cannot play 54.&e5 without
Se8 47.Sh4f5
allowing 54...&e7, when Black survives.
Black holds after another few moves.
52...C6!
In the following example, the black rook is
The king would be useless on f8, even
already active, but the king can be improved.
inviting a winning check on e7 in some lines.
Black did not sense this.
But on f7, it is ready to jump into the action.

53.Sd4
53. d6 Hdl is also harmless. Another side-
effect of 52...f6!.
 Chapter 1 3 - The King’s Role 203

53...&f7 54.d6 &e8 Vladimir Malakhov - Mikhail Demidov

Sochi 2015

8
7
6
5
4
3
2
abcdefgh 1
Black holds. abcdefgh

Hie king helping to win

53...&e7?
Passive, giving White a chance to hold.
Which he did, although there was a little
Let’s move on from the positions where the
bump in the road.
king has to keep the passed pawns under
control in a defensive setting and look at
54.&f3 Ea4
positions where the side with the advantage is
looking for a way to make the most out of his 54...5g3t 55.&e4 h3 56.§h6 followed by
&f5 would give White sufficient counterplay
handling of the king.
to secure the draw.
In the first example, Black has an extra pawn
55.Sg6 Sa3t 56.i?g4 Sg3t 57.i?f5 h3
and high hopes of winning with his connected
58.SE6 g4 59.Sh7t 6O.e6 Sgl 61.&g5?
passed pawns. But as we shall talk about in the
6 1 .e7f was the right way to go, 8 with a draw
chapter on Connected Passed Pawns, it is rare
like the game.
that the pawns can be promoted without the
help of both the king and the rook.
8
7
6
5
4
3
2
1
abcdefgh
204 Conceptual Rook Endgames

61...5g2? 55...Sf4t
Black had a surprising win. 55...Hxe5? 56.&g4 would allow the white
king to smother the black pawns and paralyse
After the accurate 61...d?e8! 62.e7 &d7! White Black. With the king out of play, it will not be
is too late to create counterplay. For example, possible for him to make progress.
63. ci?f4 Eg2 and the pawns will cruise to the
bottom rank. 56.&g2 g4 57.Sd6 Se4 58.e6 59.e7t
&xe7 6O.Sh6 g3
62.e7t &e8 63.&F6 3£2f 64.&e6 Se2f 60...h3t?61.&g3 is a draw again.
65.&d6 h2 66.Eh8f &f7 67.Sh7t * e8
68.Sh8f &f7 69.Eh7t &f6 7O.Eh6f &F7 61.&h3
71.Sh7t &f6
Vi—Vi

Black needed to involve the king in the

support of the pawns in order to play for a win.
We shall see this in many examples elsewhere
in this book, for example in the chapter on
Promoting Passed Pawns (page 115), so here
we shall keep it to just this and the next
example.

53.. .6g7!
The points of this move are several.

54.Sa6 abcdefgh
54.&f3 would now lose to 54...Hg3f. A key 61...Sg4!
point is that after 5 5. 4 h3, the white rook A nice trick. If the rook is taken, the g-pawn
cannot get to h6. It may feel like a small way promotes.
for the black king to help, but it is sufficient.
55. f2 Ha3 62.&g2 &£7
Black wins by a series of zugzwangs. The
54...5e4 55.&B
white rook cannot leave the h-file, as ...h3t
would win on the spot.

63.Eh8 &e6 64.Eh5 &f6 65.Sh8 &c5

66.Eh5t i?d4
The king will come to assist the promotion
of the pawns. Black wins.

The final example of this chapter is terribly

complex. I believe that the key ideas from
earlier are clearly visible, but I shall not point
to them at every turn.

abcdefgh
 Chapter 1 3 - The King’s Role 205

Ma Qun - Xu Yinglun 40.gxh5t

As said, 40.a4 h4 41.a5 Bf3! with the idea
China 2022
...Hf6 gives Black enough counterplay.
40...6f5I!
This is the surprising move. We have talked a
lot about endgames being about promoting
pawns. The e-pawn is the cornerstone of the
black defensive action (which is offence), it
does not matter if he is two or three pawns
down.
41.Be7 &f4 42.a4
42.Hxg7 with the idea to give up the rook
is perhaps the sensible choice. White will
be able to set up the triple pawn fortress
discussed on page 99.
abcdefgh 42...gxh5 43. b4 He5 44.Sf7t &e4
White is the one who has to hold the draw.
37.Se4?
He can and he will. But this is still a success
This looks entirely convincing, but offers
for Black.
Black a fantastic chance to save the game.

37.c4! was best, when the a-pawn becomes 39.g?Ji5 &h6

seriously weak. 37... loses to 38. c5 &f7
39. He4. And 37...Sf4 38. c5 Hc4 39. b4 axb3t
40. xb3 Hxc5 41.Hxe3 also wins for White.

37...Sxh3 38.Sxa4

abcdefgh
4O.Se4
4O.Hg4! Hxh5 41.Be4 Hh3 42.a4 would have
won a tempo and decided the game quickly.
The h-pawn does not serve much as a barrier
abcdefgh
for the advance of the g-pawn, as the e-pawn is
38.. .h5? providing all of the counterplay. But at move
38.. .6g6! 39.Se4 h5 40 such finesses are hard to see and even harder
Threatening ...h4, which would give to commit to.
sufficient counterplay. It is hard to deal with
two passed pawns...
206 Conceptual Rook Endgames

40...&g5 41.&b3 $6 42.ge7 &f4 44...&d3 45.Sd7t &e4 46.gxg7 Sxh5

47.&c4 e2 48.ggl Sh8 49.a4 gc8f 50.&b4
&d3 51.a5 Sb8f 52.&a3 Sb5 53.b4 &xc3
54.§clt &d2 55.Sbl &c3 56.gclf &d2
57.Sbl &c3 58.Eclt &d2 59.ghl &c3
6O.gh3t &d4 61.Sh4t &d3 62.Sh3t &d4
63.Sh4t &d3 64.Sh3t &d4 65.gh4f
Vi-Vi

As we are progressing in this book, the examples

are becoming more and more complex. I hope
that at the same time you will see that the ideas
remain simple and applicable. The king’s role is
often to make sure that the rook can be active.
abcdefgh This means that the king will defend against
dangerous passed pawns, often by blocking
43.Sf7t?
43.&c4! would have won the game. White them, or helping the pawns advancing.
wins quickly after 43... &f3 44. &d3 and Generally, the king is useful when the fight
43...Sxh5 44.£d4. And after 43...Shl 44.b4 is local, while the rook is strongest when it is
&f3 45.a4 e2 46. a5 e l = ® 47.Hxel Hxel 48. b5 working from afar (Checking Distance, page
the pawns are running away... 1 63) and when it can shift from one side to the
This was by no means easy to calculate, but other and thus often take split responsibilities,
the risks White is taking are small, so it should time permitting. Here we looked mainly at the
have been tried, even if Ma Qun did not feel in defensive roles the king can play. In the next
complete control. chapter, we will focus on the moments when
we need to activate the king.
43...&e4 44.Se7t
White is not really trying to win, but also
seems to have problems letting go of hope. The
game continued for 20 needless moves.

44.£c4 Hxh5 45.Se7t &f3 46.&d4 e2 White

has to hurry up forcing a draw with 47.gf7t,
as he would lose after 47.a4? S h i .
Chapter 14

King Activity
Following up on a shadowy theme from the previous chapter, we shall now look at situations
where the rook is already active and the chances can only be improved by activating the king. This
may seem basic - and I want it to! Because the implementation of basic ideas in an exceedingly
complex setting is what we are doing here.

The first example is glorious. Because it is funny.

J. Deepan Chakkravarthy -
Balasubramani Sekar

Kolkata 2018

abcdefgh
45.Sb2?
A decent-looking move, similar to many situations we have seen in this book.

45.Hb3! was stronger. After 45...d3 46.&b6 White is winning. The key point is that the d-pawn
is eliminated. After 46.. .d2 47.Hb2, Black will have to move the rook and White simply wins.

45. ..Sal 46.&b6 d3 47.a7

1-0
208 Conceptual Rook Endgames

Black resigned, which is easy to understand. The game ended in a draw quickly.
He is about to lose his rook and had been
lost for a while, so the fact that the position 50.<&dl? Sc2 51.a6 Sxb2 52.8b7 Sa2 53.a7
is a draw is entirely surprising. If he had been Salt 54.&d2 b2
forced to play on, he would have made the Vi— Vi
draw easily:
White would have won easily after 50. d2!
47.. .6ft>! 48.£b7 Hc2f 51. <&d3 Hxb2 52.&c3, when Black
48.Sd2 is met with continuous checks. If the cannot both keep an eye on the a-pawn and
white king hides on a8, White will even lose. defend the b-pawn. The main loser is of course
the black king, which is entirely cut off.
48.. .6e5 49.a8=® Hxa8 50.1$?xa8 d4
In the next example, White goes for a pawn
8 advantage, but this was not enough to win.
7
Alexander Ipatov - Levente Vajda
6
Golden Sands 2012
5
4
3
2
1
abcdefgh
The black king arrives just in time to make
a draw in the pawn ending. It helps that the
white king is as far away as can be.

Petro Golubka - Witold Klepek

abcdefgh
Krakow 2020
67.Sxb6 a3!
Creating a passed pawn, which is what will
hold the game.

68.Sb5 Sb2 69.Sxc5 Sxb3 7O.Sa5 &d7

71.i?g7 Sg3t 72.£f7 &c6
This position is a draw, though several
mistakes were made before this was the result
on move 109.

We have discussed all the aspects of the winning

line earlier in this book. We need to make the

abcdefgh
 Chapter 14 - King Activity 209

most out of our most important asset, which is Let’s see a high-level example.
the passed pawn. The rook belongs behind the
passer, which also means we want to make it Levon Aronian - Maxim Matlakov
hard for the opponent’s rook to get there. And
Tbilisi (3.5) 2017
we want to cut off the opponent’s king, so it
cannot help with the defence, as it would be
the best piece to do so.

67.bxa4! Sxa4
67...&d7 68. &f7 Sxa4 69.Se4 wins without
too much trouble. White can always play Se7t
at the right moment.

abcdefgh
Matlakov lost because he was trying to make
the king do all the work.

50...&g3? 51.Sb6 &x£3 52.Sxc6 g5 53.Sg6

1-0

abcdefgh Black could have supported the king by

making something out of the rook: letting it
68.<&e5I!
take the defensive role...
The key point. Black cannot take on c4, as
the f-pawn would sail straight to the end. And
5O...Sd7I!
at the same time there is no time for ...Sal,
5O...Sd8? would not work. After 51.Sb7,
as White would take the b-pawn under much
White wins.
better circumstances.
51.Sb6 Sc7
68.Se4 does not work now, because of
Black is fine. He will win the h-pawn and
68.. .b51.
the king will provide counterplay.

68.. .<£d7 69.f6 Sal

The key point of this chapter is that it is our
Now forced. White is totally in control and
instinct to focus on the possibilities of the rook.
can win in a lot of ways. The simplest would be
And we should! But when they are limited, it is
to win a second pawn.
important to look at the king as well. Or to say
it in a different way: we should have a plan for
7O.Hxb6 Hfl 71.Sd6| &e8 72.&d5 3f5t
all of our pieces.
73.&e6 Sfl 74.Sc6
Black can only resign.
210 Conceptual Rook Endgames

Sergei Lobanov - Nikita Meshkovs 30.<&£2! b4l?

This idea still feels critical.
Riga 2019

30.. .a5 is met with the powerful 31.Sdl!,

when the king is still aiming for d4, while
31...b4?! loses on the spot to 32.cxb4 axb4
33.Sd8f &h7 34.Sf8, where Black would
be doomed to passive defence, and 34...&g6
35.g4 would lead to a straight descent into the
abyss.

31.<&e3 bxc3 32.<&d4 Sb5 33.bxc3 Sb2

34. c4!
34...Sxg2 35.c5 Sxh2 If Black does not
remove this pawn, he will not have created a
passed pawn, but White can win slowly with
abcdefgh
36.Scl, exploiting the c-pawn.
3O.Sal?
With this natural move, White probably 35.g3 Sxh2 36.c5 &e7
expected to tie Black down. But Meshkovs
saves the game with active play, which is more
or less the only way to do it when you are a
pawn down. (More about this discussion in
Chapter 22, page 315).

30.. .b4! 31.cxb4 Sc4 32.f5

Trying to obtain winning chances with four
against three, but it leads nowhere.

32.Ea4 Sxf4 33. b3 Ed4 34.&f2 Ed3 Black

also holds, as the rook is beautifully active.

32.. .5xb4 33.Sxa6 Sxb2 34.Sa8f &h7 abcdefgh

35.g4 Se2 36.Se8 h5 37.gxh5 Se4 38.Se7
37.Sbl!
Sg4f 39.&hl &g8 4O.h6 gxh6 41.Se8f
White is winning. The focus on activating
<&h7 42.Se7 &g8 43.fb h5 44.h3 5f4
the rook should be consistent in our thinking,
45.&g2 h4
and here it happens last, but still it was always
The white king is cut off.
on the cards as the way to win the game. It was
just that, in order to make the most out of the
46.Sa7 SB 47.Sa4
advantage, White needed to activate the king
Vi-Vi
first. Play could continue:
The winning idea is thematically relevant, as
37...*&e6
well as prophylactic and beautiful.
37 ...§c2 38. f5 is the simplest.
 Chapter 14 - King Activity 211

38.Sb6f &f5 39.c6 £g4 4O.£d5 Sc2 57...£f6 58.£d2 £e6 59.£e3 £d6 6O.£d4
41.£d6 £xg3 42.c7 £c6 61.Sbl
And it is all over.

White missed a win in the next example.

Robin van Kampen - Sergei Tiviakov

Canada 2014

abcdefgh
61...16!
The only move. But it is there and it holds.
A key point is that ...g5 will lead to a pawn
ending where White only has the h-pawn and
will not have enough tempos to keep the black
abcdefgh king away.

Because the white king is so passive, it is

62.Sfl Sd8f 63.£c4 £xb7 64.Sxf6 Sg8
difficult to have the b-pawn and the rook do
65.&d5 g5
all the work. Something extra is needed.
Black holds.

56.b7? 66.§h6 gxh4 67.gxh4 £c7 68.£e6 £d8

56.§b5? is also poor. The issue is that Black 69.Sxh5 £e8 7O.£f6 Sffif 71.£g7 Sf7t
will also be able to bring the king in after:
72.£g6 £ffi 73.Sh8f £e7 74.h5 Sf6f
56...£f6 57.£d2 £e6 5 8 . £ e 3 (58.b7 Eb8
75.£g5 £(7 76.Sh7t £g8
would allow Black to bring the king over to
Vi-Vi
eliminate the b-pawn.) There are a few draws
available here. The thematically consistent The winning path was based on defending
would be: 58...£d6f 59.£f4 £c6 60.§b3 the b-pawn and keeping the black king out of
£b7 Black is fine. action for long enough for the white king to
enter the game.
56...Sb8 57.Sb5
Trying to keep the black king a little under
control, but it is already too late.

After 57.Sc7, Black will hold with 57... £16

5 8 . £ d 2 £f5! 5 9 . £ e 3 £g4 6O.£f2 f6, when
White is tied down on the kingside.
212 Conceptual Rook Endgames

56.Sc6! Swapnil Dhopade - Abhijeet Gupta

New Delhi 2019

abcdefgh
56...Se4 abcdefgh
Trying to put the rook behind the b-pawn,
79.Sg7?
but it costs time.
79.Sxd3? is entirely hopeless and easy to
discard. The pawn ending after 79...e4f wins
The black king could be kept under control
easily for Black. The king steps forward and
after: 56..T6 57.b7 Eb8 58.Ec7t &h6 59.&d2
follows up with ...f4, winning.
g5 60.&e3 &g6 61.&e4 The black king is
under control. White will win in various ways,
79...f4 8O.Sxg6t <&d5 81.Sh6 Sxe3t
depending on what Black tries.
82.<&£2 Sg3 83.Sxh5 e4 84.Sh8 e3t 85.<&e2
<&e4 86.Se8t 6 87.Sffif &g4 88.g6
57.&c2 Sb4 58.<&c3 S b l 59.<&d4 ft 60.<&c5
Sg2f 89.<&d3 e2 90.<&d2 91.Ee8 Sxg6
g5 61.&d6 &g6 62.&c7 <&f5 63.b7 Sxb7t
92.Sxe2 Sh6??
64.&xb7 &g4 65.Sxft gxh4 66.gxh4 &xh4
A horrible blunder in never-ending time
67.Sg6!
trouble that allows White to escape.
White wins.

The simplest of several winning moves was

The following example is from a game between
92... <&g3! followed by advancing the pawns.
two talented Indian GMs. White is clearly
under pressure, but a route to salvation exists.lt
93.<&el &g3 94.<&fl
involves active play, using the king to keep an
The game was drawn on move 144.
eye on the opponent’s pawns from a surprising
location, as well as using it to keep the black
king out of the game. As so often, the rook The drawing idea is based on a lot of things we
and the pawns are slow, giving White enough have already seen: checks, checking distance,
time to create counterplay. But first, let’s see a keeping the rook active and shouldering.
failed strategy, where the black king is allowed
to get active. 79.Sa7!I f4
Obviously the critical line.
 Chapter 14 - King Activity 213

Black could keep the game going with In the next example, from the Soviet
79.. 8O.Ea6t & d 5 81.Sa5t Ec5, but Championship, Black lost in perhaps the most
White should still hold with decent play. For illustrative way, but also without any serious
example: 82.§a6!? resistance.

8O.Sa6f! &d7 Lev Polugaevsky - Evgeni Vasiukov

The pawn endings would hold this time.
Tbilisi 1967

81.&e4!!
The key idea. The king has to get active, in
order to slow Black down.

It is not possible to check the black king to

the other side of the board and hold. After
81.Sa7t? &c6 82.Ea6t & b 5 83.Exg6 Exe3t
Black will bring the king to d3/e4 quickly and
put the rook on g3. The important idea is to
cut off the black king on the 6th rank.

81...3xe3t 82.<&d5 B 83.3xg6

The idea is not to check the black king to the abcdefgh
8th rank, but just to keep it out of the game.
45.. .6h7? 46.b7 &g7 47.&e3 e4 48.&F4
&h7 49.*&e5 &g7 50.<&d5
Threatening 51.&C6!.

50.. .5b2 51.&xe4 Sb4t 52.<&d3 Sb3t

53.&c4 Sbl 54.f4 Self 55.&d3 Bbl 56.f5
Sb6 57.ft>t
1-0

Passive play is entirely hopeless.

45...h5?
White would often keep the pawn on b6, to
abcdefgh give the king a place to hide from checks on
b7, but in this case, there is no need.
83...e4 84.H66 &e7
After 84...§e2 85.g6 Black will have to
46.b7 &h7 47.*&e3 &g7 48.&d3 &h7
abandon his pawns to stop the g-pawn.
49.&C3
The black rook cannot keep the white king
85.&d4 Se2 86.Sf4 £2 87.g6
cut off on the 4th rank.
Black has to abandon all hopes of anything
but a draw.
49.. .5bl 50.&c4 e4
214 Conceptual Rook Endgames

Black cannot afford to lose the e-pawn, or he 47.. . <&h5! 48.<&e3

would lose as in the game. White also cannot make progress after:
And after 5O...*i?g7 51.<&d5 S b 5 t 52.&C6 48.&g3 Sb3t 49.8 Sb2
White wins.
48.. .6E4 49.&d3 &xh3
51.&d5 Sb4 52.<&c6
The simplest.

52.. .5c4f 53.&b5 Sc2 54.Sf8

White wins.

Trying to activate the king is the only chance,

but it should be done with some attention to
detail.

45.. .6g6? 46.b7 &h5

This looks natural, but after:

47.g4f! hxg3 48.ftg3 Sb2 49.g4f &h4

abcdefgh
5O.g5!
White wins. Despite being limited to one file; the black
king is making headways.
The way to activate the king is based on
prophylaxis. Black has to prevent this g4f idea. 50.&c3 Sbl 51.&c4h5
51...e4? would be a serious mistake. White
45. ..h3!! 46.gxh3 &g6 wins after beautiful triangulation: 52.< ?d4
Sb4t 53.£c5! §bl 54.<±>d5 Sb4 55.&c6 and
8 White wins.

7 52.&d5 h4 53.&xe5
6
5
4
3
2
1
abcdefgh
47.b7
Black also draws after: 47. eS » ?h5 48. dS
<±>h4 49.&c3 § b l 50.&c4 &xh3 51.&d5 h5
52.&xe5 h4 53.f4 &g3 54.b7 S b 5 t 55.£e4 abcdefgh
Sb4t
 Chapter 14 - King Activity 215

53...5b5t! 48. h4!! Exf2!?

The key move. White cannot be allowed the Black has to try something. The king can
keep the f-pawn. escape after 48...g4 easily.
And the variation after 48...gxh4 49.Sd8
54.&e6 S d l 5O.d7 &g7 51T4 &h7 52.f5 &g7
Or 54.&d6 &g2 55.F4 <±>f3 56.F5 Sxf5! 53.f6t £f7 54.Sh8 Exd7 55.§h7t &e6
57.Sh8 S b 5 58.&c6 Sxb7 with a draw. 56.§xd7 &xd7 57.&h3 appears to be quite
straightforward. White wins.
54.. .<£g2 55.f4 h3 56.3g8f &£3 57.b8=® 49. hxg5t &xg5 50.§d8! §f7
Hxb8 58.Sxb8 &xf4 50...§d2 51.d7 seems to go straight to the
Black holds. pawn ending, but 51... h4 would be a
stalemate trick, hoping for 52.§g8?? §xg2f!.
In the next example, a former World Champion However, after 52.& g l White will soon have
was in desperate trouble. a g2-g4 tactic to decide the game.

Tamaz Giorgadze - Mihail Tai

Tbilisi 1978

abcdefgh
51.Se8!
Cutting off the black king and threatening
Ee2-d2.
51...Sd7 52.Se5t! £f4
52...&f6 53.Sxh5! wins.
abcdefgh
53.Sd5 &g4 54.Ed4f &f5 55.<±>h3 &e5
Giorgadze was much better early on and had 56.Ed2
plenty of wins, giving his opponent only one White wins. The rook is a poor blockader.
chance to escape in the early middlegame. But
with the transition into the endgame, Black’s 48.§d8 h4! transposes.
drawing chances increased all the time, until
finally he was within drawing margins, where 48.. .3d3t 49.8 h4f 5O.&£2
we come in. 50.&g4 &g6! and White has to be ever so
slightly careful.
47...Sd2? 48.&g3?
Missing his chance to beat the hero. 50.. .6e6 51.Sg7 &f5 52.Sd7 &e6 53.&e2
gd5 54.Sd8 Sa5 55.&£2 Sa2f 56.&gl
Sd2 57.Se8f &f5 58.Sd8 &e6 59.d7 Sxd7
216 Conceptual Rook Endgames

6O.Sxd7 &xd7 61.&12 &e6 62.&e3 &e5 49... e6!

63.&f2 64.&e2 <&e6 49...§d2f? would lose an important tempo.
Vi-Vi It does not matter if the king is on the 1st
rank or second, but it does matter if it is on
For this reason, Black should have restricted the h-file or g-file. 50.& g l &e6 51.Sd8!
the white king immediately. Sxd6 Otherwise White will eventually swap
the pawns and have two connected passers.
47...h4! 52.Exd6t &xd6 53.&f2 £e6 54.&e2! White
47...&e6 followed by 48. ..h4 also worked. wins. Classic opposition.

48. g3 hxg3t 49.fcg3 5O.Sd8

49.&xg3 is visually a draw. White will not be 50.§g7 &f6 goes nowhere.
able to make any progress beyond exchanging
the d-pawn for the g-pawn, ending up in a 5O...Sxd6! 51.Sxd6f &xd6 52.&g2 &e5
theoretical draw. 53.<&£2 <&f5
53... &f6 with opposition also works.

54.&B g4f!

abcdefgh

abcdefgh
With a draw.
Chapter 15

Checkmate
Checkmate is the only way to win a game of chess against a non-cooperative opponent. Therefore,
we all play for mate, all the time, when we try to win a game of chess. The difference is if we take
the direct path, or the scenic route. In this chapter, we shall look at situations where a shortcut
to checkmate appears.

Checkmate is not a big topic in rook endings, although it does come up from time to time. I do
not have any specific deep ideas about this theme, but I do have a few nice positions I want to
show.

I. Krikheli
Suomen Shakki 1985

abcdefgh
This study does not contain surprising moves or deep ideas, but it does have a nice, although
a bit orthodox, finish.

I.b3!
I.§d2? could sort of make sense if you are focusing too much on keeping the black king
dominated. But endgames are about pushing and promoting pawns, so the move is rather odd,
besides just bad.
218 Conceptual Rook Endgames

l...<&g2 2.<&e5 <&£2 3.<&d4 <&e2 4.Sh3 <&d2 The next example sees White dominate the
4...Bd8f 5.&c3 would allow the king to opponent with a combination of mating
come around the queenside way and help threats and the creation of a passed pawn.
advance the pawn.
Kirill Alekseenko - Bilel Bellahcene
5.Eh2f £cl
Warsaw (rapid Wch) 202 1
5...&dl 6.&c3 Ec8f 7.&b2 and the rook
and king together will push the pawn forward.
The black king will remain cut off.

abcdefgh
45.Sf4! Eh5
abcdefgh After 45...&xh2 46.&f2 Bh5 47.Eh4f, the
h-pawn runs away.
Black has managed to prevent the white
king from hiding behind the pawn and thus 46.<&£2 a5
protecting it. He is ready to harass the king If 46...d3 47.Sh4t Exh4 48.gxh4 £xh4
endlessly, without allowing the advance of the 49. b3! the pawn ending wins. (However,
pawn. But White has a simple trick to win. 49.&e3? c4 50. b3 cxb3 leads to a long variation
where Black draws by a tempo.)
6.b4!
The pawn will advance and the black king 47.b3
is cut off. Controlling the black pawns. White now
won a game with more only moves to win than
6...Exb4f he had needed to. But when you play them all,
Is thus the critical line, but it loses there is nothing to complain about.
immediately.
47...Eh8 48.&B Sh7 49.a4 Bh6 50.EB
7.&c3 <&xh2 51.Exc5 Ee6 52.Bh5t &gl 53.Ed5
Not a new trick, but still a checkmate trick. Se3t 54.<&f4 <&£2 55.Exd4 B B f 56.<&e4
Exb3 57.g4 &g3 58.<&f5 <&h4 59.g5t
&h5 6O.Ed5 SBf 61.&e6 &g6 62.Bxa5
Bf4 63.&d6 Ec4 64.&d5 Bf4 65.&c6 Bb4
66.&c5 E b l 67.Eb5 B al 68.a5
1-0
 Chapter 1 5 - Checkmate 219

The following position arrived in my analysis Sergey Fedorchuk - Sophie Milliet

and it became a nice little prophylaxis exercise.
Paris 202 1
Once you understand White’s idea, you are
able to choose the right pawn.

Vladimir Kramnik - Alexander Grischuk

Stavanger (analysis) 2014

abcdefgh
49.&d6? Sg2! 50. ?e5
5O.Sg6 was a bit more dangerous, but Black
can get counterplay with 50...c5! 51.&xc5
Sxa2 and Black has enough time to get the
abcdefgh pawns going, as the white king is a bit offside.
39...b5!
If you choose the a-pawn, mate is inevitable. 5O...Sxg4
The winning chances are gone, as the king is
40.&h3 b4 41.&h4 b3 42.&h5 b2 blocking the passed pawn.
The point. If Black is allowed to queen the
pawn, h7 will be defended. 51.Sd6 &e7 52.ft>t &e8 53.Se6f &f8
54.Se7 Sg2 55.Sxb7 Sxa2 56.Sc7 Se2f
43.Sgl b l = ® 57.&d6 Sd2f 58.&e6 Se2f 59.&d6 Sd2f
Black has to defend against the renewed 60.&e6 Se2t
threat of mate. White now has to force a draw. Vi— ¥2

44.Sg8f! Sxg8 45.&g8=®t Sxg8 46.Sxbl White would have won the game by making
Black is a pawn up, but White is active and the most of the connected pawns.
will make the draw.
49.g5! Sxa2
The next position is completely winning, This feels principled, but loses in a number
but White ruined his chances using the wrong of ways.
strategy.
5O.Sh7t
50.&xb7? would be a distraction. After
5O...Sa5! Black holds.
220 Conceptual Rook Endgames

But also after 50.&d7 Sd2f 51.Sd6 Sxd6f Adriana Nikolova - Iva Videnova-Kuljasevic
52.&xd6 a5 53.g6t &f6 54.g7! &xg7 55.&e7
Sofia 2023
White wins.

And 50.f6 Eg2 51.Eh7t &f8 52.&d6 Exg5

53. &e6 and 54.Sxb7 would also win. Again,
the threat of mate is important.

50...&g8 51.g6 S£2 52.SI7 a5 53.16 a4

54.&d8 a3

abcdefgh
4O...d5? 41.Sb6f! &f7 42.Sb7t &e8
43.Sxb5??
Blundering the rook. Instead 43.Sb6! would
have held the game.

43...Sf5t
In view of 44.&e3 d4f!, White resigned.
abcdefgh
0-1
55.&e8!
Threatening 56.Sf8#. The winning idea was:
40...&d5!!41.Sxb5t &d4
55...Se2f 56.Se7 S£2 57.f7t &h8 58.f8=®t
SxfBt 59.&xf8 a2 6O.Sh7#

The next game was only lost after a massive

mistake by White.

abcdefgh
The pawn ending is unavailable for White,
and after 42.Sa5 d5! Black is winning. But the
key point is of course: 42.§b8 Sf5#
 Chapter 1 5 - Checkmate 221

Zeng Chongsheng - Zhao Jun 59.. .Bg3 6O.Bxc3

60.&c2 Bgl and the pawn queens.
Xinghua 2015

60.. .Bxc3 61.f5 Bg3

0-1

A pity. 61...&c4 62.g7 &d3 63.g8=® B e l #

would have been a thematic finish.

Black’s winning plan relies on checkmate in

the main line.

55.. .6c5! 56.Bbl

White has to go back.

abcdefgh 56.Bc7t &b6 (or 56...&d6) 57.Bc8 Bg3

would queen the d-pawn.
55. ..&c4?
55...&a4? happened later in the game. 56...Be3 57.&c2 Bel 58.Bdl <£>d4!

56. Bc7t &b3 57.Bb7t

White had a few ways to make a draw, but
none of them were simple.

For example, 57.g7 Bg3 58.Bb7t &a4 59.&c2

Bgl 6O.Ba7t! & b 5 61.Bd7! and Black will not
make substantial progress.

57.. .6a4 58.Ba7t?

58.Bbl! Be3 (58...Bg3 59.&c2) 59.&c2 Bel
6O.Bdl Black has to look for a way to make
the draw: 60...&a3 61.g7 Bxdl 62.d?xdl &b2
With a hoover draw on the near horizon. abcdefgh
59.g7 Bxdl 60.&xdl &d3 61.g8=® c2#
58.. .6b5?
0-1
Black would have won with: 58...&b4
59.Bal Bg3 60.&c2 Bxg6
A young Magnus Carlsen found a sensational
mating pattern in the next game.
59.Bc7?
59.Bal and 60.&c2 would have made the
draw once again.
222 Conceptual Rook Endgames

Magnus Carlsen - Zbynek Hracek

Germany 2007

abcdefgh
65.&c6! &e5 66.Se8f &f4 67.&d5 f6
68.2ft
abcdefgh 1-0
59.&d6
In the next example, Black was winning right
Forced.
before the time control, but failed to find the
way.
Black would be lost in the pawn ending after
59..T6 6O.He8!.
Husain Aziz - Mariya Muzychuk

60. f4!I exf4 Gibraltar 2022

Black has to take twice. After 6O...gxf4
61.gxf4 Sd2f 62.&c5 Kb2, White has
63.fke5t &g7 64.e6! fxe6 65.&d6, winning.

61.gxf4 gxf4 62,Hg8!

This is the key idea, without which White
does not win. With it, he does.

62„.Sb6t
62...Sxb7? 63. e5# is easy.

63.&c7 Sxb7t 64.&xb7 13

This is the only way to offer any resistance,
but White can continue using mating threats abcdefgh
to kill the counterplay. 35...Scl?
Also after 35...h5? 36.Ed7 h4 37.Ef7t,
White would be out of the woods.

36.2d4f &g5 37.2d5t &f4

37...&h4?? is worse now. 38.g3#
 Chapter 1 5 - Checkmate 223

38.Sd4t $5 39.&g3 Bc2 4O.Bf4t &e5 38.Bxh6 Bc2! 39.a4 b4! 4O.Bh5 Bc3
41.Bxa5 g3f! 42.£h3 Bel
White is checkmated. 38.Bd5 b4 39.Bxa5
g3t is similar.

For a change, we shall start in the middlegame,

where Fier played a compelling combination.

Dmitriy Kushko - Alexandr Fier

Reykjavik 2023

abcdefgh
Luckily for Black, after the time control,
fortune swung back in her favour.

41.Bxg4?
41.Ba4 with the idea 41...a6 42.Bb4 b5
43. a4 and 41.Bf7!? would both have held.
Black is not able to win on the kingside.

41...Sxa2 42.Sg7 b5 43.&h4 &d4 44.&h5 abcdefgh

±c3 45.Sg3t &b4 4O...jLd2! 41.Bxe8!
In view of 46. <&xh6 $2a3 47.Sg7 xbS, This queen sacrifice is a strong practical
White resigned. chance that worked out well in the game.
0-1
41.®xd2 Bxe5 would have lost in short order.
The winning idea had all to do with luring the
white rook to the 5th rank, where it would be 41...jLxc3 42.Ble7t &h6 43.Bh8f &g5
unable to harass the black king from the back. 44.Bg8t Sg6 45.Be6!

35.. .a6!
35...b6! 36.Bd7 a5 37.Bd6 b5 transposes.

36.Bd7
36.Bd4f ?! <±>g5! 37.Bd5t &h4 would bring
White under a permanent threat of checkmate.

36.. .b5 37 .Bd6 a5

Black is threatening to win the a-pawn with
...Bc2. The key point is that after:
224 Conceptual Rook Endgames

45...®xd5t!
Correctly returning the queen.

46.cxd5 Sxg8 47.bxc3 gxh2 48.&xh2 &f4!

The king often belongs in front of the pawn,
as we know. Here the white king is cut off.

49.d6
49.Kxb6 &e3! would have won quickly for
Black.

49...6B 5O.d7 Sd8 51.Se7 &£2!

The mating net is established. abcdefgh
It was now the final chance for Black to
52.Sh7 make the draw.

57...&xc3?
The drawing line goes: 57...f2! 58.Sf7 &e2
59.&e7

abcdefgh
52...&e3?
Just when it was the time to shine, Fier loses
his way.
abcdefgh
52...b5H wins without too much hoopla. 59...b51! Creating much-needed counterplay.
Black creates a passed pawn on the queenside 6O.axb5 a4 61.b6 a3 62.b7 a2 63.d8=® Hxd8
and White cannot stop it, as his rook is busy 64.&xd8 a l = ® 65.b8=® f l = ® 66.ExfI ®xfl
defending against checkmate on the h-file. and the queen ending holds.

53.&g3 £4f 54.&g4! Sg8f?! 58.Se8 Sg6f 59.&d5?

Fier has clearly lost his grip. Giving Black a random chance.

54...6xd3 would make an effortless draw. 59.&e5! was correct: 59...Sg5t (59...f2 6O.Sf8
wins.) 60.&e4 f2 61.d8=®! (But now 61.Sf8?
55. 6! f3 56.Se7t &d2 57.&e6 does not work on account of 61...Sg4f
 Chapter 1 5 - Checkmate 225

and the rook comes behind the d7-pawn.) 67. &b5! to bring the king back to deal with
61...fl=® 62.®xg5! and wins after a lot of the c-pawn would have won after a few semi-
extra moves. accurate moves.

67...Sb3t 68.&xa5 c3 69. &a6 &b2 70.® d4

£bl
A miracle escape in a tumultuous game.
Vi-Mz

The next example is highly elaborate and I do

not want to pretend I have it all under control.

White has managed not to lose a pawn in the

short term, but his long-term prospects are dire.
Inarkiev did not defend in the most energetic
way (31 .a4!), but instead decided that winning
abcdefgh the g6-pawn was important. Remembering that
endgames are almost always about queening
59...Sg5t?
The chance was: 59...f2 6O.Sf8 Sg3! 61.2xf2 pawns, it is hard to see why attacking the
Sxd3t 62.&c6 &b4 Black would lose his rook, g6-pawn was what Inarkiev found important.
but easily make the draw. I am being cheeky. This is a rapid game where
White has already been defending for quite
60.&c6 Sf5 61.S e i ! Sf6f 62.£c7 Sf7 a while and had to make plenty of difficult
63.Sfl &xd3 64.Sxf3t Sxf3 65.d8=®t decisions, so it is fair to presume he was short
sl?c2 66.&xb6 c4 of time by now and making moves with the
hand more than after reflection...

Ernesto Inarkiev - Boris Gelfand

Magas 2016

abcdefgh
67.®d5?
White stumbles on the finishing line.

abcdefgh
226 Conceptual Rook Endgames

31.Sg7? 35...Sf5 36.g4 hxg4 37.Sxg4f

31 .a4! would have fitted perfectly with the
principles we have talked about up till now
and is not surprisingly the drawing move. The
a-pawn will eventually get exchanged for the
d-pawn, with a drawable position. 31...Sf4
32.&g3 Sg4f 33. &h3 Sxd4 34.Ea8 &f5
35.a5 Ka4 36. a6 d4 37. a7 and there is no way
to avoid the exchange of pawns.

31.. .6f5 32.Sd7 &e4 33.Sg7

After winning a few tempos with the king,
Black is happy to exchange the useless g-pawn
for the d4-pawn, creating a passed pawn.
abcdefgh
33.. .6xd4 34.Sxg6 37.. .6e3?
As Sam Shankland described in Small Steps
to Giant Improvement, the king is often best
placed in front of the passed pawn, as it is the
best place to fight for all the squares in front of
the pawn. It is a sort of shouldering, at times,
but can also be basic flexibility, aimed at being
able to support the pawn from either adjacent
file. 37...&d3! 38.Eg3t &c4 39.a4 d4 Black
wins.

38.Sg3t &e2 39.Sg2f &d3?I

Black could have repeated with 39. .. e3!
4O.Kg3t &e4 41.Sg4f and now found
abcdefgh 41...6d3!, which is unlikely to happen in a
The most important position of this rapid game.
endgame. We will give it more attention
below. Gelfand played a natural move, but 4O.Sh2?
gave White a chance to create counterplay. Under a lot of pressure, Inarkiev misses the
last exit o n the path to defeat. After 4O.Sf2!
34.. .'&e4? 35.&gl? Sh5 41.Kxf6 Sxh4 42.Sd6 White holds.
White needs counterplay and could again
have made the draw with 35.a4!, based on
35.. .d4 36.a5 Ea2 37.Sxf6 Hxa5 38.&g3 d3
39. &£2 Ka2f 4O. e l ! Kxg2 41.Kh6 with
enough counterplay. For example, 41...&e3
42.Se6t &f4 43.Kd6 with a draw.
 Chapter 1 5 - Checkmate 227

35.a4 d4 36.a5 d3 37.a6 gf4!

The rook is needed behind the pawns.

37...Sa2? 38.Sxf6 would allow the white rook

to fight against the d-pawn from behind,
making the draw.

38.a7

abcdefgh
4O...Sh5! 41.&E2
After 41.a4 d4 42.Ka2 &c3 43. a5 d3 we can
see White is too late.

41...d4 42.&B &c3 43.&g4 Sd5 44.SE2

f5t 45. &g5 d3 46.h5 d2 47.Sfl &c2 48.h6
d l = ® 49.Sxdl £xdl 5O.h7 f4f 51.&g4
Hd8
0-1
abcdefgh
38...Sa4!!
Black has quite a number of ways to get to the The difficult move. Black attacks the
same point, and probably a few others ways a7-pawn immediately, ignoring the free (and
to win. As far as I can see, the following is the irrelevant) pawn on h4.
direct path and the one humans (like Sam
Shankland) will find. 38...Sxh4f 39. &g3 Ka4 4O.Kg7 would lead to
a draw. Black does not have a way to push the
34...&e3! pawn to the back rank and nowhere to hide
the king.

39.Sg3t
39.Kxf6 Kxa7 looks bad and does lose.
Black is threatening ...Ha4, dominating the
opponent entirely. 4O.Ke6f &£2 41.Kd6 <&e2
42.Se6f & d l 43.Sd6 d2 44.g4!? So, this has
to be tried, but it does not work. 44...Sa4!
is now simplest, although taking on g4 also
works.

39...&e2 4O.Sg7 d2 41.§e7t

abcdefgh
228 Conceptual Rook Endgames

41..J&£2!!
The reason Black did not take on h4 earlier.

42.Sd7 Exh4#

abcdefgh
Chapter 16

Stalemate
While promotion of pawns is the key aim of most endgames, there are some other elements we
strongly associate with the endgame. The active role of the king is an obvious one, but fortresses,
zugzwang and stalemate also stand out. Fortresses are quite rare in rook endings, while zugzwang
is a big theme and stalemate is a medium-sized theme. A big part of the stalemates is based
around one common pattern, while the rest are divided into two other themes, “random” and
perpetual rook. Let’s start with a “random” stalemate.

The 2020 and 2021 British Championships were played online, with a classical time control.
Peter Finn played remarkably in both tournaments and beat several GMs, despite being rated
below 2000.

In this game Mark Hebden had a chance to escape with a trick.

Mark Hebden - Peter Finn

Internet 2020

abcdefgh
56.Sb8?
56.f6! gxf6 57.Sxb2! would have drawn immediately, due to stalemate. Black can give the
check on c3 and play some moves, but because of the g4-square being covered, Whites draw
should be close to effortless.
230 Conceptual Rook Endgames

56...±xf5
White is done. The last chance would have
been a waiting move, hoping to get a second
chance of Sxb2! after 57...&e4.

57.&B ±e5 58.Sb7 ±d4 59.±f4 Sg2

60.&B Sd2 61.±f4 ±c3 62.±g5 Sd5t
63.&g6 Sd4 64.§xb2 d?xb2 65. xh5 Sd6
66.d?g5 c3
0-1
abcdefgh
At the end of the chapter there will be some is stalemate.
curious and unique stalemate scenarios, similar
to the one above. But the key stalemate scenario 80...&e5 81.Se7t 82.Sa7 Sc6 83.&g3
we have to add to our image vocabulary Sc3t 84.<&h4 Sc6 85.Sb7 Sa6 86.&g3
consists of the king standing right in front of Sa3t
an enemy pawn, with the king cutting off the Mb-1/!
line behind. Lets see a few examples to really
rub it in. The following Armageddon game decided the
2018 French Championship. Black missed the
Rob Perez - Vincent Tsay chance to draw, which in this case would have
New York 2018 won the tournament.

Tigran Gharamian - Romain Edouard

Nimes 2018

abcdefgh
78...Sb4f 79.±g3 Eb6I! 8O.±h4
White cannot win, as 8O.Sxh6f &g5
81.Sxb6 abcdefgh
59...±b4? 6O.Sb8f ±a5 61.Sb5t ±a6
62.§xc5 Sxh6 63.Sg5 b6 64.b4
1-0
 Chapter 1 6 — Stalemate 231

59...Sh2f! 60. &a3 (6O.£bl <±>b4 61.h7 &c3 8O...Sa8 81.Sh5

would also lead to a draw.) 6O...Sxh6 and
Black would have eliminated the dangerous
8
passed pawn. White can try to push a bit, but
even after 61.Sb8 §b6, Black is fine. 7
6
The following example is a bit more elaborate.
5
Javokhir Sindarov - Marat Dzhumaev 4
Uzbekistan 2021 3
2
1
abcdefgh
81...6b6?
The stalemate trick is the same as always.
81.. .5xa6! and so on.

82.±b4!
White wins.

82...Hg8 83.Hh6f &a7 84.&b5 Sb8f

abcdefgh 85.&a5 Sbl 86.Sh7t &a8 87.c5 &b8
88.Sd7 Saif 89.&b6 S b l f 90.&c6 Shi
75.Sb5?
91.Se7 Sgl 92.Sh7 Sdl 93.Sb7t &a8
75. &b2? with the idea < > b2-a3-a4 looks
94.Sd7 Shi 95.&b6 S b l f 96.&c7 &a7
attractive, but as in the game, Black has
97.&c8t &a8 98.c6 Sb8f 99.&c7 Sbl
75...a4!, when the a- and c-pawn ending is
100.&d8 Shi 101.c7 Sh8f 102.&e7
a theoretical draw, while 76.b4 Sg3 is also
1-0
impossible to win for White, with the king cut
ofFentirely.
The following study is based on the same idea
(definite spoiler), with exceptional details on
Thus, the winning idea was to reroute the king,
the way. This is one of my favourite studies
without allowing ...a5-a4. White should have
of all time. I have used it with students for
played: 75.Sa6! &d7 76.&b2 &c7 77.&a3
blindfold “guess the move” exercises, or as a
d?b7 78.He6 and after £&a3-a4, he will have
straight playing position. It is suitable for
either Ha6 or Sd6-d5, winning.
this, as a lot of decisions can be made with
the favourite defensive tool of elimination,
75...Sg3t 76.&b2 a4! 77.bxa4 &d6 78.a5
without seeing the full outcome from the
&c6 79.a6 Sg8 8O.£c3
beginning. Especially as Black’s resources may
Black is also making a draw after 8O. ?b3
come as a surprise a lot of the time.
Sa8 81.Sh5 Sxa6 82.Sh6t &b7!.
232 Conceptual Rook Endgames

Black is threatening ...<&g6-f5, which 2...±g8!

combined with an attack on the f-pawn would A strong prophylactic move.
win quickly. There is only one try.
After 2...&f8?!, the simplest way for White
A. Gilboa &Y. Afek to secure the draw is 3.Sh7! Sxf5 4.&b7 and
the white king joins the fight - something the
1st prize Kalandadze MT 2019
black king cannot do.

abcdefgh abcdefgh
1.5! White needs to bring the king out without
The black king cannot escape, as something letting the black king out, which requires that
like l...&g7 is met with 2.Sg4f and 3.Sg6, the rook moves a bit across the 7th rank. The
when Black cannot make progress. So, he has b7-square is obviously needed for the king,
to attack the pawn immediately. leaving only three options. The right one can
be found by elimination.
l...Sb5!
Definitely the difficult move to deal with. 3.Sd7!!
The white king is kept cut off, while the pawn 3.Se7? is wrong, as after 3...Sxf5 4.&b7
is attacked. Se5! Black wins a tempo. 5.Sd7 Se6! The
white king is cut off again and Black is ready
l...Sfl?! is less challenging, as after 2.Sa7t to play ...f5 and ...Sf6, winning.
&g8 3.&b7! the white king comes back in
good time, after the black king has had to 3.Sc7? is also a mistake, as it is clumsily placed
retreat. after 3...Sxf5 4.&b7 Sg5! 5.&c6 Sg7, where
Black wins an important tempo, forcing the
2.Sa7t white rook to the inconvenient back rank.
White has no choice but to seek to dominate 6.Sc8t &f7 7.&d5 Sg4! and the white king is
the black king, at least for a moment, so he can cut off, leaving White lost.
get his own king back in the game.
3...Sx5 4.±b7 Eg5
2.Sf3 &e7! with the idea 3.Se3t &d7 4.Sd3t Still the most dangerous try.
&c6 would be hopeless.
5.±c6 Hg7
 Chapter 1 6 - Stalemate 233

Something special

Sometimes life imitates art. And when that

happens, it usually outperforms it. The
following game is fantastic, at least in the
analysis. Black is threatening to come out with
the king, similar to the previous example.

Ernesto Real De Azua - Vitaly Sivuk

Lorca 2019

abcdefgh 8
White is faced with the same threat as in 7
the line with 3.Sc7? above. He has only one
6
defence.
5
6.Sd4I! 4
The only way to prevent the black rook from
cutting off the white king. 3
2
6...&£7
1
6...Sg5 7.Sd7! also holds for White.
abcdefgh
7.&d5 2g5t 8.&e4!
59.2a8f?
If 8.&C4? then 8...Hg3! is the winning move.
59.Ha3? £f7 6O.'i?xd5 Hg4! is another
(My previous notes gave 8...He5? as winning
version of what we saw above.
for Black, but after 9.Hdl &g6 10.&d4 White
will draw with defence from the front.
59...6.7 6O.Sa7t
bO.&xdS Sg4! wins.
8...Sg4f 9. 6! 2xd4
60.. .*ii?g6 61.Hal Hd7
Black is winning.

62.2glf 63.&d4 Hd6 64.2flf &g5

65-Hglf £f4 66.Hflf &g4 67.&c5 Sd8
68.±d4
68.Sxf6 d4 and Black wins.

68.. .B 69.Hglt &f4 7O.Hflf &g5 71.2glf

72.2ft

abcdef gh
234 Conceptual Rook Endgames

Black will also find it impossible to get the king

out to help with the conversion after: 59...Sg5
6O.Sa7! Se5 61.&d4 and the king is cut off.

60.d?xd5
Taking the pawn luckily works, as otherwise
Black would win with ...Sg5-e5.

6O...Sg5t

abcdefgh
72...Sa8! 73.&xd5 Sa4
The white king is cut off and he realised the
futility of his position.
0-1

Instead, White can draw with:

59.Sa4!!
The main line has to be the attempt to bank
on the f-pawn. But Black can also try to hold
on to the pawns, which requires a look.
abcdefgh
61.d?e4! Sg4f 62.±f5 Sxa4
59...Sd7 60.&d4 £f7 61.S a l &e6 62.Self Having burned this image thoroughly into
£d6 63.Sfl Sf7 your retina, it is time to move on to a few
random, yet delightful, examples.

S.L. Narayanan - Marcel Kanarek

Stockholm 2013

abcdef gh
64.§f5 White wins the d-pawn and secures
the draw. For example, 64...&e6 65.Sxd5 Sd7
66.Sxd7 &xd7 67. &e4 with the most basic of
draws.

abcdefgh
 Chapter 16 - Stalemate 235

The Indian star lost without a chance after: Aryan Tari - Jaime Santos Latasa
74,Sg8t? &f6! 75.h4 Sxf2f 76.&gl Sd2
Mayrhofen 2022
77.h5 f2t 78.&g2 e3 79.h6 S d l 80. W &g6
0-1

Instead, White could have saved the game

with:

74.h4f! ±g4
White will be able to bring out the king
to attack the e4-pawn after both 74... &g6
75.&g3 and 74.. .&f5 75.W &e5 76.Se8f
&d4 77.Sd8t &c3 78.&g3. In both cases,
White is making the draw.

75.Sg8f &xh4 abcdefgh

The only try that is different from the lines
above. 62...Se8? 63.&g6 Sb8 64.Sf6 Se8 65.&h6
Sb8 66.gg6
1-0

Instead, Black could have used the stalemated

situation of the king to achieve what is
popularly referred to as “perpetual rook”, a
special sub-section of the stalemate theme,
with:

62.. .d4!! 63.cxd4 c3 64.bxc3 Sflf

8
abcdefgh 7
76.Sg4f! 6
76.Sh8t &g5 77: &g3 Sglf 78.&h3 &f4 5
79.Sf8t &e5 and Black wins.
4
< >
76... xg4 3
Stalemate.
2
Under a lot of pressure, the talented Jaime 1
Santos missed an escape road. abcdefgh
The checks will only end with the capture of
the rook and stalemate.
236 Conceptual Rook Endgames

65.&e6 8
65. g6 Sf6f! draws immediately.
7

65.. . S e l f 66.±d5 Se5t 67.±d6 Sd5t 6

With a draw. 5
4
In the next example, the heavy underdog 3
missed a chance to save the game with a 2
beautiful stalemate trick.
1
Lars Hauge - Ludvig Carlsson abcdefgh

Oslo 2020 59...&a3! 6O.f6 Sa2f 61 .<±>bl Sb2f 62.Sxb2

Stalemate.

54.f5t ±d3 55.±cl a3 56.±bl Sg2 57.§f4

±c3 58.(6 ±b3 59.SBT ±a4 60.(7 Eglt
61.±c2 Sg2f 62.±d3
1-0

Boris Gelfand found the next example in

an old edition of 64 and allowed me to use
it in this book, when we did not find an
appropriate place to put it in our joint book
Decision Making in Major Piece Endings. As I
have not found it in any database or online,
abcdefgh I am unsure if this is a variation or what was
actually played, but the position is nice.
52...a5! 53.Sxg4 a4?
A natural move, but one that allows White Shirokov - Rubin
to reroute the rook to the 3rd rank, after which
it is game over. Nordic Youth Championship 1984

The most human drawing line is based on

stalemate (they all are) and a lot of natural
moves.

53.. .6d3! 54.&cl

White should be careful about running from
the draw. 54.&el e3 55.&fl? &f3 would
be unfortunate.
54...6c3 55-<±>bl Sb2t 56.&al Sg2 57.Sg8
57. f5?? &b3 would lead to mate.
57.. .a4 58.f5 &b3 59.Sb8t

abcdefgh
 Chapter 1 6 - Stalemate 237

White can try a lot of decent-looking moves. 5.Sxel

One by one, we can disregard them.

I.&b5? fails to simple play: l...&d4 2.c5

Otherwise 2...Se5t is coming. 2...el=®
3.Sxel Sxel 4.c6 S b l f 5.&a6 &c5 6.c7
Sxb6f 7. a7 Sc6 8. b7 with a draw.

After l.Sel? &d4 2.&b5 (2.c5 &d5 and

3...Se3 makes the draw easily.) 2...2e5t! Black
can draw in whichever way he chooses. 3. a6
could be met with 3...&xc4 4.b7 Se8, when
abcdef gh
5.Sxe2 Sxe2 6.b8=® Sa2f is another hoover-
variation. 5...Sb6t! 6.®xb6
Stalemate.
The most important variation to calculate goes
like this: For this reason, the winning move is: l.d?c5I!
I.c5?&d5 2.c6
White is seemingly winning.

abcdef gh
But an amazing resource exists: I...d?d3 2.b7 White has managed to avoid all
2...Se6!! 3.b7 Sxc6 4.b8=® el=®t!l the traps and will win the game.
The engine says that Black can also draw
by promoting to a bishop, but honestly, in
1984 no machine would understand what
was happening...
Chapter 17

Zugzwang
The demand to move against your own interests is one of the features that makes chess a special
game. This is not exclusively an endgame phenomenon, but when it arises in the middlegame,
it always feels both sensational and misplaced, as all associations with this theme are from the
endgame, where it is so entrenched that there are positions we don’t even think of as zugzwang,
but rather as opposition - which is zugzwang by another name. Zugzwang is not an uncommon
phenomenon in rook endings either. There are many positions where the defender is suffering
partly because of the inactivity and inflexibility of his rook’s positioning. In the endgame we should
always have a feeling of what the opponent is intending to do, or can do. If it is nothing, often
we have the chance to strengthen our position before changing it, a theme akin to “evolution/
revolution” known from attacking chess (see Attacking Manual 7, Chapter 7). At other times all
moves would worsen his position, in which case it would be important to ask the opponent to
move. This is what we call zugzwang.

The first example sees an experienced IM lose a drawn endgame against a strong G M , because he
was unaware of the dangers of getting into zugzwang in this position.

Ali Farahat - Robert Hovhannisyan

Sharjah 2022

abcdefgh
240 Conceptual Rook Endgames

59.c7? and a b-pawn), Gelfand resigned instantly

White would have been able to hold the against Anand in the play-off of their 2012
game if he had just waited. 59.Sc7! may look World Championship match.
strange, but the c-pawn is still defended and
if the black king tries anything, the rook will 66.Sf8 &g2 67.Sg8f
come out from its awkward position with a 0-1
check, leading to a situation where the c-pawn
actually becomes useful on the 7th rank. We should be careful with trapping our own
rook in front of a passed pawn. It often needs
59...Sd7 a little bit of air...
Black keeps the tension. White can only
make a move with the rook, after which he Aleksey Goganov - Nikita Vitiugov
loses his greatest asset.
Nizhny Novgorod 2013

6O.Sb8 Se7f!
An important move that asks the white king
if he would prefer to enter a mating net, or if he
would prefer to make way for the opponent’s
king.

61.&d2 Bxc7 62.Bxb5 Sa7 bS.&el

White had to defend against the black king
going forward and controlling all the squares
in front of his passed pawn. Instead, he loses in
a well-known way.

63...Sxa3 64.Sf5 S a i f 65.&d2 abcdefgh

48.h7?
48.Sh7? was the remedy in the previous
example, but is not a panacea. In this position
it loses to: 48...<±>e3 4 9 . £ e l d3 5O.Sh8 f3
51.Sh7 (Or 51.h7 f2f 52.<±>fl Se7 with
zugzwang.) 51...f2f 52.&fl d2 53.Sd7 And
now the simplest is 53...dl=®t 54.Sxdl
Sxh6, winning.

48...Se7! 49.&d2 <±>e4! 50.&e2 d3f 51.&E2

It makes little difference if we mirror the
situation with the king blocking the other
abcdefgh pawn.
65...SA!
51...B
This setup is valuable to remember. In a
Zugzwang.
similar situation (albeit with reversed colours
 Chapter 1 7 — Zugzwang 241

White needed to prevent ...&e3 more than

anything.

48.&d2! &e4 49.&e2 d3f 50.&d2 B

51.Sh7!
The needed waiting move. Black has no
viable way to play for a win.

51.h7? Se7 would lead to zugzwang again.

And all the attempts to activate the rook would
lead to pawn-down, lost endings.

abcdefgh Aagaard

White cannot move the king, as allowing Original 2019

...&e3 would lead to lines similar to those
given above.

52.Sa8 Sxh7 53.Sa4f &d5 54.&e3

54.&xf3 Se7 would see the white king cut
off.

54...Bf7 55.Bal £2 56.BA &c4 57.&d2 Ba7

58.Bclf &d4 59.Bal

abcdefgh
This composition came around through
rubbish analysis of Levenfish - Rabinovich,
Leningrad 1934, leading to an interesting
situation where triangulation can force Black
into zugzwang. It does not have great value as
a study, as White has two ways to win. But it
is a nice illustration of how it is possible with
carefully refined finesses to put the opponent
abcdefgh in zugzwang, without there ever being a
More a joke than a trap. zugzwang moment the other way (we will deal
extensively with mutual zugzwang in the next
59...Sa2f chapter).
I like 59...fl but some would argue
this is playing with one’s food... The starting point to solving this study is
0-1 to realise that if it was Black to play in this
position, he would lose. If the rook retreats,
242 Conceptual Rook Endgames

2.&f3 and 3.g4f just wins. Thus, Black is G M Rakhmanov and many other GMs solved
reduced to two attempts. this position, and all in the same way. They
realised the theme was mutual zugzwang and
1...6g4 2.g6 Sxc3 3.Sxa2 Sxg3t 4.&f2 is gave the following variation:
winning for White. The key factor in this
position is that the pawn is on c5, not c4, l.&h2 Sa8
so that 4...&xf4 loses on the spot to 5.Sa4f. l...c4 2. h3! puts Black in zugzwang. The
4... 5 b 3 loses a bit more elaborately after rook has to retreat. 2...Sa8 3. g2, when we
5.Sa8! Sb7 6.Sf8! and the connected passed have previously seen that White is winning on
pawns decide. account of 3...Sa3 4.&f2, although 3... e4!?
is a bit more complicated, as we shall see by
I...c4 2.&f2 is a position of mutual zugzwang. transposition below.

2.&gl!
White is in the middle of traditional
triangulation. If Black plays 2...Sa3, 3. g2!
would have passed the “right” to move to Black.

abcdefgh
If White was to play, his advantage would
dissipate. Black is perfectly set up to react
against everything. If the king moves to the
3rd rank, the c-pawn is taken with check, and
if to e2, the g3-pawn is undefended.
And after 3. g2 Sxc3 4.5=£xa2 <&g4 the
g3-pawn will fall with check. I n this variation abcdefgh
there is no check on a4, so Black can follow
2...Sa6
up with 6...&xf4. Alas, it is Black to play
So Black must try a waiting move.
and whatever he does, all these avenues of
counterplay evaporate.
3.&£2
Threatening 4.&f3.
After 1 .<&f2? c4! 2.<&g2 Black can play 2...Sxc3
and 3...&g4, making the draw. With the pawn
3...Sa3
on c4, there is no check on a4, as already
3.. .'&g4 4.g6 and White will soon be two
established.
pawns up without complications.

And l . f l ? with the idea l..,c4? 2.<&f2! would

4.&g2!
allow Black to play l...Sxc3 picking off the
The move has been passed to Black, who is
white pawns.
now in zugzwang.
 Chapter 1 7 — Zugzwang 243

But one variation is not fully satisfactory, The “simplest” winning method appears to be
although it wins. Black can seek active to give up the queen for the c-pawn and queen
counterplay with: an additional pawn. If White only had one of
his pawns, the position would be a simple draw.
2...c4!?
9.&B c2 10.®b8f &c3 ll.Wc7t &b2 12.f6

win long and in need of accuracy, something

computers take for granted, but humans rarely White will eventually queen one of the
deliver. pawns and then convert the queen vs rook
endgame (at least in theory).
4.g6 &d3 5.g7 &xc3 6.f5!
6.g8=®? Sxg8 7.Sxa2 <&b3 would allow So, in order to avoid this long winning line
Black to draw by one tempo. with queen vs rook, the strongest way to play
would be:
6...&b2 7.Bxa2f Bxa2 8.g8=® c3
l.<±»h3!

abcdefgh
abcdefgh
244 Conceptual Rook Endgames

No one ever suggested this move. If the long Paul Dargan - Julien Shepley
variation above was leading to a draw, this
England 2022
would be a fantastic study. As it does not, it
will have to do as a great illustrative position.

1...5.8
1...c4 loses to: 2.<&h2!Sa8 3. g2 Sa3 4.&f2
and we have another of our key positions.
White wins.

2.&112! Ha6 3.&gl!

3. < ’g2 Ha3 would bring us back to where we
started.

3.. .ga7
3.. .c4 4.<&g2!Sa3 5. f2 wins again.
abcdefgh
46.a4!
Finally, threatening 5.&f3. The pawn leaves the 3rd rank, making it less
likely it will drop off later.
4.. .ga3 5.&g2!
We have returned to the starting position, 46...a5
with zugzwang. White wins. There were other options, as so often, but
nothing better.
The final example of this chapter was played in
the Four Nations Chess League (4NCL) by a 47.ge7?
student of the Killer Chess Training academy. A bit of foreshadowing to the next chapter
I had picked up on the game without paying occurs here.
attention to the names, as I often do, and it
was only when it came to discussing the theme 47.Sf3I!
of zugzwang at the academy that I realised the This would prepare to put the rook on f5,
coincidence. which is the ideal position. From there the
rook would attack the f6- and c5-pawns
(as well as the a5-pawn down the line) and
prevent Black’s only pawn break, ...h5. But
if it was White to play, he would have no
way to get it there. His only “pass” move
would be h3-h4, which would weaken the
g-pawn and lose time in a later race, should
the kingside fall apart. But since it is Black
to play, and only the rook is realistically able
to move, White will be able to move the king
and vacate the f5-square.
 Chapter 1 7 - Zugzwang 245

47...Sdl 5O.Sxc5 Sxa4 51.h4?

Other squares for the rook exist, but they all Paul missed a chance here. H e could have
lead to similar outcomes. played 51. e4! Sa3 52.Sh5!> when the
48.<±>e4 Sd4f 49.&e3 h5!? 5O.Sf5! counterplay on the kingside will come much
too late and the c-pawn wins effortlessly.

51...Sal 52.h5 Sflf 53.&e4 Sgl 54.&B

Sflf 55.&g2 gf4
55...Sell with the idea 56.Sc8 f5 57.gxf5
&f6 and Black will draw effortlessly.

56.&g3 Sd4 57.Sc8 f5 58.gxf5 59.Sc5

<±>g5 60.&B &xh5 61.&e3 Sdl 62.Sc7 fib?!
To me 62... g5! 63.Sxf7 h5 with active
counterplay offered by the h-pawn feels like an
abcdefgh
easier way to draw, although a few more moves
Black will lose the pawns on the 5th rank, would have to be played.
leaving White with the a- and c-pawns, which
will perform spectacularly against the doubled
f-pawns in a race.

47...Bd3 48.ga7 Bxb3 49.Bxa5

abcdefgh
63.Sg7!
Keeping the black king under control - the
beginning of Black’s later problems.

abcdef gh 63...Belf 64.&d4 ge5 65.c5 Bxf5 66.c6

49...Sb4? Sfl 67.Sg3
49...Sxh3 5O.Sxc5 h5! would have created a
passed pawn and secured enough counterplay
to draw the game.
246 Conceptual Rook Endgames

8
7
6
5
4
3
2
1
abcdefgh abcdefgh
67...Sal? 69.&d5?
This could have lost to a fabulous idea. Analysing the game afterwards with his
teammates Paul found the following staircase
Better was: 67...Sdlf 68.Sd3 Scl 69.Sc3 zugzwang mechanism.
S d l t 70.&c5 Sd8 71.c7 Sc8! 72.&d6 £g4
73.&d7 69. e4!! Sc8 7O.£f5! &h4 71.Sc3 h5

abcdefgh
72.<±>f4
abcdefgh The first zugzwang.
Black should play 73...Sxc7t! and 74. ..h5!, 72...f5 73. S c l
making the draw. (Usually, we would look A small waiting move, similar to the
triangulation above, albeit simpler and more
to have the pawn queened as far away as
beautiful.
possible, but in this case, it would improve
White’s fortunes. After 73-..Sg8?? 74x8=9 73.Sc2? <&h3 would put White in zugzwang.
Sxc8 75.Sxc8 h5 he has the additional option Black escapes.
76.§g8f! &f3 77.§h8, winning a crucial 73.. .6.3 74.Sc2!
tempo in the race.) The king is contained. Another zugzwang.
74...h4 75.£f3 f4 76.Sc3 <±>h2 77.Scl!
Another zugzwang.
68.c7 Sa8
 Chapter 1 7 — Zugzwang 247

77...H3 78.&f2 f3 79.§c4 £ h l 80.§c6 &h2 74...Sxc7f?

81.5cl This time 74...Sa8! was the right approach.
After 75.c8=® Sxc8 76. <&xc8, the white king
is a move further away, allowing Black to draw
by a tempo.

75.&xc7 &h2 76.&d6 h3 77.&e5 £hl

78.&f4 h2

abcdef gh
The final zugzwang. White wins.

69...Sc8 7O.Sg7 &h4 71.&e6 h5 72.&xfi6

&h3 73.&e6 h4 74.&d7

8
7 abcdefgh
6 79.&g3
5 White wins.

4 79...&gl 8O.Sa7 hl-4it 8 1 . & B &h2

3 82.Sg7
1-0
2
1
abcdefgh
Chapter 18

Mutual Zugzwang
In my statistically unrepresentative database of interesting rook endgames, there are some
zugzwang examples, but even more with mutual zugzwang. Even the final example of the previous
chapter was not free from mutual zugzwang! Mutual zugzwang is used to describe a position
where the side to play cannot avoid a deterioration of his position. In most cases, and all the
examples in this chapter, the difference is between one side winning and the other side drawing.
I cannot remember any positions with mutual zugzwang in rook endings, where the side to play
loses, off the top of my head, but the following pawn ending springs to mind:

abcdefgh
The side to play will lose.

The examples in this chapter will all start in a place of relative innocuousness, before leading to a
surprising scenario of mutual zugzwang.

Mark Dvoretsky noticed the following example almost immediately after its publication and
wrote a big article about it, calling it a paradox. Whether the first move of this obvious race of
the a-pawn to the last rank is a contradiction to logic or simply to expectations I will leave up
to others to decide. But it is true that the first time people encounter an idea like this, they are
looking into a dimension of chess thinking that was previously hidden from them in plain sight.
250 Conceptual Rook Endgames

Van der Heijden The surprising solution to this danger starts

with one of the weirdest-looking moves you
2001
will ever have seen.

I.&b4!!

abcdefgh
l.&b5? &f3 2.a4 &e4 3.a5
When showing this example to students at abcdefgh
Killer Chess Training, 3.&c6 was suggested.
While still making space for the a-pawn, the
Within minutes, a student came up with the
white king is seemingly losing a tempo for no
smoothest solution. 3...&d4! 4.a5 &c4 5.a6
good reason. As the rook remains occupied
Sh6f 6.&b7 &b5! and Black wins. with looking after the h-pawn, the black king
still has to rush off in a diagonal line towards
3...&d5 the opposing corner to stop the a-pawn.

2.a4 &e4 3.a5 &d5 4.&b5

abcdefgh
This is the key position. As it is White to play,
he is in zugzwang. After 4.&b6, Black wins at/cdefgh
with 4...&c4! 5.a6 Sh6f. And after 4.a6, with
With White having lost a tempo, it is now
4...Sb2f 5.&a4 <±>c4! 6.&a5 &c5 7.&a4 Eb8
Black to play. Without concessions from
and both the white pawns are under control.
White, Black cannot win.
 Chapter 1 8 — Mutual Zugzwang 251

4.. .5b2f Jacob Aagaard

The greatest surprise is probably the lack of
The Problemist (983) 2008
waiting moves.

4...2h3 would bring the rook too close to the 8

white king. 5.a6 2b3t 6.&a4 is no longer 7
the same. 6...&c4? would not threaten mate.
6
Instead, Black has to rush back with 6...2b8,
when White will make the draw after: 7.a7 5
2a8 8.&b5.
4
And after 4...S h i we see the real point of the 3
study. 5.a6 Sb I f 6.&a5 2
1
abcdefgh
l . ei! a2 2.Sa8 &g4 3.Sa7! h3 4.Sg7t
&h4 5.&f4! &h5 6.Sgl &h4 7.Shl
And White wins.

Let’s see another mutual zugzwang.

abcdefgh
6...&c5? no longer works, as 7.h8=® would
see the queen covering the mating square on
the a-file.

5.&a6 Sb8 6.&a7 Sh8

The final try. 7.a6? &c5! would lose.

7.&b6! abcdef gh
But this secures the draw. White cannot make progress. If the rook
moves, the pawn drops and the king can
Without noticing it, some doofus made what only move away. Black to play would make
is called an “accidental recomposition” and a concession. If his rook goes backwards, the
had it published... I thought I had found white king comes to c6. And if his king moves,
something great, but the key point is entirely the white king can come to c4. Sideway rook
the position from above. I have removed the moves like l...Sh6, allows 2.2c II, with White
variations, which are all rather repetitive. untangling himself from the inconvenient
sideways defence of the pawn. The rook would
no longer be in the way of the white king,
which will carry the pawn over the finish line.
252 Conceptual Rook Endgames

A. Mandler

Berliner Tageblatt 1954

abcdefgh
2.gg5!
White still has to count the steps. 2.§d5?
abcdefgh &g3 3.&g5 &f3 4.<±>f5 &e3 5.&e5 Sh6!
6.Sc5 &d3 7.&d5 Sb6 would leave White to
Let’s see how we got there. After l.&g5? &g3,
play.
the direct path with 2.&f5 &f3 3.&e5 &e3
4.&d5 &d3 would land White in zugzwang.
2...Sd6
The path to the same position is a bit more
2.. . <&h2 3.&g4 would see the black king
elaborate after 2.Bf5, where Black has to
entirely out of the game.
choose the right squares for the rook with care.
After 2...Sd6! 3.Se5 &f3 4.&f5
3.SB! &g3 4.&g5 Se6
4...5b6 is where Black wants to end up, but
not too early. After the basic shuffling on the
third and 5th ranks, starting with 5.Sc5 we
can easily see Black arrives first and will be in
zugzwang. And if he tries 5...&f2!?, White will
follow with 6.&f4, making no difference.

abcdefgh
4...Sh6! 5.Sd5 &e3 6.&e5 Sg6 7.Ec5 &d3
8.&d5 Eb6, again, it is White to play.

Since this is a study, it is not a big surprise that

the best move is:

l.&h5! &h3
1 ...&g3 2.' ?g5 would see Black arriving first
to the d-file, and soon thereafter in zugzwang. abcdefgh
 Chapter 1 8 - Mutual Zugzwang 253

5.Sd5! &f3 6.&f5 Sh6 the black rook down. Luckily the executions of
Black is trying to lose a move, but this time these two plans are not situated close to each
White has a real threat ready. other geographically and Black will be able to
Or 6...Sb6 7.&e5 &e3 8.Ec5 &d3 9.&d5, swing the rook back and forwards. First it has
winning. to go back to limit the advance of the white
king.
7.Sd3t! &e2 8.Sb3
The white rook has been rerouted and 56...Sa8?
the obstacle to pushing the pawn has been As with the paradox example, this move is
overcome. certainly the natural move any human would
make, if he did not have the foresight to think
Usually, when you illustrate a theme with something out of the ordinary was on the way.
a study, you are presented with something
mind blowing. Switching to examples from 57.&h5 Sh8t
recent games, you would expect to be a turn By now Schroeder must have realised the fate
to the mundane. In this chapter, this is not the awaiting him, which explains his somewhat
case. The first example is from a Bundesliga unexplainable moves.
game between two experienced grandmasters.
Presumably tired and short of time, Black did 57...Sg8 would have been the intention.
not see the coming mutual zugzwang until it White continues with 58.&h6! Sg5 and all
was up close and personal. looks good, until 59.&h7! comes.

Pavel Eljanov - Jan-Christian Schroeder

Germany 2018

abcdefgh
Black is in a terrible zugzwang. Quite an
unusual one and the game is done. There is
a small trick that can be tried. After 59...f5
6O.Sxf5t Sxf5 61.gxf5 &xf5, White should
play 62.&h6, when resignation is reasonable.
abcdefgh
But after 62.&g7? g5! Black would be able
Black has to rely on a frontal defence to avoid to take the opposition and keep the white king
White’s major threat of &h5-g6, followed by at bay.
Sf5t- At the same time, he has to keep an eye
on the f2-pawn, in case White wants to play 58.&g6 Sg8t
2f5 t and &g3-f4, attacking the e4-pawn, tying Trying to urge the white king to h7.
254 Conceptual Rook Endgames

59.&h6! Sa8 Black lost the following game after a few

59...Sg5 60.&h7 is covered above. mistakes. On the surface the game seems to be
nothing special, but looking under the hood,
60.&g7 we see a lot of details that are quite surprising.
Black loses the f-pawn and resigned.
1-0 Kazybek Nogerbek - Artem Tyurin

Novi Sad 2022

Black could have made a draw by the
paradoxical-looking:

56.. .5a7I! 57.&h5

57.2f5t e6 58.&g3 Black is in time to
return with 58...Sa2!, diminishing the damage
of &f4xe4 enough to secure the draw.

57.. .5g7 58.£h6

58.Sf5t &e6 59.&h4 again shows Black to
be in time with 59...Sa7 60.&g3 Sa2.

58...5g8 59.&h7 Sg5

abcdefgh
42...&g7?
Hoping to bring out the king via g6-f5, but
this is too optimistic and loses an important
tempo.

43.Sb6! 44.&d3 &e7 45.Sd6! Sa3t

8
7
6
abcdefgh
5
White is in zugzwang and cannot improve
his position. There is no triangulation or 4
waiting moves, leaving a friendly handshake as 3
all that remains to be done.
2
1
ab cdefgh
46.&C4?
This natural move is a simple blunder.
 Chapter 1 8 - Mutual Zugzwang 255

The winning method was 46.&e4!! Sa4f 44.. .6d8!! 45.Sd6t

47. d5 Sa5t 48.£c4 Sxe5 49.Sd5! S e i 45.2f6 &C8! transposes to the main line
5O.Sa5 S e l f 51. d5 Sc8 52.a7 Sa8 53.&c6 below.
and White wins, as the Black king is cut off
on the 5th rank and would have to go to g5 45...6c7!
to come out, which is both slow and on the 45...&c8? 46.Sf6 Another mutual zugzwang
wrong side of the pawn. position. If Black was to go with the king to
the d-file, it would be too far away from the
46...Sa5? a-pawn. And after 46...& b 8 47.Sxf7 Sxa6
Both players missed 46...f5!I, which would 48.Sf6! S a l 49.Sc6 the king gets cut off too
have made a draw. far to the other side.

47.&d4 Sb5 48.Sd5 Sb4f 49.&c5 46.Sfi6

1-0

Black needed to keep the king active, ready to

stop either of the white pawns.

42...&e7! 43.Sb6 &d7

43. ..&d8!? 44.&d3 &d7!! works in the same
way.

44.&d3

abcdefgh
46.. .6c8!!
Mutual zugzwang. The black king will follow
the white king. But if the black king would
have to go first, the white king would go in the
opposite direction.

47.&c3
47.&d2 Sa31? does not improve things for
White. But Black could also go for the e-pawn.
abcdefgh
Black is entering a minefield with mutual 47.. .6b8 48.&b3 Sa5
zugzwangs all over the place. For example, Black draws.
after 44...&c7? 45.Sd6!, when the king has to
go to c8, when 46.Sf6! is covered below.
256 Conceptual Rook Endgames

The next example is between two important

people in British chess. Chapman is a successful
businessman who has supported English chess
in various ways over the years. Most recently
as a supporter of the English over-50 national
team, which has been greatly successful in the
major events. His opponent is on the other
end of the age spectrum. Freddy Gordon is the
youngest ever player to play for the Scottish
national team. He was selected at the age of 10
to play in the European Team Championship as
reserve, where he drew against a grandmaster in
the first round, became top scorer and had the abcdefgh
second-best rating performance. This game was 62...&d5?
played only a few months before this selection. You can sense the inexperience of the young
Scot. Chasing the d-pawn has no future and
Terry Chapman - Frederick Waldhausen Gordon
led to a relatively unsurprising and quick end.
Internet 2021
Returning to move 62, Black would have been
able to make a draw by going for the white
pawns on the kingside with:
62...&d4 63.Sel e3 64.d7 &e4!
64...&d3? would set up a check on d l after
65.&e7.

abcdefgh
Chapman chose a natural continuation and
won the game.

61.&c7? abcdefgh
61.&d8? would also have allowed Black to 65.&e7
escape. After 61...&d4 62.S e i e3 63. d7 &e4! A somewhat more complicated attempt
Black would draw as seen below. is 65.g4l, when White is close, but not
close enough. 65...Sa5 66.&c7 Sc5t
61...Sc5t 62.&d8 67.&d6 S d 5 t 68.£c6 Sd2 69.h5 g5’
For later we notice that 62. &b6 S b 5 t ! is an 70.&c7 &f3 71.d8=S Sxd8 72.&xd8
important resource for Black. After 63. &a6 Sd5 &xg4! 1 73.Sxe3 xh5 74.&e7 &g4!
64. Sxe4f &c5 White should be a bit cautious. 75.&f6 <±>F4 76.Sa3 g4 77.Sa4f &f3 78.&f5
 Chapter 1 8 - Mutual Zugzwang 257

g3 79.Ea3t &f2 80.&F4 g2 81.Sa2f <£fl This is the point. Above we saw that Black
82.&f3 g l = & t Black makes the draw with had ...Sb5t. With this gone, lost in the
not a lot of margin. zugzwang, Black is done for.
65.. .5e5t 66.£d6 Bd51 67.£c6 Sd2 68.&c7
Black cannot keep checking the white king. 65...Sd5 66.Sdlf &e4 67.Sxd5 e2 68.d7
But after: White wins.
68.. .<&f3 69.d8=® Bxd8 70.&xd8 &f2 71. S a l
e2 72.&e7 &xg3 73.&f6 g5 In the usual time scramble we see at the end
Black is in more than good time. of rapid games, Black had to choose between
three possible squares for his king.
63. <17 &d6 64.Bxe4 Bd5 65.&e8 &c7
66.Sc4t &b7 Giga Quparadze - Manuel Petrosyan
1-0 Warsaw (rapid Wch) 2021

The win was pointed out by Keith Arkell. If

he found it, or his friend Chapman told him,
I don’t know.

61.g4!
This is a waiting move, pure and simple.

61...g5
61...h5 62.g5 is so similar it is almost
identical, while 61...&d4 62.S e i runs adjacent
to the main line.

62.h5! abcdefgh
This is the position of mutual zugzwang.
56...&d7?
We will start by considering a losing option
62...&d4 63.Sel e3 64.&c6 Sc5t 65.&b6
which Black avoided in the game.

57.&b2! S d l
57. ..Sei 58.e5 Se2f 59.& a l also exposes
the downside of putting the king on d7.
59...Sd2 6O.Sxh6 Sxd4 61.Sd6t would see
White queen a pawn.

58.d5 Bel 59.&xa2 Bxe4 6O.Bxh6 Bd4

61.d6
White wins. Black cannot take the d-pawn
and White can bring the king in to help. It
may at first glance look as if it will be difficult
for White to make progress, once the black
abcdefgh
rook is behind the h-pawn. But with the white
258 Conceptual Rook Endgames

king on f6 or g7, he would give up the d-pawn By method of elimination, I am sure all have
and get the h-pawn going. realised that the right move was:

The game went: 56.. .617!

It is difficult for White to hang on to both
56.. .6e7? 57.e5 the central pawns. Black is threatening ...Sei,
57.&b2? does not work this time around. picking up the e-pawn.
Because the king is not on d7, there will be
no check on d6. 57. ..Sei! 58.e5 S d l 59.&xa2 57. e5
Sxd4 6O.Sxh6 Se4 Black can take the e-pawn, 57. &b2 shows Black’s main idea: 57...Sdl
as the king this time is within reach of the 58. d5 S e i 59.&xa2 Sxe4 6O.Sxh6 Se5 61.d6
h-pawn. &g7 62.Sg6f &f7 and White will soon have
to yield.
57.. .6e8
If the king goes to f7, White has e5-e6f. If 57.. .6e7!
the king goes to d7, there is a check on d6.

58.Sa5
58.Sa7 &d8 59.Sa4 was also strong. The
a-pawn falls quickly.

58...6d7

abcdefgh
We have reached the position of mutual
zugzwang. If Black was to play, he would have
to put his king into harm’s way. Alas, he does
not.

58.Sa4l?
abcdefgh A clever waiting move.
59.Sa4?
A crude blunder. 58.d5 S e i picks up the e-pawn.

White would have won instantly with 59.&b2 58...&e6 59.&b2 S h i 6O.Sa6f!
S d l 6O.Sd5t, with two extra pawns. The only challenge.

59...&c6 6O.e6 &d5 61.e7 S e i 60.&xa2 & d 5 would cut the line short. White
The game was later drawn. will not be able to put any challenges to Black.
 Chapter 1 8 - Mutual Zugzwang 259

60...&d5 61.§d6f &c4 62.&xa2 Sxh5 48.h3? g5 49.&g3 h5

White is already in dire straits. If the king
retreats, 50...h4 will put White in zugzwang.
White will have to move the rook, after which
the black king will surround the e5-pawn all
by himself.

5O.h4
Natural desperation.

5O...gxh4t 51.&112 &e8 52.Sa3 &d7

53.Sa6 &e7

abcdefgh
White has two connected passed pawns.
But being under observation, they are slow to
advance, whereas the h-pawn is poised to run
unopposed down the board.

63.e6 Sh2f 64.&a3 Se2 65.d5 h5 66.Sd7

h4
White will have to eliminate the h-pawn
sooner or later, giving up his two pawns on the
way.
abcdefgh
The following example is quite complicated, White will have to let the black king pass
even if the game was not. the 6th rank, but chose to resign instead.
Then we come to a basic theme of theoretical
Sanan Sjugirov - Aleksey Goganov rook endings. Black will play ...h3 at some
point. White has to play gxh3, after which the
Sochi 2020
f-pawn advances and will eventually have to be
eliminated, when the black rook comes out to
win the game.
0-1

White’s defence had to start with:

48.&g3! &e8
After both 48...g5 49.&g4 and 48...h5
49.&h4, White will hold.

abcdefgh
260 Conceptual Rook Endgames

White has to make a concession.

49.Sa8t
There are other reasonable tries, which
generally lead in the same direction.

49...&d7 5O.Sa6 &c7 51.2a7t &b6 52.Sa8

&c5

abcdefgh
Either the h4-pawn will be lost, or White
will have to weaken his position further.

57.&h2 &f4 58.g3t 59.Sa3t &e4

60.§a5 &f5 61.&g2 &e6 62.&h2 g5!

abcdefgh
We have reached a critical position. To
understand how White can save it, we will first
consider some losing options.

53.h4?
This would lead to a terrible weakening of
the white position.

53.Sa7? is also inadequate: 53. ..&d5 54.Sa5t

&e6 55.&g4 (After 55.&h3 h5 56.&g3 g5, abcdefgh
the black king will soon arrive on f4, causing
The purpose of this is to imprison the white
zugzwang.) 55...Shi and Black wins.
king.

53.Sa5t? &b4 54.Sa8 &b3! is similar.

63.hxg5 hxg5 64.&g2 g4 65.&h2 &d7
66.Sa6 <&c7 67.Sa8 &b6 68.&g2 &b5
53...&d5 54.Sa5t &d4 55.&h2 &e4 69.&h2 &c4 70.&g2 &c3 71.&h2
56.&g3 &f5
 Chapter 1 8 — Mutual Zugzwang 261

abcdefgh abcdefgh
71...Sdl! 72.Sxa2 Sd2f 73.Sxd2 &xd2 A necessary move. If the black king comes to
The pawn ending wins. f5 and later f4, White will end up in zugzwang
and lose the e5-pawn. After the text move
White needs to defend against losing the Black can circle around via the b- and c-files
h-pawn and should play: and get to e4, but White would wait with the
king on g3, so ...&e4 can be met by &g3-g4,
53.h3! and the black king cannot make it to f5.

I have not been able to put together a

meaningful variation, as I cannot see a serious
try for Black (virtually only the king is able to
move). It is easier to show how White would
lose, if White was to play; which makes sense,
as White is able to put up a blockade here, but
zugzwang would create a crack in the wall...

55...(pass) 56.h4
A minor concession, but enough to tip the
balance.

abcdefgh 56. g3 loses on the spot, after 56...h5t, when

53...&d5 57.&g5 Sgl wins pawns, and 57. h4 &f5
The critical try. leads to zugzwang.

White is also fine after both 53. ..h5 54. h4 and 56...6d7
53...g5 54.£g4!. The black king will circle around via the
queenside. White can play in a lot of different
ways, which all lead to the same place.
54.Sa5t &e6 55.&g4!
262 Conceptual Rook Endgames

57.&g3 &c6 58.Sa8 &d5 59.Sa5t &d4 63.g3t &e4 64.<£>g2 &d4 65.&h2 &c3
The zugzwang train starts. 66.Sa8 Sdl !

6O.£g4 &e4 61. g3 £f5 62.<£>h2 &f4

8
7
6
5
4
3
2
abcdefgh
1
67.Sxa2 Sd2f
abcdefgh Black wins the pawn ending.
This is the painful moment. White has to
make more concessions.
Chapter 19

Tempo Wars
Endgames often go through two phases. In the first, manoeuvring is happening. Perhaps one
player is in control and the other is preparing himself against the various ideas he can be faced
with. But once an attempt to convert a positional advantage into a full point begins, it has an
irreversible nature and the game goes from a period of preparation into a period of consequences.
Often this happens with a pawn break or a pawn race. Endgames are mainly about queening
pawns. Rook endings are no different. In this chapter we will look at situations where the players
can utilize finesses to prepare for the race and shift the balance in their favour.

We know how these Internet rapid games go. The players run out of time and are presented with a
meagre 10 seconds per move, which never adds up to much. Before you manage to have a second
thought, you have to play your first one... Navara played everyone’s first idea, running with the
queenside, creating a passed pawn and then seeing it arriving too late.

David Navara - Loek van Wely

Internet 2021

abcdef gh
48.b4? e4!
The e-pawn is well supported by the black rook.
264 Conceptual Rook Endgames

49.fice4 fixe4 5O.a4 e3 5 1 . § e 2 &f5 52.b5 61J&c6 &£2 62.g4 hxg4

axb5 53.axb5 &e4 0-1

White could have won a tempo by making it

harder for Black to create a passed pawn.

48.§£2!!

abcdefgh
54.b6
54.§a2 is an interesting idea in some
positions, where White wants to get to the
far side to harass the king. But after 54...&d3
55.Sa3t &c4, the e-pawn looks formidable.
abcdefgh
Obviously this looks artificial, but the
54...&d3 5 5 . § e l &d2 56.Sal e2 57.b7 natural line went straight to the mortuary, so it
Se7t 58.&c6 Sxb7 makes sense to look at something less obvious.
The simplest.
48...&e7
59.&xb7 e l = ® 60.§xel &xel At the academy we tried hard to get other
lines going for Black, but nothing was working.

Analysis of 48...&g6 49. b4 e4 5O.fxe4 fxe4

proved that here white has to play 51.g4!,
to make use of the extra tempo. It draws
uneventfully. Although 49.&d7! is simpler and
underlines the futility of the king move.

49.b4
White has won a tempo for the race. The
rook will have to block the e-pawn no matter
what, while the black king is farther away.

abcdefgh 49...e4 50.fxe4 fxe4 51.a4 e3 52.§e2

The white king is way offside for a pawn 53.b5 axb5 54.axb5 &f5 55.b6 &e4
ending.
 Chapter 1 9 — Tempo Wars 265

62.Sa5!
Not an overly difficult move, but absolutely
essential. The rook comes behind the passed
pawn, and anticipates ...&xg2 which can now
be met by Sxg5t.

62...&fl
62...g4? 63.Sxh5 g3 would be trying too
hard. After 64. h4 the white pawn arrives so
early that Black will lose.

63.Sf5t &gl 64.Se5

abcdefgh White makes the draw.
56.8a2!
An important idea. The next position has many similarities with
the previous example, except this time winning
56„.&d3 57.Sa3t! the tempo war means winning the game.
The most natural is to harass the black king.
Denis Khismatullin - Ernesto Inarkiev
White can also opt for 57.b7I? Se7t 58.&c6
Sxb7, but would then have to play 59.Sa3f! Minsk 2016
&d2 6O.Sa2f! to make the draw.

57...&d2 58.Sa2f £el

Eventually the king has to go to an
inconvenient square.

59.b7 Se7t
Black does not get more out of 59...Se8
6O.Sa8, although 6O...Se7t would still hold.

60.&c6 Sxb7 61.&xb7 e2

abcdefgh
If Black plays with his hands, he will already
have lost his head:

46...b5? 47.e5 b4 48.§xd6f &e7 49.Sxh6

abcdefgh
266 Conceptual Rook Endgames

play, as the rook is no longer hanging on the

7th rank. At the same time, it asks a pertinent
question: what is White actually doing, besides
waiting to meet ...b5 with e4-e5?

47. h3
Nothing, it seems.

47.&c4 b 5 t 48.&b4 &e5 would see Black

switch to the kingside similarly to what
happened in the game.

abcdefgh It is tempting to look at active defence for

We see the issue. White is threatening White. Usually this is what works, so it should
5O.Sh7t and Black will have to lose a tempo be our first instinct.
to avoid losing the rook.
But after 47.Sa7 b5 48.Sh7 b4 49.Sxh6f
49.. .5b8 5O.Sa6 b3 51.Sal b2 52.Sbl &d7 50.<&d5 Black can push the b-pawn, as
White is in time and the draw is not far away. White will not succeed in threatening mate.
Although the 3-second human solution
Inarkiev understood all of this well and for this 50...5.5t 51.&c4 b3! is both simple and best.
reason played a beautiful preparation move,
winning the tempo war. 47.. .b5
A number of our students wanted to play
46.. .5b8!! for zugzwang with 47...Sb7?> but the rook
is poorly placed on b7, as we remember.
After 48.Sa8 b5 49.Sh8 White has enough
counterplay.

48.Sa3!?
Khismatullin plays passively out of need,
rather than desire. He must have seen that
48.e5 b4 49.Sxd6f &e7 was hopeless for the
reasons mentioned above, and decided to hope
for his opponent to make a mistake down the
line.

48.. .b4 49.Sb3 Sb7

abcdefgh Putting White in zugzwang. No mistakes are
forthcoming.
The idea is simple. In the line above, 5O.Sh7t
was a nuisance, but here it will not come into
50.&d3 &e5 51.&e3
 Chapter 1 9 - Tempo Wars 267

54...a5? 55.h6 Se2f 56.&xd3 Sh2 57.g5

Sh3t 58.&d2 Sh2f 59.&d3

An interesting try was pointed out by Aradhya

Garg: Black can play 54...2e2f? 55.&xd3
Sh2, with the idea of controlling the white
pawns from behind in classical style.

abcdefgh
51...Sc7!
The simplest is to play actively.

52.Sxb4 Sc3t 53.&£2 Sxh3

Black has too many threats. ...&f4 and
...Sd3-d4 are the first to mention.
abcdef gh
54.&g2 Se3 55.Sb5t &f4 56.Sd5 &x g 4 The problem is that Black is not actually
57.Sxd6 h5 threatening to take any of the pawns. White
White resigned. Excellent technique from should now advance the pawn not controlled
Inarkiev. by the rook, which looks as if it would be a
0-1 tempo up - and simply is a tempo up. 56. g5!
&b2 57.g6 and White makes the draw.
Carlsen did not believe that he had any
winning chances and played a random move However, it is also possible to lose to the
without serious engagement. following classical trick: 56. h6? &b2 57. g5
cl=W 58.Sxcl &xcl 59.&e4 a5 60.<&f5 a4
Fabiano Caruana - Magnus Carlsen
61.g6Eh5t!
Karlsruhe 2018
The winning idea was unsurprisingly
connected to slowing down the advance of the
white pawns. In this position this is unusually
happening from the front.

54...Sh7!

abcdefgh
268 Conceptual Rook Endgames

abcdefgh abcdefgh
55.&xd3! Black will have to play this idea at some
This is the only serious try. point to win the race. It makes more sense to
do it earlier, when far advanced pawns would
The passive options all look the same, so we not provide White with a bit of a mess, if he
may as well choose the least embarrassing chooses to take the rook.
of them to illustrate why it does not work:
55.Sb8f &a2 56.Sc8 Forced, in order to 58.Sc7
prepare &xd3. 56. ..a5 57.&xd3 & b l The This move is forced, but its a sad concession.
c-pawn is immune, which is part of the reason Black has won another important tempo, as it
why the rook is so well placed on h7. 58.Sb8f will take one more move for the white pawns
& c l 59.Sf8 a4 O r something like that. Black to be able to threaten to make it to the back
wins in all the variations. rank (as both will have to make it to the 7th
rank now). After:
55...Sd7t 56.&e4
56.&e2 is not as silly as it looks. White is 58...a4
trying to take the d l -square away from the The black pawns arrive first.
black king and trying to create a race between
the pawns. This loses, but with the king placed The last game in this chapter has loads of small
passively, 56...Sdl! is simpler. Black arrives tempo war manoeuvres. It is not a perfect
with lots of time to spare. game, but it has a lot of nice ideas in play from
both players and should be highly instructive.
56...a5 57.g5 Sd8!
 Chapter 1 9 - Tempo Wars 269

Aleksey Goganov - Maksim Schekachikhin

8
St. Petersburg 2021
7
6
5
4
3
2
1
abcdefgh
36.±e4?!
Goganov definitely considered 36.Sc5, but
abcdefgh must have failed to notice that the d-pawn
drops after 36...Se7 37.Sc4!.
31.Sa6!
The rook is perfectly placed here, ready
36...Sc2!
to quickly create passed pawns, which is
Black manages to create counterplay.
persistently the most successful strategy in the
Goganov now decided to go for the a-pawn,
endgame.
which creates the beautiful connected passed
pawns.
31...&g7 32.&B h5 33.h4!?
This move is fine and Goganov follows it up
37.Sa5?
with a number of accurate moves that should
The problem with connected passed pawns
have won easily, except in the end, he had
is that they both have to move in order to
to make more than he was able to do to win
exploit their combined power of 1 +1=®. And
without “complications”.
meanwhile the white rook will be far away
from the action on the rest of the board after
I prefer 33. e3, quickly attacking the d-pawn,
taking on a7.
ready to meet 33... h4 with 34.gxh4!, when
White is far ahead in a potential race between
The core fault of going for the a-pawn is that
the two kings.
White unnecessarily enters complications. It
was possible for him to win a pawn and keep
33...5e7 34.Sa5!
control. He would be far better placed for a
An important tempo-winning move. The
future race, if he had already eliminated the
rook attacks the d-pawn, while at the same
d-pawn. Instead, he decided to provoke the
time keeping an eye on the a-pawns. One
race immediately. Both players now made
to defend and one to remove, given half the
mistakes and White won anyway, but it is
chance.
worth noticing how this happened. Black was
34.Sd6 Sc7 35 .Sa6! would lead to a repetition. unable to act with energy in creating the one
thing that would give him counter-chances: a
34.. .d4 35.Sd5 Sc7 passed pawn.
270 Conceptual Rook Endgames

Correct was 37. f3! Sg2 (37...Se2t!? 38. &f4! is 40.Ha5!! g4

similar) 38. &f4 Sa2 39.Sa5 and White wins a Black cannot take on g3, as the white king is
vital pawn. close enough to stop the h-pawn.
41.a4 Sxg3 4 2 . S g 5 t h6 43. b 5 f6
37...Sx£2 38.Sxa7

An important tempo move, as the rook is

abcdefgh not anticipating assisting in promoting the
b-pawn.
38...g5?
38...gg2! 39. b5 (39.&f3 Sb2! would slow 44.Sc5? Sa3! would hold.
down the white pawns sufficiently.) 39...Sxg3 44...Hb3
40. <&xd4 g5! Material is often a less important 44... <&h7 45.Sc8 Sa3 also loses spectacularly.
White plays: 46. b6 Sxa4f 47. &C3! Avoiding
factor in endgames than people realise. Because
a mess with two new queens on the board.
endgames are centred around promoting
pawns, creating passed pawns is a high priority 47.. .5a3t 48.&C4! White wins. For example,
after queening first and picking up the new
in most situations. Black makes a draw with a
black queen with a check from behind.
small, but significant, margin.
45. c4 S b l 46.a5 &h7 47.Se8 g3 48.He2 f5
49.b6 f4
39.hxg5?
After 39. <&xd4! Sg2, White would have a
rather beautiful winning variation, with lots of
small tempo-gains, at his disposal.

abcdefgh

abcdef gh
 Chapter 1 9 - Tempo Wars 271

5O.Se5!!
White wins additional time. The threat of
Sxh5t followed by Sg5t is enough to slow
Black down. And after 50... hb 51.Sg5! £3
52.Hxg3 f2 53.SF3 fl =Wf 54.Sxfl Sxfl 55.b7
the white pawns sail through.

abcdefgh
44.&c3!!
By kicking the rook with the king (which
would be happy anywhere), White ensures
that the a-pawn will win a tempo as it lunges
up the board!

abcdefgh
44...Sbl 45.±c2 Sb6 46.a4 h4 47.a5 Sb5
39...&g6? 48.a6 h3 49.Sa8 h2 5O.Sh8
Using four moves with the black king to take 1-0
the g3-pawn is an incompetent waste of time,
only forgivable because of the lack of time
Black had to make this decision. White won
with a nice little tempo-gaining finesse.

39...5g2! would drive the white king to

a worse square. 40.&f3 (40. b5 Sxg3 and
41...5b3 would leave the h-pawn to make
the draw - which it would.) 4O...Sb2! 41.Sa5
The connected pawns are really slow, when
controlled. 41...Sb3t The simplest. 42.&f4
d3 43.&e3 d2f 44.<&xd2 Sxg3 The h-pawn is
hard to get to. Black makes the draw.

40.b5 ±xg5 41.b6 Sb2 42.b7 ±g4 43.&xd4

&xg3
Chapter 20
Anticipation
One of the most basic tools I use to understand chess is that of The Three Questions:

Where are the weaknesses?

Which is the worst placed piece?
What is the opponent’s idea?

The value of these questions is focus. Chess games are decided on weaknesses. We need to make
the most out of our pieces. And we need to understand the opponent’s ideas as well as find ideas
for ourselves, yet we frequently forget.

Anticipation is a sub-section of prophylaxis, which is the understanding of the opponent’s ideas,

with various levels of reaction to them, scaling from full-on prevention to complete apathy. With
anticipation, we are looking at small finesses, often deep into the variations. Understood as such,
this is a deep and irrationally random way of looking at deep calculation - and as such, not really
useful. But let’s add one of the core pieces of advice I give students on decision making with the
examples in this chapter, and a path to using this way of thinking constructively will appear.

Detlev Fischer - Ing. Jan Matusek

email 2012

abcdefgh
274 Conceptual Rook Endgames

In our first example, White is in danger. He 67.Sf7!

is a pawn down and has to fight for the draw. This is the alternative, and much preferable
Using the three questions (which is easier to in reality. In a correspondence game this is
use when there are more pieces on the board), perhaps not even noticed in this modern age,
we get these observations. The f-pawn is a but it goes against the common instinct of
weakness for both sides. Its threat to promote is extending the checking distance as much as
a disability for White, and Black wants to keep possible.
protecting it. White’s king is cut off and the
black rook on f4 is awkward and not as active 67.. .5Bt 68.&d4
as it should be. Thus, Black’s idea is ...Sf3t and 68.&d2? would lose to 68. ..f4 69.Sf8 &g2.
...&g3, followed by ... f4, or ...Sfl, ...f4 and
other ways to advance the pawn without the 68.. .f4
rook getting stuck in front of it. Sometimes it is hard to guess why a game is
played out, in this case until stalemate. I am
67.Hg6? would allow the rook to escape the slowly coming round to the idea that all games
awkward position. Simplest is 67...2a4. should be played out, and the draw offer
should be put in the bin of outdated artefacts.
67. &d2? fails similarly to 67...2d4f 68. &c3 The key point is that 68...&g3 no longer
Sd5 69.&c4 Sa5. works wonders. After 69. e5 He3t 70. <&xf5
Sf3t 71. <&g6 the rook is protected.
But the most important point comes after
everyone’s first idea: 67.2f8?, where Black
69.SB ±g3 7O.Sg5t ±h4 71.±e4 ±xg5
plays 67. ..Sf3t 68. <&d4 <&g3 as intended. 72.±x6 ±6 73.±g2 ±e4 74.±£2 ±6
If the king can come in front off the pawn, 75.±e2 ±g4 76.±£2 6 77.± fl ±f4 78.±£2
keeping the white king cut off, Black wins. ±e4 79.±fl ±e3 8O.±el ±d4 81.±£2 ±e4
Which he can, due to 69. e5 Be3*|T 82.±fl ±e3 83.±el £2f 84.±fl ±B
Vi-Vi
If we take the following practical advice to
heart, we will have a real chance of finding the
This idea is also surprising to players at the top
extra finesse in this position - and all others
level.
in this chapter: “Always try hard to find two
serious ideas in all positions.” Moving the
king and moving the rook sideways were not
serious ideas, as we saw above. If these are the
alternatives you find, you have not tried hard
enough. Failing to consider the best option on
move one is an obvious weakness for our play,
but also a common one. Even among strong
players. Here, it would lead to one alternative.
We can then try to find the difference. This is
commonly referred to as comparison - when
we look at one variation, but have one piece
that could be on either of two squares. We
don’t calculate it as different lines, but as one,
with an alternative reality attached to it.
 Chapter 20 — Anticipation 275

Viswanathan Anand - Nodirbek Abdusattorov usually good at applying different approaches

to different scenarios with success. In this
Internet 2020
case, noticing that there is a quirky alternative
is enough as a start, meaning that there is a
reason to think. We would then calculate the
outcome.

Piotr Piesik - Sabina-Francesca Foisor

Stockholm 2019

abcdefgh
Black is threatening mate in one, but more
importantly, also to meet 59.a7? with a simple
perpetual check. There are only two ways to
avoid this, so Anand played the natural-looking
59.Sal?, when after 59...Sb5t 60.&a4 &b6,
due to the threat of ...Ba5t, he had to abandon
the a-pawn and soon after the hope of winning
abcdefgh
the game. The first line we would see would most likely
be the way the game went:
59.Ha2! would have won the game. After
59...Hb5t 6O.£a4 <£b6, White has 61.a7!, 58...&b2 59.d7
when the rook is within reach. After 61...Sa5t We quickly see that going with the rook to
62. b3 Hxa2 63. <&xa2 &xa7 64. b3 b6 a3 is not a serious option, because of a check
65.&c4 <&c6 66.&d4 d6 67. h4 the pawn on a8.
ending wins trivially.
59...5d3 6O.sl?e6 al=W 61.Hxal xal 62.f5
This principle can be applied to many positions. White is threatening to put the pawns on the
In the following, White is threatening to take 7th rank and zigzag the king back in response
the a2-pawn (with check), leaving Black with to checks, never allowing the rook to make it
seemingly only one reasonable move, which to the 8th rank. Black thus has to start giving
will defend the pawn, help advance the pawn checks now.
and prepare to block the access to the pawn
with the rook. There is no such thing as a 62.. .5e3t 63.&f7 Sd3 64.<&e7 Se3t 65.&f8
perfect way to calculate — and especially not Gaining time. The king belongs on d8.
one that is equally effective in all types of
positions. But human reason and experience is 65.. .5d3 66.&e8 Se3t 67.&d8 SB
276 Conceptual Rook Endgames

abcdefgh abcdefgh
Hopefully we would reach this position The point is that 63. dS can be met with
in our calculations without difficulty and 63. ..Hf3!. Black has not spent a tempo winning
carefully check if there is a way for White to the rook and the f-pawn is still on the 4th rank,
progress. There is. a step too far from the f6-trick seen above.
White can repeat with 64.&e7 (and should),
68.ft>! ±b2 but can also choose to go down the drain with
68...Hxf6 69.&e7 and the d-pawn queens. 64.&c7 Hc3t 65.Hc6? a l =©! 66.d8=® ®a5t
and Black wins.
69.&c7 Ec3t 70.<&b6 Sb3t 71.&c5 Sc3t
72.&d4 &c2 73.f7 Sd3t 74.&e4 Sxd7 White thus has to find:
75.fl8=® Sd2 76.Bc5t £dl 77>c4 Se2f
78.&d3 Sd2f 79.&e3 Sb2 8O.®d3t £cl 63.Se6! al=W
81.Wc3t £ b l 82.<&d3 &a2 83>d4 <&b3 The queen ending after 63...Hxe6f 64.&xe6
84>c3t &a2 85.Wa5t & b l 86.®a4 gh2 a l = ® 65.d8=® is lost. “The further the pawn
87>b3t Sb2 88>dl| &a2 89.&c3 Sh2 is from the centre, the better the drawing
90>a4| £bl 91.®b5t &al 92>a6f &bl chances” is one oversimplification of when
93.Wb6t &a2 94>a7t the endgame is holdable and when not. The
1-0 bishop’s file is a not.

Having gone through this variation, the natural 64.Sxe3

thing to do would be to check if we missed White is temporarily a queen down, but the
something on the way. Having noticed the pawn is getting close.
usefulness of the checks later in the variation
(despite its failure), we may reasonably come 64...Wa7 65.Se5!
up with the idea to give them earlier.

6O...Se3t!? 61.±f7 Bd3 62.±e7 Se3t

 Chapter 20 - Anticipation 277

58...£bl!

abcdefgh
The rook controls a lot of key squares from
here.
abcdefgh
Although this looks a bit odd and otherwise
65...&c3 66.&e8 ®a4 67.f5 ®d4 prone to transposition, we should not
After 67... <&d4 68.Se7, Black cannot stop disregard the move without investigation. We
the advancing f-pawn. already know that the key advantage of putting
the king on b2, the ...Ba3 ideas, would not
68.Se6 Wh8f 69.&e7 Wg7t 70.&d6 Wd4f work anyway. And surely enough, after going
7L±c7Wf4f 72.Sd6 through the same line with the king on b l
White wins. During the game it would be slowly, we notice one advantage:
difficult to say with certainty if Black has a
draw or not in this line, something which may 59.d7 Sd3 6O.±e6 Se3t 61.±f7 Sd3
be hard to emotionally understand when we 62.&e7 Se3t 63.Se6 al=W! 64.Sxe3
know he does not. But certainly, it looks like a
better try, due to the lack of clarity. If nothing
else, there is room for mistakes.

But we should not forget that there was a

second option on move one.

abcdefgh
64...Wg7t!
The king is not blocking the long diagonal
and the queen comes into the game quickly,
securing the draw. For example:
278 Conceptual Rook Endgames

65.±d8 WfSt 66.Se8Wxf4

White cannot hide from the perpetual.

Noticing this type of small difference does not

come easy. It takes experience to get a feeling
for how minor differences can lead to vastly
different outcomes. A student at the academy
expressed how he desperately had tried to find
options for the queen in the line with the king
on b2, but never even considered that more
options would occur with the king on b l .

In the next example, White seemingly only has abcdefgh

one sensible move. Areshchenko played it and Without the e-pawn, Black would not win.
lost a tight race. But it is there.

Alexander Areshchenko - Niclas Huschenbeth

54.. .e5t! 55.±g4 ±£2 56.Sh3
Germany 2021 56.Sf8t & g l 57.Sh8 g2 wins.

56...e4
0-1

Let’s imagine we had calculated the outcome

of the race to a sufficient level of accuracy to
be concerned about its outcome. This does not
require us to see everything clearly to the end,
only to recognise the feeling of “hope” related
to how the line is functioning. This should
inspire us to look for a different direction.

The first one that comes to mind would be

abcdefgh to start the race with advancing the king.
However, this does not work out well:
43.b7?
43.Sg8? g6 would make no progress at all.
43.&e4? ±g4 44.±e5
44. b7 xg3 and 44.Sg8 g5 are both
43...Sb3t 44.&e4
nonsensical options.
44.&F2 &g4 followed by an opportune
check on the 2nd rank was hopeless as well.
44...g5
44...±g4 45.&e5 &xg3 46.±d6 h4 47.&c6
g5 48.Sg8 g4 49.b8=Wf Sxb8 5O.Sxb8 h3
5L±d6 h2 52.Sh8 ±g2 53.±e5 g3! 54.±f4
 Chapter 20 - Anticipation 279

47...&B!
Black wins on account of 48.Sxh5 Sb5t-
(47. ..&h4 48.Sgl Sb6 also wins, but looks
rather clumsy.)
0-1

All of this, and a persistent search for additional

options, should lead us to a peculiar move,
which like we have seen before.

43.Sb7!
The rook is seeking to go behind the black
pawns, but not on the 8th rank, where it has
45.b7 no secondary function.
It is hard to find alternatives.
43...g6
45.&f6 loses to 45...e5! 46. *xe5 xg3. After 43. 44.&e4 g5 45.< >d4, White
would be able to use the many free tempos to
45. xe6 *xg3 is clearly dead lost. make a draw in the race.

And 45.Sg8 Sxb6 is depressingly far from 44. &e4 &g4

serious. Black can play waiting moves, but White is
happy enough to play 45.&d4, heading for the
45...&xg3 46.Sg8 Sxb7 queenside.
The difference from the previous line is that
the white king is bang in the centre and Black
cannot afford to lose the rook.

47.Sxg5t

abcdefgh
45.Sg7!
The point.

abcdefgh
280 Conceptual Rook Endgames

45...g5 46.b7
With the pawn on the 7th rank, White
will make an effortless draw. The rook to the
side of the pawn does not help advancing the
pawn, but this is not too important, as the job
of the pawn is not to advance, but to demand
complete attention from the black rook.

Makan Rafiee - Armen Ar. Barseghyan

Germany 2022

abcdefgh
3O.b3
White has a winning position. Although all
can see this is a better version, it makes sense
to show a plausible path towards the full point.

3O...Sc8 31.Scl &xh7 32.Sdxc6 Sxc6

33.Sxc6
The black king is too far away.

33...Sg7 34.SFS Sc7 35.&b2 &g7 36.SB

abcdefgh a5 37.&a3 &g6 38.&a4 Sc5 39.a3
White is in check and has to find the right White wins.
space for the king. This requires calculation
and anticipation of the opponent’s intentions. The following example is quite deep.

28.£c3? Dmitry Bocharov - Vasily Usmanov

This looks natural, but failed.
Sochi 2017
28.£ c l ? fails for the same reason.

28...exd6 29.Sxd6 Sb8 3O.Sxc6

30. b3 Sxa2! is a key point. So White tried
something else and got nowhere.

30.. .5gxb2 31.Sc8f Sxc8f 32.&xb2

33.&b3 Sf4
The game was drawn 20 moves later.

The lesson from the game . was that White

needed to defend the a-pawn:

28.&bl I! exd6 29.Sxd6 Sb8 abcdefgh

 Chapter 20 - Anticipation 281

Failing to find a win in the rook vs pawn 66.&dl!I

endgame, White settled for: This surprising move is the only path to
victory.
64.Sxh4? a3 65.Sh3 a2 66.g7 al=®
67.g8=®t &c2 66.&d2? h2 67.Sb7t (67.Sg3t &b2! would
He tried to win this endgame for 52 moves keep the white king at bay.) 67. ..&a2 68.§h7
without success. (This is a nice trap, but Black does not have
to fall for 68. ..a3?, with the idea that 69.&c2?
You could argue that from a practical point of h l = ® leads to stalemate, but also without a
view, White should have won the bishop and reply to 69.&cl!.) 68. .. b2! Shouldering is
see if he was able to find something later on almost always the answer. 69.Sxh2 a3 Black
that eluded him at this point. I often discuss holds the draw.
this topic with top rated students. Although
they have fantastic calculation skills, they often Black also makes the draw after 66. &d3? h2
give up on variations they cannot completely 67.Sb7t &a2 68.§h7 b3!, preventing the
close down and chose something they do not white king from coming to c4.
believe in, rather than allow intuition a larger
say on which path to choose. Sometimes we 66.. .h2
should trust ourselves a bit more. As in this 66. ..a3 67. & c l ! and wins.
case.
67.Sh7
64.g7 &xg7 65.Sxg7 h3 We now see the key point. 67... &b2
65...a3 is hopeless and loses to 66.&d3 and 68.§xh2f is a useful check.
67.Sb7t.
67.. .a3 68.&C1! a2 69.Sh3t
White wins.

I am not sure to what extent Deac made the

next move with the hand rather than the
head in the next example. But it is easy to
imagine that he never even considered the only
winning move. It is not something you would
consider without understanding the principle
of anticipation and having seen examples such
as this... which is always a problem: you only
understand afterwards how to think in such
a position, if you have not had the theme
abcdefgh explained in advance. To make this a paradox:
If Deac had read this book, his game would
not have been in it...
282 Conceptual Rook Endgames

Haik M. Martirosyan - Bogdan-Daniel Deac

Internet 2020

abcdefgh
68.e7??
You feel the time trouble in this move. White
abcdefgh was definitely in all types of other problems,
but there was a way out.
65...Sei?
65...Sg5? would also be mistaken. After
After both 68.&d6? S d l f 69.&e7 g5! and
66.Sf7l! the black king is cut off and White has
68.Sh8? S d l f 69.&c4 <&xe6 7O.Sxh4 g5
counterplay. Curiously Black has to quickly
White would have lost without a fight.
get the e-pawn under control, as he would lose
after 66. ..h4? 67.Sf8!.
But he could have played: 68.Sf8f! &g4
69.&d6 h3 7O.Sh8! (7O.e7? h2 71.Sh8 &g3
Everything flows naturally from this point on.
wins for Black.) The key point is that 7O... g3
is met with 71.Sg8!, with a draw. And after
66.&d5 h4 67.Se8 &f5
70...g5, we see that this pawn move is far
Black would be able to eliminate the e-pawn
less helpful. White draws with: 71.e7 &g3
with 67. ..h3!? 68.Sh8 &f5 69.Sxh3 S d l f
72.e8=® Sxe8 73.Sxe8 g4 74.&e5 And so
7O. c4 <t£?xe6, but the drawing method
on...
for White is well known to us. 71.Sg3 &f6
72.Sf3t &e5 73.Sg3 Sd7 74.Sgll! (See
68...6f6! 69.&d6
Chapter 10 on Checking Distance, page 163).
69.Sh8 Sxe7 is equally hopeless.
It is possible to argue that 67... h7!? was
the best practical chance. After 68.SF8! Black
69.. . 5 d l f 70.&c6 <&£7
The white king is cut off and Black wins.
cannot play 68. ..h3, due to 69.Sf3!, but after
68. ..g5, White has to find 69.Sf5! &g6 7O.Se5
71.Sd8 Sei 72.Sh8 g5 73.&d6 Se6t
Sxe5t 71.&xe5 h3 72.<&d6, where he reaches
74.&d5 Sxe7 75.Sh7t 76.Sh8 &f5
a queen ending, which is really difficult to
0-1
draw in practical play.
 Chapter 20 - Anticipation 283

The winning move was simple to explain, Aagaard

but hard to anticipate - without the idea of
Original 2021
anticipation.

65...Se2!! 66.&d5 h4 67.Se8

abcdefgh
White seems to be unable to do
abcdefgh anything. If he plays l.c4?, Black will play
l...h4!, and we can easily see which of the
Black now has a choice of three winning
pawns has the better prospects. Therefore,
lines. In all three of them, the rook on e2 plays
most would follow my hand and give checks.
a crucial role in defending the pawn on h2:

l . S g 7t? &h3 2.Sf7 &g2 3.Sg7t 4.Se7

67...h3 68.Sh8 h2 69.&d6After 69*7 Sxe7!
The queen ending after 4.§b7 h4 5-Sb8
7O.Sxh2 £f5 71.Sf2t &g4 72.gg2t £f4!
£gl 6.a8=B fl=Bt 7.&c2 Exb8 8.Bxb8
Black wins, because the white rook is on the
h3 is entirely hopeless, mainly due to the
2nd rank and not the 1st. 73.Sg6 £5 74.§gl
unfortunate placement of the white king. The
g5 75-Sfl t &g6 76.Hgl Sh7 77.&e4 Sh3 is
c-pawn does nothing to help White.
the concrete variation. 69.. .&ft> 7O.Sf8f &g5
71.Sh8 &g4 72.e7 &g3 73.Sh5 g6! Black
4...Sd8t 5.&c2 h4 6.Sh7 h3 7.Sxh3 &g2
wins.
8.Sd3 Sa8
Black just has to avoid 8...fl=®? 9.Sxd8
67.. .6f5!? 68.Sfl8t &g4 69.&d6 h3 70.SE8
®f5t 1 0.§d3, when White will build a fortress.
h2! 7 1 .e7 &g3 Black wins. The g-pawn is
coming.
9 . S d l Sxa7
Black wins by cutting off the white king,
67.. .6h7 68.Sf8 h3 69.SB h2 also sees the
or because it has to go to the b-file (no
pawn defended on h2. After 7O.Sh3t &g6
shouldering!) to avoid that.
71.&d6 72.SBt &g5 73.Sh3 &g4
Black wins again.
White needs to deal with the manoeuvring of
the king to e l , which is the winning idea.
The following position came up in my analysis
of Maxime Vachier-Lagrave - Wesley So,
Internet 2020.
284 Conceptual Rook Endgames

l.§£6!! h4 2.Sg6f &h3 3.Sf6 &g2 4.§g6t The next example is beautiful.
£fl
Nicolas Brunner - Francois Godart

France 2022

abcdefgh
5.Sa6! ga8
5...‘£>el?? is impossible due to: 6. S a l # abcdefgh
We enter this game where Black missed the
6.&e3! h3 7.Sa2 &gl 8.Sx£2 Se8t 9. £3!
chance to force a draw by setting up a mating
Sf8f
net around the white king.
9...h2?? would even lose: after 10.Sg2t
£hl l l . S a 2 Sf8f 12. eS Sa8 13.&d4 & g l
33.. .5A8?
14.&c5 the king arrives on b6, in time to
33... f7!! 34. h5 The king could retreat and
defend the a-pawn and support the advance of
take all his hopes of an advantage with him.
the c-pawn.
34.. .5d4 35.h6 e7! Threatening ...Se4#.
36. S e i Sh4 37.b3 Sh5 38.&f4 Sh4f 39.&f3
8 &f7 White has no advantage to speak of.
7
34.b3?
6 Lazy...
5
34.Sh3!! would have anticipated 34...f4, when
4 35.Sf3 is winning.
3
34.. .f4 35.SA B 36.&e4
2
Paying the price.
1
36.Sf2 h5 and Black draws by a tempo.
abcdefgh
10.&g3 Eg8f l l . £3 h2 12.Sxh2 &xh2 36.. .E2 37.&e3 Sh8 38.Sx£2 Sxh4 39.Sg2
13.&e4! Sd4
The c-pawn secures the draw. Black holds.
 Chapter 2 0 - Anticipation 285

Let’s dive full into the other side of the looking 48.SBI?
glass... In the next example Black is a pawn I don’t like passive rooks, although here it is
up, but converting it into a full point proved good enough to draw.
difficult.
48.Sgl! Sxf4 49.S a l was cleaner.
Evgenios loannidis - P. Iniyan
48...B5 49.SA a6 5O.Sal Se3t 51.&xd4
Plovdiv 2021
Se4f 52.&d5 Sxf4 53.Sxa6 Sxb4 54.&e5
&g7 55.&x£5 Sxb2 56.h4 Sc2 57.h5 Sc5t
58.&g4 Sc4f 59.&f5 Sc5t 60.&g4

Let’s have a look at the alternatives.

43...h5I?
This looks strong and would require an
amazing degree of accuracy of White to hold
the draw:

44.Sh3 &g6 45.Sg3t &h6 46.Sg5 Se3t

abcdefgh 47.&c4! h4 48.Sxf5 d3 49.Sd5 Sh3
5O.Sxd3 Sxh2
43...Se4? 44.Sh3! &g7
44...§xf4 45.Sxh7 also gives a lot of
8
counterplay.
7
45.Sg3t! 6
The checks are annoying.
5
45...&T6 46.Sh3 &g7 47.Sg3t &h8 4
3
2
1
abcdefgh
51. d4!!
Another beautiful piece of anticipation.

51.' ’d5? I ’h5 52.Sa3 Sf2! and Black wins.

abcdefgh
286 Conceptual Rook Endgames

51...6h5 The winning move was the cryptic:

51...2xb2 52.&e5 with a draw.
43...h6!I
52.Sa3 S£2 53. e3!
Winning an important tempo.

53...5xb2 54.Sxa7 Sb3t 55.&e4

But the king was always meant to go forward.

55.. .5xb4f 56. >f5

abcdefgh
44.Sh3 &g6 45.Sg3t &h5
With no check on g5, Black wins.

46.Sh3t &g4 47.Sxh6 &xf4 48.h4

48. b5 Se5! wins.

abcdefgh
48...Se3t 49.&xd4 Se4f 50.&c3 &e3
White holds. The threat of mate and the The f-pawn decides the game. Notice that, as
passed f-pawn offer enough counterplay. in so many conversions of an extra pawn in the
rook ending, the pawn is given up in order to
have the rook and the king being able to help
the stronger passed pawn race up the board.

In the next example, Black looks at first glance

to be a big favourite. He is a pawn up and it
is quite advanced. But the problem is that his
king is cut ofFand the white king has potential
counterplay with b4-a5, attacking the black
queenside pawns. So, is White close to making
a draw after all? In the game he lost quickly.
 Chapter 20 - Anticipation 287

Vladislav Krasikov - Sergei Lobanov White’s drawing move comes right out of the
left field, but makes sense once you have seen
Moscow 2022
the variations above.

45.a3!! S d l
Still the only try.

46. Bxh3 Bxd4 47.&c3 &d5 48.Bh8 e5

49.Sa8 Ba4 50.&d3! e4f 51.&c3
White holds. Narrowly. More moves can
definitely be tried.

In the game below, White is clearly fighting for

a draw. The white rook is lacking a way to deal
with stopping the g-pawn and defending the
a-pawn. White lost quickly.
abcdefgh
45.&b4? Bdl! Danyyil Dvirnyy - Nils Grandelius
Black has to switch strategy.
Jerusalem 2015

46.&c3
A key variation is 46.Bxh3 Bxd4f 47. c3
*d5 48.Bh8 e5 49.Sa8 Ba4, when Black is
threatening both ...Bxa2 and ...&e4.

46.. .Belt!
Exploiting the odd positioning of the king.

47.&d2
47.&b3 Sc4 48.Bxh3 &d5 49.Sh6 Bxd4
also wins for Black.

47.. .Shi 48.&c3 a5 49.a3 h2!

abcdefgh
Played the moment where White will get
into zugzwang and lose the control of the 5th 81. e6?
rank. Looking for a long checking distance with
81.Bb5? does not work. The king can hide on
50.&c2 a4 51.Sh3 &d5 52.Bh4 &c4 h3.
Zugzwang.
81...g3 82.Sf5t &g4 83. fi6 g2 84.Sg5t
53.b3t axb3t 54.&b2 &d3
0-1 White resigned. The king is too far away to
assist the a-pawn later.
288 Conceptual Rook Endgames

The drawing line is nothing but sensational.

White has to find a way to draw with the
a-pawn after giving up the rook. And as no
straightforward path exists, he has to find a
crooked one.

81.Sd5!!

abcdefgh
85.Sd5!!
The great moment of anticipation. The rook
defends the a-pawn and stops the g-pawn. If
the pawn is queened, White gives up the rook
for it. But there is nothing else Black can try.

Move 40 is a terrible place to have to make the

abcdefgh biggest decisions. White did not get this one
We shall soon see the reason for the rook right...
belonging on the d-file.
Yuval Yaniv - Ediz Gurel
81...g3 82.&c6! g2 83.§d4f &g3 Cala Gonone 2022
The critical try.

Both 83...&e3 84.Eg4 and 83...&g5

84.Sd8 lead to immediate draws. And
going forward just leads to an extra check.

The reason the rook belongs in the d-file is that

after 83. . .&f5 8 4 . § d 5 t there is space for all the
checks. Black will have no other idea than to
hide on h2.

84.Sd3t >h2

abcdefgh
4O.Sd2?
You flip a coin and it ends on its side...

4O...a5?
Move 40 is equally difficult for Black!
 Chapter 20 - Anticipation 289

Black misses a chance to hold with: 4O...Sc6!

41.gc2 a5 42.&e3 a4 43.&d4 a3 44.£c4 Sa6
45.Sa2 &d7 46.&b5 Ha8

41.&e3?
Move 41 is often another place for mistakes.
Especially if you are White. The adrenaline
from time trouble is clouding your mind and
you cannot easily recompose.

White was winning after:

41.Ha2! Sa6
abcdefgh
46...hxg4?
A horrible mistake, giving White a distant
passed pawn for no reason whatsoever.
46...f6 would have held and is not surprising
at all.
If 47. g5, Black takes and pushes the h-pawn.
And after 47. gxh 5 gxh5 48.f4 he needs to
find 48...Sa6! when White is unable to make
progress. But this is not so difficult, as White’s
abcdef g h only ideas are c5-c6f and e4-e5. Now after
49.c6t Sxc6 5O.Sxa3 Sd6t 5I.&c4 Sc6t
42.Sa4l!
52.&d3 Sd6f 53. &e3 Sb6 Black will still
Blocking the pawn as early as possible, as we
have to find a few moves, but hold he will.
have previously discussed.
54.Ha7t &e8 5 5 . W Sb3t 56.£d4 gf3 57. f5
42...&d7 43.&e3 &c6 44.&d4 & b 5 45.Ha3 &f8 58.Sxh5 &g7 is the most entertaining
Luring the black king forward. White is
variation. The white rook is in house arrest.
ready to play &d5-d6-e7. Now after 45...&b4,
And after 59.e5 fxe5t 60.&xe5 Se3t Black
White wins with 46. Hal a4 47.&d5! and the
holds this often seen theoretical endgame,
c-pawn will roll.
which you can read all about in Theoretical
Passive play with 45...&c6 loses to 46. g4
Rook Endgames by Sam Shankland.
f6 47. g5 fxg5 48.hxg5 h4 49.f4 a4 50. f5 and
White arrives first. 47.fxg4 ga6 48.e5 ga4 49.h5! gxh5 5O.gxh5
ga6 51.h6 gxh6 52.gxa3 g h l 53.c6f &d8
41...a4 42.&d4a3 54.Ha8f <&c7 55.Ha7t &d8 56.Hxf7 g d l f
Black’s drawing chances have increased 57.&e6 g e l 58.gd7t &c8 59.&d6 g d l f
greatly by having a passed pawn on the 3rd
60.&e7 gel 61.c7 g h l 62.e6 g e l 63.gd2
rank. gfl 64.gc2 Hel 65. f7 gflf 66.<&e8 Sei
67.e7 gfl 68.gc4 g£2 69.Hg4 &xc7 7O.gc4f
43.&d5 Ha6 44.ga2 <&d7 45.g3 Ha5 46.g4 <&b7 71.&d7 Sd2f 72.<&e6 ge2f 73.&d6
g e l 74.gc5
1-0
290 Conceptual Rook Endgames

By knowing what is coming, we can find the Nico Zwirs - R. Ashwath

right square for the rook:
Hoogeveen 2022

4O.Sd3!! Sc6 41.Sc3 a5 42.&e3 a4 43.&d4

Sa6 44.Sa3!
The rook is able to stop the a-pawn earlier
and after:

44...&d7 45.&d5

abcdefgh
55...&g7?
55...Sd7? loses to: 56. &f3 &g7 57.Sd3!
Preparing &e4, when the rook and the d-pawn
are both defended. 57. ..c6 58.d6 &f6 59.&f4!
c5 (59...&e6 60.&g5 Hxd6 61.Hxd6t &xd6
abcdefgh 62.&xg6 c5 63. &f5! and White wins.) 6O.Hd5!
&e6 61.Hxc5 Hxd6 62.&g5 White wins, but
White is winning.
after a good deal of moves. Yet, this was enough
to convince Sam of the inevitable result.
45...&c7 46.g4 ft> 47.g5! fxg5 48.hxg5 h4
49.f4
56.Sc4 Sd7 57.Sc5 58.f4 Sd6
And so on. The key point here is about
58...&e7 59.&f3 &d6 6O.Hc6f and after
time. And about passed pawns being a bigger
White takes the g-pawn, the h-pawn is a great
influence the closer they are to the 8th rank.
liability.
After:

49...h3
White is in time to play:

5O.Sxh3 a3 51.Shl a2 52.Sal

Winning.

I gave the final position of this chapter to Sam

Shankland in a training session. He solved it
by the method of elimination. Meaning, he
tried everything... and I eliminated them.

abcdefgh
 Chapter 20 — Anticipation 291

59.&B? 64.. .§(8 65.Sxg6 &h3 66.d6 h4 67.f5 &h2

Giving Black an undeserved chance. 68.d7 h3 69.Sd6 Sd8 70.&B
1-0
59.Hc6! won. 59...Sxc6 6O.dxc6 &e6 61. f3
&d6 62.g4 hxg4f 63.&xg4 &xc6 64. &g5 The drawing move borders to the land of the
&d6 65.&xg6 c5 66. h5 c4 67. h6 c3 68. h7 absurd:
c2 69.h8=® c l = ® 70.®d8t &e6 71.f5t &e5
72.f6 with a winning endgame. The guideline 55...6h6!!
for these endgames objectively and practically It turns out that Hxc7 cannot come with
is different. With the h-pawn it is a draw - and a check, and that the g-pawn still needs
you may hold it in classical, but not likely in defending. So, the king hides. Black’s play is
rapid chess. With a g-pawn it is a narrow draw now ...Hd7-d6 followed by ...c6, eliminating
and you will not hold it in practice. With the his weakness.
f- and e-pawns, it is a win. The long checks are
too few. 56. f4
56.Hc4 can now be met with 56...Hf5!
59...&f5 60.&e3 &g4 61.&e4 57. Hc5 c6! and Black draws.

56...Sd7 57.&B

abcdefgh
61...&xg3?
Black could have drawn with an exceptional
abcdefgh
manoeuvre, drawing on everything we have 57...c6!
talked about throughout this book: 61...Hd7I! 57 ...2d6? is too slow now. After 58. e4 c6
62.Hc3 Se7t 63.&d4 &h3! 64.&c5 S e i 59.&e5! White wins.
(64...5d7!? is less natural to me, but works
too.) 65.&c6 Hdl and White has to give up 58.d6 &g7!
the kingside to make progress. The black king returns at the right moment.

62.§xc7 63.Sc3t &xh4 64.§c6! 59.&e4

The key move. White wins back the pawn Black holds.
and cuts ofF the king.
292 Conceptual Rook Endgames

The topic of anticipation is a difficult one and clearly builds on a lot of the knowledge and ideas
from previous chapters. The most valuable idea you can take from this chapter is that the rook
does not necessarily have to go to the end of the board in all situations. Look at what happens if
it goes only part of the way too...
Chapter 21

Connected Passed Pawns

As this book is nearing its end (don’t check how many pages are left; I promise we will get
there...), we get into more and more hazy topics. The following positions with connected passed
pawns are not followed up with strong unique concepts, but rather deeply complicated positions,
where you will see themes from previous chapters appear with a clarity you would not otherwise
be able to see.

The first game comes with a bit of a story. Martin is a friend and in the mid-90s we were discussing
a lot of chess and hung out. We both wanted to make it to the IM title at the time and were
fairly clueless about how to go about it. The game in question was played in the Championship
of Copenhagen and had gone to adjournment after a terrible first 40 moves for Martin. For some
reason or another I became obsessed with analysing this position and found a lot of ideas. At the
time I may have believed that I had managed to find a draw for White, which is obviously wrong.
But the things I showed Martin were of great practical use once the game was resumed.

Martin Olesen - Jorgen Hvenekilde

Copenhagen 1996

abcdefgh
41...Sxh3 42.Sc7
294 Conceptual Rook Endgames

Focusing on my key idea, which worked

8
wonders in the game.
7
42.. . 5 h l 4 3 . & d 4 S b l ? 6
Push the passed pawns! For example:
43. . .h5 44.Sc6 Sxc6!? 45.dxc6 E c l 46. &d5 g4 5
and Black will arrive in very good time. 4
44.Sc6 3
Threatening to push the d-pawn, but this is 2
not the main intention.
1
44.. .5b8? abcdefgh
I guess Jorgen was heading for a8 to push 5O...Sbl?
the a-pawn, but this is a very time-consuming 50...f5?! is tempting, but the white c-pawn
manoeuvre. will still reach the 7th rank with a tempo,
making the tandem incredibly fast. 51.d6 f4
52.c7 Sc8 53.Sd2! f3 54.&c6 f2 55.d7 Sxc7t
56.&xc7 fl=® 57.Sd6t! & h 5 58.d8=®
should hold for Black, but there are a lot of
pitfalls on the way to the coveted half point.

5O...Sb3!
This was the strongest move, putting White
under pressure. Only one move holds:

abcdefgh
45.S£2!
The key idea of my analysis. White suddenly
has connected passed pawns. And as Black has
graciously refrained from advancing his pawns,
they are fairly dominating.

45...Sxc6 46.dxc6 g4 abcdef gh

However, this does not mean that White is 51.&c4!
out of trouble. The idea is not to attack the rook, but to
avoid being in check when the pawn reaches
47.£c5 h5 48.d4 g3 49.Sg2 h4 5O.d5 g1 -
51.c7? Sc3t would allow the black king to
join in and decide the game.
 Chapter 2 1 — Connected Passed Pawns 295

51.&d4? h3 52.Sxg3t Sxg3 53.c7 h2 We have previously seen the idea 54.a5 Hxh4
54.c8=® h l = ® (54...£g7? 55.®f5! and 55.Hb3!, which would have won on the spot.
White holds.) 55.®g8f &f6 56.®xg3 Sadly But even without this, you should consider
for White, his pieces are unfortunately pushing the furthest advanced passed pawn
placed, and he will have to exchange queens first anyway.
(no!) or accept that after 56...®alf! 57.&c5
®a5t 58.&c6 ® b 5 t the d-pawn is lost and 54...Sxh4 55.a5 Sh2f 56.<&g3 Sa2
Black will win the queen ending. Although the white pawns are further
51...h3 advanced, Black is already better. The white
Black has nothing else. pawns will be clumsy to advance, while
52.c7! hxg2 53.c8=® g l = ® 54.®g4f the black pawns will advance with check,
White gives a perpetual. supported by the rook from the side.

57.Sb6 &g5 58.a6 h4f 59.&B?

51.d6
59.&h3! would make the draw.
And suddenly, the white pawns are
unstoppable.

51...Self 52.&B6 Sdl 53.d7

1-0

In the next example we shall see a number of

things that are not new to us, as well as how
poor the king is in front of connected passed
pawns, assisted by a rook.

Mads Vestby-Ellingsen - Mahel Boyer

Cala Gonone 2022

abcdefgh
59...Sa3t?
59...h3 wins, as can be seen from move 63
onwards.

60.<&£2 Sa2f 61.&B?

61.&gl ! would hold.

61...5a3t? 62.<&£2 Sa2f 63.&B??

The error here is double. Not only is this a
losing move, if the game continues, but White
also missed the chance to claim a draw from
three move repetition.
abcdefgh
63...h3! 64.Sc6 h2 65.Scl &h4! 66.b5 g5
54.b4?
67.Sbl g4f 68.&e3 g3 69.Sb4f <&h3
A move any hand can make.
0-1
296 Conceptual Rook Endgames

While the positions with connected passed

pawns on both sides are certainly complicated,
they are still governed by the simple principles,
themes and concepts we have discussed so far.
The most basic of them remains the need to
make the most out of your passed pawns.

Manuel Apicella - Vladimir Hamitevici

Chalons-En-Champagne 2021

abcdefgh
60.. .g4??
It is quite easy to follow Black’s thinking.
The white rook cannot get behind the g-pawn,
so it makes sense to push it. But the b-pawn is
ahead in the race and Black needed to respect
that.

Black’s winning method can be found in the

chapter on Tempo Wars (see page 263).
abcdefgh 60...5c7!61.a4Sb7I!
Advancing the pawns now requires the
59.b4g5?
support of either the white rook or king. The
How to distinguish between which pawn
same is not the case for the Black pawns.
to push? Simple. If you cannot find anything
62.Hd5 f4 63. a5 f3 64.a6 Sa7!
wrong with it, you should push the most
Tying down the pawns yet again.
advanced pawn. Soon enough it will become
65.&b3 g4 66.Hf5
a dominating presence.

8
59...f4! won without too many difficulties.
But you had to see the key idea. 60. b5 Hc7! 7
(6O...He4?! also wins, but is entirely absurd. 6
The reason for it winning is beautiful though: 5
61.b6 He2f!! and Black wins, on account of 4
62.&b3 He6!, or 62.& a l He7 63. a4 He4!.) The
3
text move wins more straightforwardly after
61.a4 Hf7! (rook behind the passed pawn... 2
did I not see that somewhere?), or 61.b6 Hb7 1
62.Sd6 f3. abcdef gh
66...Sa8I!
6O.b5
A fantastic move, whose idea can be found
in the chapter on anticipation. White will
 Chapter 2 1 — Connected Passed Pawns 297

put the king on a5 and play b5-b6. This now 62.<&al??

comes without a gain of tempo. A move I find very hard to explain, since it
White makes a miraculous draw in the draws after some difficulty, but the obvious
following line: 66. ..h5 67.&b4 h4 68. &a5 alternative draws any which way you like,
g3 (Black would even lose after 68...h3? and on top of that wins with a bit of accuracy.
69. b6 Sxa6t 70.<&xa6 h2 71.b7 h l = ® White only had 45 seconds to decide, but this
72.b8=® S a l t 73.<±>b6!) 69.b6! Hxa6t does not make it a less bizarre move.
70.&xa6 g2 71.b7 f2 72.Hh5t (72.Hxf2
g l = ® 73.Hf7t g6 74.Hd7!! is also a draw. 62.&b3! Sc6 63.Sd7t &g6 64.b7 Sb6t
And a rather spectacular and confusing one. 65.&c2! Not allowing ...Hbl. White is now
Let’s move on!) 72...&g6 73.b8=® fl=®t threatening Hd6f and Black cannot play
(73...&xh5 is another draw, but this is 65...5b5 due to 66. a4. And after 65...&h5!?
getting too absurd already!) 74. Hb5 g l = ® (otherwise Hg7t is always there) 66. a4 g3
75.®g8t With a perpetual, as the black king 67. a5 g2 68.axb6 g l = ® 69.b8=® there is no
is unable to cross the 5 th rank. perpetual.
67.&b4 <&g6 68.gf4 h5 69.&a5 g3!
A theme we have touched upon frequently. 62.. .Sc6 63.b7 Sb6 64.b8=® Sxb8 65.Sxb8
The goal is not to have most pawns, but to g3 66.§bl f4 67.a4 g2 68.a5 (5 69.a6 £2
queen one of them. 7O.a7 gl=® 71.a8=® Wxblf 72.<&xbl
7O.Hxf3 h4 71.b6 g2 72.b7 Sd8 73.a7 g l = ® fl=®t
74.b8=® Sc5t As said earlier, this endgame is very difficult
Black wins. On the next move a rook check to hold in rapid, but can be held in classical.
follows. Which is what happened here. A draw was
agreed on move 98.
61.b6 Sc2f
After 61...Hc6 62. b7 Hb6 63.Hd7t! g6 The last two examples in this chapter are
64.a4 the a-pawn advances with a gain of both entirely bonkers, and recent top-level
tempo. games played in the 2021 Champions Tour
on Chess24. These are rapid internet games
and the quality therefore weaker. But they are
8
still fascinating fights with a lot of recurring
7 themes for us to enjoy.
6
In the first, White is obviously dead lost. A
5 quick glance at the pawns should assure us
4 of this. But soon he was back in the game,
when Abdusattorov showed some gaps in his
3 technique.
2
1
abcdefgh
298 Conceptual Rook Endgames

Dommaraju Gukesh - Nodirbek Abdusattorov

Internet 2021

abcdefgh
48...&g5?
This is the move that feels most strange to
abcdefgh criticise, but it is actually the move that throws
away the win.
41.&g2 a3 42.Sa6 a2?!
I strongly prefer keeping all the pawns
After 48...c3! 49.h6 Sd2! 50.h7 Sd8 51.Sxa2
independently defended. If Black had played
&e4 Black would win.
42...b4’, the win would not have been difficult.
Take for example: 43.Ha8f &g7 44.Ha7t &f8
49. h6?!
45.Sa8t &£7 46.%a7t &e6 47.h5 c3 48.h6
49.g4! with the outside chance that you get
and Black can reroute the rook to d8 without
to play 50.&g3 and deliver checkmate on the
giving up a pawn, with an easy win.
next move was my preference. After 49...&xg4
50.h6 Hd2 51.h7 Hd8 52.Hxa2 White will be
able to pick up one of the remaining black
pawns and make the draw without too much
trouble.

49...al=® 50.§xal &xh6

abcdefgh
It just so happens that at this point he also
has the spectacular 48...Hxf2tl, when after
both sides queen, Black will have . . .®b2f!, and
the a-pawn will queen on b l .

43.Sa8f &g7 44.Sa7t && 45.Sa6f &f7

46.Sa7t &g6 47.Sa6f &f5!? 48.h5
abcdef gh
 Chapter 2 1 - Connected Passed Pawns 299

51.Sa6f? 57.Ec5t and the pawns will create counterplay)

The black pawns are too strong. The rook White has to play 57.&c2! Eb2f 58.&d3, after
cannot control them alone and needed help which Black cannot make progress. 58...Ed2f
from the king, as well as counterplay on the 59.&c4 White holds.
kingside.
56...&e6 57.g4
51.Ecl!! with the idea &f3-e3 would have It feels wrong to leave the f-pawn hanging.
held. 5 1 ...Eb3 is the best try. Now White must
play: 52.f4! &h5 (52...c3 53. &f3 b4 54.&e4 57T4 &d6 58.Ec8 Bbl! 59.&c2 Eb2f 60.&d3
and the king will arrive in time.) 53. f5 &g4 &d5 61.Ed8f &c5 62.Ec8f & b 5 The black
54.Efl! Exg3t 55.&h2 The f-pawn creates king will enter the position and help the pawns
enough counterplay to draw. (See more in the promote.
next chapter).

51...&g5 52.Sc6 &B!

The king comes to help.

53.&B Sb3t 54.&e2

abcdefgh
57...Sb2?
Thinking with the mouse, not the head.

57...&d6! 58.Ec8 Ebl!! was the key idea once

again. White is done for. If 59.f4 b3! and a
abcdefgh pawn promotes. Or 59.&c2 Eb2f and the
54...c3?? f-pawn is lost.
Pushing the pawn way too early. Black
needed to bring the king in. 58.Sc4??
58.f4 would have held along the lines given
54...&e4! is completely in control, and above.
winning.
58...&d5 59.Sd4f &c5 6O.Sc4| &b5
55.&d3 b4 56.Sc5t? 61.Sf4 Sd2f 62.&e3 b3 63.SfB b2 64.Sb8f
Missing another chance to draw. &c4 65.g5 Bd5 66.g6 c2 67.Sc8f Sc5
0-1
After 56. f4! the most dangerous idea would
be 56...Ebl!, threatening ...b3!. (56...Eb2
300 Conceptual Rook Endgames

The final game is insane. Earlier Whites 38...Sa6! would have removed a lot of
position was overwhelming, which probably counterplay and placed the rook behind the
affected his mood. At this point Sam took the passed pawns.
idea of placing the king in front of the passed
pawns a bit too literally. 39.h6! b4

Sam Shankland - Levon Aronian

Internet 2021

abcdefgh
4O.Sel
Arguably the most natural move. But I
abcdefgh believe Blacks task would have been even more
difficult after:
33.&h4?
It was better to play for activity with: 33.&f4!
4O.Scl!? d3!
Ed2 (33...Sh3 3 4 . S h l followed by g4-g5
There are many ways for Black to go wrong.
and &g4 looks fine for White.) 34.h4 Sxb2
Here we shall only give one. 40.. .Sa5 ? 4 1 .Se 1
35.Se I ! Other moves are possible but this feels
c4f 42.&g6 Sa6f 43. &g5 Amusingly, Black
most logical. White has a lot of activity and
cannot win. White will give perpetual check.
does not stand worse at all.
41.Sxc5 Ec2! 42.Eb5
White would go down easy after 42.Sd5
33.. .5d2 34.h3?
Ec8! 43.&g6 b3 44.Sd7 Sc6f, and Black
First here the game falls into the lost category.
wins.
42...Ec8! 43.Exb4 Ed8 4 4 . E b I
34. &g3! Sxb2 35.h4 would have offered
White enough counterplay, despite the loss
of time compared to the previous note. Black
cannot move all his pawns forward at once.

34.. .5xb2! 35.&g5 d4 36.Sdl c5 37.h4

Sxa2 38.h5 b5?
Natural and sloppy. What is Black going to
do with a third passed pawn? Moving arm-in-
arm, these pawns are slow.

abcdefgh
 Chapter 2 1 - Connected Passed Pawns 301

The most difficult moment.

44...a5I!
It is important to advance the a-pawn first, as
there are some lines where gaining a tempo
with ...a2 is necessary to win.
44...d2? would not win: 45.Sdl a5 46.&F6
a4 47.&e7! Sd3 48.g5 a3 49.g6 a2 5O.h7t!
<±>g7 51.S h i Se3t 52.&d7 Sh3 53.Sxh3!
d l = ® t 54.&e7 and Black cannot get the
queen past the 3rd rank defended by the
rook after 54...®e2f 55.&d8!, or 54...Welt
55.&d7!.
45.&f6 a4 46.g5 abcdefgh
The point behind the 44th move is seen in Unlike the line after 4O.Scl, Black only has
this line: 46.&e7 Sa8 47.g5 a3 48.g6 a2 to make one difficult decision in this line. But
49.S a l d2 and Black wins. this was still one too many.

46.. .b2?
This faulty move attracts the rook to b8,
where it can take the checking piece later.

46... c2! is the right choice. You can easily

explain it with long variations you would
never calculate during a game, or we can try to
understand it conceptually: the white rook will
go behind the first advanced pawn. Also, White
abcdefgh will want to give up the rook for the queen
when it is promoted. So, we want the white
46...d2!
rook on the c-file and queen the b-pawn first,
It is necessary to pacify the white rook and
so the c-pawn is promoted with a discovered
win a tempo doing so.
check, rather than promoting the b-pawn with
46...a3? would be a mistake now. 47.g6 a check, where it is then captured, as in the
and Black has to quickly play 47...Sd6f! to
game.
avoid losing. After 47... a2? White wins with:
48.h7t & h 8 49.g7t! &xh7 5O.Shlf &g8 47.g6t &e7 48.Sb8 c2 49.g7 b l =® 5O.g8=®
51.Sh8#
cl=®f 51.Sxbl Wxblf 52.&e5 d3
47.Sdl a3 48.g6
With little time on the clock, White lost this
48.&e7 Sa8! wins.
lottery.
48...Sf8t’ 49.&e6 a2
Black will continue with 5O...Sa8, winning.
53.®e6f?!
53. &d4! would have won the d-pawn
4O...b3! 41.Se8t &f7! 42.h7 Sh2! 43.h8=® instantly.
Sxh8! 44.Sxh8 c4! 45.&f5 c3! 46.g5
302 Conceptual Rook Endgames

53...&d8 54>d6f &c8 55.Bc6f &b8 Not that he is so in our first position. The
56.&d4 d2 57>d6| &b7 58.®e7t &b6 white rook is heading for the ideal g6-square
soon. But there is still a trick available.

Saha Neelash - Rinat Jumabayev

Novi Sad 2022

abcdefgh
59.®d8*|?
White could have given a perpetual with:
59.®c5t &a6 60.®a3t & b 5 61.®c5t &a4
62,®xa7t &b3 63.®b6t &c2 64.®c6f £ d l abcdefgh
65.®f3t and so on.
81...Sflt?
A natural-looking mistake.
59...&a6 60.®c8f &a5 61.®c5t Wb5!
62.®xa7t &b4 63.®e7t &a4! 64.®a7t 81...Sgl? loses to 82.Ed8!.
&b3 65>f7t &b2
Poor Sam... With this introduction to But Black could have made a draw in a
connected pawns, lets move on to the one sensational way. 8 1 ...Sb4!!, with the key point
area where these endings really stand out. The being that 82.&f5 is met with 82...Sf4f or
miracle draws had against them... 82...Sxg4. Alternatively, after 82.g5t &xh5
0-1 83.Sh7t &g4 84.g6, the white pieces are
misplaced and Black holds. 84...Sb6f 85.&f7
Drawing against Connected Passed Pawns
Sb7t 86. &g8 E b 8 t 87.&g7 &g5! With a
draw.
In this section we shall see two recurring
drawing methods when fighting against 82.&e5 S e l f 83.&d5 S d l f 84.&e6?
connected passed pawns. The first is when the 84.&c6 was the winning move, the key idea
king is close to the pawns and it is not possible being 84...Sgl 85.Sd6f!, winning.
for the stronger side to bring in the king and
make the pawns go forward. We shall see 84...Sgl!
different versions of this. The second relies on Shocker!
a single passed pawn creating counterplay that
will occupy the enemy rook or king. 85.Sd4
Lets start by looking at situations where the Neelash obviously missed: 85.&f5 Sxg4!
king is controlling the passers. 86.&xg4 Stalemate.
 Chapter 2 1 — Connected Passed Pawns 303

85...&g5 Momchil Petkov - Vlad-Cristian Jianu

Hybrid 2021

abcdefgh
The stalemate stuff was fun, but this is the abcdefgh
key image I want you to take with you. Black
42.&C3?
makes an easy draw. The black king is close
Visually the most appealing square for the
to the white pawns, like a boxer hugging his
king to go to. But this is a serious mistake. The
opponent in the ring, avoiding getting hit.
king should have gone to d3 or c l , when after
Because of shouldering, the white king cannot
42...a5, White has 43.Sd5, with a draw.
get close. The position is a prototype draw.
42.. .5a2?
86.&f7 Sal 87.Sd7 Sa4 88.&g7 Sxg4
42...a5! is winning. After 43.Sd5 Black has
89.116 &h5t 90.&E7 Sg5 91.Sf7 Sa5
43...b4f! and the rook will make it to a3, giving
92.Se7 Sb5 93.Sd7 Sa5 94.Sf7 Sb5 95.S16
him a winning position (explained below).
Sb7t 96.&g8 Sb8t 97.&g7 Sb7t 98.Sf7
Sxf7t 99.&xf7 &xh6
43.&b3 Sa4 44.Sh6 &e5 45.Sg6 &d5
Vi-Vi
46.SK6 &c5 47.Sg6 &d5 48.Sh6 &e5
An important point is that after 48...&C5
In the next example, Black would be winning
49.Sg6 Sd4 5O.Sxa6 Sd3t, White has time to
if the pawns were on a5 and b5, and the rook
regroup the rook. But if the pawn was a row
on a3. O r if the rook was on a4 and the king
further down, he would be lost. 51.&c2 Sh3
on the 7th or 8th ranks, where it would be
52.Sa8 &b4 53.Sb8 If the white king was on
possible for it to come to b7 and free the black
the back rank here, Black would have ...Shlf
rook from the responsibility of defending the
and ...Sbl, winning. Now there is nothing.
a6-pawn.
White holds.
304 Conceptual Rook Endgames

49.Sg6 $6 5O.Sh6 &g5 Mohamed-Mehdi Aithmidou - Li Chao

Batumi Olympiad 2018

abcdefgh
51.Sd6! abcdefgh
It is important for White to stay on the 6th
rank. See more about this in Kravtsiv - Ding,
62.Bc7t 63. e4f &e5
Li Chao (in the database referred to as
(page 372).
Li Chao B - the B simple means that there
is another one, if I am not mistaken) was
51...&f4 52.Sh6 &e3 53.Sh3t
enormously resilient in this game.
54.Sh2f &g3 55.Sh6 &g4 56.Sb6 &g5
57.Sc6 &f4 58.Sh6
63...&g6? would fail to 64.Sc6t &f7 65. b6
V2-V2
with the idea d6(t)-b5.
The final example of this method was played
in the first round of the 2018 Olympiad. I 64.Se7t!
White avoids the trap. After 64.%xg7? Sa3t
saw the position we start with live, as I was
65.&g2 &xe4 66. c5 Black makes a fantastic
wondering around the playing hall, spectating
draw. 66...Sb3! 67. b6 &e3 68.Se7t &d4
the few remaining games, which included
69.Sc7 &e3 70.&h3!? f3 71.SI7 f2 72.&g2
Tania Sachdevs game (I was the team captain
&d4! 73.Sf5 fl= 74.£xfl Eb2 and White
of the Indian Womens Team). It is unusual for
is unable to win. After 75. & e l &c4 76. & d l
a player from Morocco to be winning against
<±>d4 77.&cl Eb5 we see that there is nowhere
a player from China in the first round of the
for the white king to go. With the rook stuck
Olympiad, but this was indeed what was
on the 5th rank, progress is impossible.
happening here. I calculated a few lines and
was overwhelmed with the ways White could
64...&B 65.&d6t
finish his opponent off.
 Chapter 2 1 — Connected Passed Pawns 305

68.. .6xd6?
After having managed to create drawing
chances, Li Chao takes back after only 12
seconds (half of it spent writing down the
opponents move and hitting the clock). I
always preach to my students that they should
aim to have two serious options on every move,
seeing how often playing something instantly
turns out to be a mistake. At this point,
Li Chao might have realised that he could delay
taking the knight and force the white king to a
worse square by giving a series of checks first.
abcdefgh
During a game we would just play the idea
66.Sf7t?
After 66. b6?!, Black has a strong defensive and see where it takes us. In the analysis we
resource: 66...&xe7 67. b7 &xd6 68.b8=®t would confirm that it is indeed the difference
between drawing and losing.
&e7 69.®c7t &f6 70.®xf4f &e7 Although
White is winning, it will still take 52 moves
against the most resilient defence to break 68.. .5a2f! 69.&f3 Ea3t 70.&g2
down the black fortress (knight on e6/e8, rook The white king is also misplaced after
on c7), and in practice it may not be possible 70.&e2 Ea2f 71.&d3 Sa3t 72.&d2 &xd6.
at all. 70.. .5a2f 71.&h3 Ea3f! 72.&h4 &xd6
73.SF7 f3 74.&g3 Ec3
O n the other hand, after 66.Sxg7! &xg7 67.b6 The black rook is perfectly placed. After:
White would have won on the spot. 75.Sf6t &c7 76.Ec6t <±>b7
White cannot make serious progress.
66...&e5! 67.Sxg7 Sa3t 68.&E2
69.Sg6t &c5
69...&e5 70. c5 would win for White.

7O.Sc6f &d4 71.b6 &e4 72.c5 Salt

73.&gl!
Preparing to escape via h2-h3-h4 with the
king.

73.. .5b2!
Dominating the white king and thus posing
the opponent more practical problems. The
strength of the move is reflected by the fact that
White responded with an immediate mistake.
abcdefgh
306 Conceptual Rook Endgames

abcdefgh abcdefgh
74.Sc7? 80...&e4?
74.Sc8! was the right idea. White is Li Chao missed a drawing idea for the ages.
threatening to give a check and then push the
c-pawn. Or push the c-pawn and then give a 80...f2!! 81.&xf2 Sb3! would have reached the
check. 74...Sb I f Black is forced to try active same position as in the note to White’s 64th
play. (Against 74...&f3 and 74...&d5, White move. Although the white king can come to
should play 75.Sc7! to help the pawns advance. c2, there is no way to convert the two-pawn
This manoeuvre may be short, but it is difficult advantage. Black can put the rook on b5 and
to see.) 75.&h2 f3 76. <&g3! White has to both shuffle with the king between c4 and d4.
prevent the f-pawn from advancing, and not
be caught up in a checking net. This is how it is 81.Sf8
done. 76...Sglt 77.&f2 Sg2t 78.<±>fl White 81.Sxf3 also won.
is winning. For example: 78...Sb2 79. c6
81...&d5 82.Sc8 Sb3 83.Sc7 &e4 84.b7
74...Sblf 75.&E2 Sb2f 76.&gl Sblf Sb2 85.c6 Sg2f 86.&h3 Sg8
77.&h2 B 78.&g3 &e3 Almost the same position as in the variation
There is nothing wrong with this move, to move 78. If the black king was on d5, Black
except that finding the draw afterwards is would draw.
difficult.
The same is also true for the following 87.SI7
sequence: 78...Sglf 79.&f2 Sg2f 80.&fl 1-0
Sb2 (8O...Sg8? 81.b7 and White wins) 81.b7
S b l f 82.&f2 Sb2f 83. <±>g3 Sg2f 84.&h3 The second way to hold a draw is to get enough
Sg8! Only once the king is on h3, as it is both counterplay with a passed pawn that can push
cut ofF and about to get checked in a nasty part of the opposing army onto the defensive.
way. 85.c6 &d5! 86.Sc8 f2 and White has to The following is a great example.
consider himself lucky that he has a draw with:
87.Sxg8! fl =®t 88.&h2 and all Black has is a
perpetual check.

79.Sf7 &d4! 80.SB

 Chapter 2 1 - Connected Passed Pawns 307

Luis Sousa Reis - Andre Ventura Sousa 44.&e2I! g4

Portugal 2020

abcdefgh
abcdefgh 45.Sd6!I Sxa3
After 45...&g5 46.Sd3 S b l , White can
After being dead lost, Andre Sousa managed
either chase the rook or get a drawn pawn/
to turn the game around and win not only
queen ending with 47.Sdl, or play for activity
the game, but also the 2020 Portuguese
in the rook ending with 47. a4!? Sxb4 48.Sa3,
Championship.
with a draw similar to our main lines.

44.&g2?
46.Sd3 Sa8
White is also lost after 44. £2?, when the
Black would be able to force a queen ending
g-pawn will advance with check.
after 46... Sa2f 47.Sd2 Sxd2f!?, but in this
case, White is well within the drawing margin.
44.&e4? g4 45.Sd6 g3 also sees the g-pawn
48.&xd2 g3 49.&e2! Winning a vital tempo.
sail through.
Black will play ...&g4-h3 and then promote
the pawn, but White will be in time and
44.. .g4 45.a4
should draw this queen ending without serious
Desperation.
difficulties.

45.. .5xb4 46.a5 Sa4 47.Sa8 &h4 48.a6

47.b5 Sb8
Salt 49.&fl g3
Blacks only serious try involves blocking the
0-1
pawn immediately.

The drawing method required a nice

48.Sb3
manoeuvre to improve the white rook.
308 Conceptual Rook Endgames

55.SK?
8
Yes, the rook is a bit better placed here, but
7 not much.
6
The winning idea was: 55.g5! &f2 (55...&d4
5 56.Sd6f with the idea 57.Sdl.) 56.Sf6f
4 &g2 57. g6 e3 and White is winning after:
58.Sa6! (But not 58.g7? Se4f!! 59.&g5 e2
3 and Black holds narrowly) 58...e2 59.Sal!
2 White wins. The pawns are too far advanced.
1
55. g3!? also wins, but feels a bit mysterious.
abcdefgh
48...Sb6
But Black will not be able to advance the
pawns in a meaningful way.

49.&£2 &g5 50.&g3 h5 51.Sb4 h4|

52.&g2 Hd6I? 53.b6 Sd2f 54.&gl h3 55.b7
g3 56.Sbl
Black has to give perpetual check.

In the next example, Renier got a chance to

save the game after a sloppy loss of tempo.

Miguel Santos Ruiz -

abcdefgh
Renier Castellanos Rodriguez In order to draw, Black needs to keep his
options open.
Forni Di Sopra 2018

55...&d4?
55...&d2? also loses after 56. h6 e3 57.g5 e2
58 .Sf2! , and the pawns win.

The saving move was 55...&d3!!, when the

king is ready to go back, as well as forward.
56.g5 (56. h6 e3 57.Sf3 is the same.) 56...e3
Play could continue: 57.Sf3 (After 57.Sd6t
&e4! the king comes back to deal with the
pawns; or if 57.Sfl e2 58.Sel &d2 White
has no more than a draw after giving up the
rook.) 57...&e4! 58.Sfl e2 Black will get the
king to f5 and create counterplay against the
abcdefgh
white king. 59.S e i &f5 60. h6 Se4f 61.&h5
Se3 Black draws.
 Chapter 2 1 - Connected Passed Pawns 309

56.h6 e3 57.SA! 63.&xd3 would be a mistake in view of

Exploiting that the king is too far back. 63. . .f3 64.g 5 t &g6 65. eS Hb3t 66.* ?f4
h5! when White cannot make real progress.
57.. .6e4
A clear point is that 57. .. <&e5 58. g5!! wins 63...6e6 64.Sxd3 &e5 65.Sa3 Sb5 66.&g2
for White. The king shoulders the black king
away.

58.<£h5 e2 59.Sel &f4 6O.h7!

Black is not in time.

60.. .5e5t 61.&g6 Se6f 62.&f7

Black resigned.
1-0

The next game is one of only two in this book

from 2023, played in the yearly January super-
tournament, sponsored by Tata Steel, dating
the completion of this book perfectly.
White has the advantage. His pawns are
abcdefgh
safer and his king is attacking both the black
pawns, meaning one of them will fall. 67.Sa8?
The winning continuation is: 67.Hh3!! &g6
Vincent Keymer - Nodirbek Abdusattorov 68.Hh8 Followed by &f3xf4, or Hh8-h5,
depending on what Black does. Black cannot
Wijk aan Zee 2023 defend both the 5th rank and the f4-pawn.
Against most moves, White puts the rook on
h5. For example, 67...Ha5 68.Hh5 or 67...Hb2
68.Hh5, in both cases followed by Hf5, &f3
and Hxf4.

67...SB3?
Missing a chance to draw with: 67... g5!
68.Hg8t (Against 68.f3, the simplest draw
arrives after 68...Hb2f 69.&h3 Bf2, and White
cannot make progress.) 68... <&h4 69.&f3
Hb3t! This is the key point. With the black
king so excellently placed, Black does not need
to hold on to the f4-pawn. He needs to harass
abcdefgh the white king and pawns. 70. <&xf4 Hb4f
62...d3t? 63.&B! 71. f5 Hb5t 72.£e6 Hb6t 73.&e5 S b 5 t
Not a difficult move. White wants to win a 74.<&d6 Hb6f 75. <&d5 Black is running out of
pawn, but without giving counterplay. Keymer good checks. But he also has 75...Hf6! winning
is a student of Yusupov and Leko, both known back a pawn, while cutting off the white king.
for their outstanding technique.
310 Conceptual Rook Endgames

68.Sa5! &g6 69.Sf5 Sa3 70.B! illusive, at least there was greater complexity in
Another strong technical move. White is the alternative.
strengthening his structure before taking
decisive action. 8O.Ef8!
This was winning. The difference from the
After 7O.Exf4? &g5 71.Sb4 &h4! Black would game is that the white rook is keeping an
have taken control over the g3- and h3-squares, eye on the f4-pawn, while at the same time
preventing White from advancing his king. cutting off the black king’s access to the
scene of the drama.
70.. .Ea4 71.&h3 Eb4 72.&h4 Ea4 73.EA8 80.. .Efl 8 1 ,<±>g4 S g l t 82.<&h5 Eg3 83.&g6!
£g7 74.SB This is the key point. The king is best placed
First thought is that Keymer is gaining time in front of the passed pawn, where it will not
on the clock before proceeding with his plan. only be ideally placed to fight for the squares
But reality was that he was running out of time in front of the pawn, but also be able to assist
and, as yet, without a plan. the rook in harassing the black king.
It is possible Keymer had seen this far and
74.. .<&g6 75.Sg5t &h7 76.Sh5t &g6 77.g5 rejected the direction on account of 83. g6?
But a plan emerges, in the lack of alternatives. Bxf3 84.g7 Eh3t! and Black will make the
draw. If this is the case, he would have been
77.. .5.l 78.Sh6f $5 79.Sfi6t &e5 better off having only a minute to decide,
After a series of quick moves, Keymer has rather than three-and-a-half.
accumulated a few minutes, which he correctly 83.. .Exf3 84.ge8t & d 5 85.&f5 Efl 86.Ee4 f3
invested in this moment. Unfortunately, the 87-Hf4!
problem was too difficult for him to solve. The rook is perfectly placed, while the black
rook is trapped.
87.. .6.6 88.<&f6
8
White wins.
7
6 80.. . B h l t 81.&g4 Bglf 82.<&h5 Sg3
Black has enough counterplay and the game
5 soon ended in a draw.
4
83.Sb3 $6 84.Sb5t &e6 85.g6 SxB 86.g7
3
Sg3 87.&h6 f3 88.Sb8 Sh3t 89.&g6
2 Vi-Vi

1
Black needed to create a passed pawn to create
abcdefgh counterplay.
8O.Sb6?
You could ponder why Keymer chose this 62.. .f3l! 63.&xf3 &e6
option after 3 minutes and 39 seconds of The white rook is driven from the d-file,
thought. The continuation coming in the preparing the improvement of the black rook.
game is not difficult to foresee and could have
been discarded quickly. Even if the win seemed 64.Sd8 <&e7 65.Sa8 d3!
 Chapter 2 1 — Connected Passed Pawns 311

abcdefgh abcdefgh
With ...Hd4 coming, White has no other 78...®b3t! 79.®xb3
way to play for a win than returning with the
rook to block the pawn. This is what gives The final game shows our theme in a slightly
Black drawing chances. more complex scenario. There are more pawns.
It starts out with a position that should draw
66.Sal Sd4 67.&g3 <12 68.Sdl &e6 69.(3 easily for Black, after which things slowly get
&d5 more dicey.
69...&e5 70.&E2 £f4 71.&e2 &g3 72.Sxd2
Hf4 is not the way to play, although it bizarrely Yuri Anikaev - Vladimir Karasev
also draws.
Severodonetsk 1982

70.&F2 &c4 71.&e3 &c3

Black will win the white rook and make the
draw in the following race. The conclusion of
the line is for pleasure as much as anything.

72.fi Sd3t 73.&e2 &c2 74.g5 Se3t

A wholly unnecessary tactic.

75.&xe3 &xdl 76.g6 & c l !

There are other draws, but none like this.

77.g7 dl=® 78.g8=®

abcdefgh
29...a5!?
A natural choice, but not the only one.

3O.Sxe6 a4 31.Sb6 a3 32.Sbl a2 33.Sal

34.&d3 &e6 35. c3 &e5 36.£3 g5
312 Conceptual Rook Endgames

37.&b3 Sb8f I More resistance would be offered by:

Forcing the white king to the a-file. 48...6f4! 49.e7
49.Hel &g4 50.e7 transposes.
38.&xa2 g4 39.Sbl Sa8f 49...He8 5O.Hel <£g4 51.&c6!
Black could also hold after 39...Hg8. For 51.Se4f &f5 is a direct draw.
example: 4O.Hb5t &f4 41.Hf5t &e3 42.e5 51...6xh4 5 2 . & d 7 Ha8
<±>f2! 43.fkg4t <&xg2 44.h4 <&g3 45.Hg5 Hb8
and the draw is getting near. The white king is
too far away.

40.&b3 gxf3 41.gxf3 Sffi

The simplest draw arises after 41... f4!
42.Hfl &e3 and Black wins a pawn.

42.SA Sc8
42... d4! would have included shouldering
and drawn easily.
abcdef gh
43.&b4 h5 44.h4 &f4 45.&b5 &e3 46.e5 53-Sal!
<&e2 47.Sgl 53.e8=®? Hxe8 54.gxe8 <±>g4!I Shouldering.
(54...&g3 would allow the white king to
march in and win.) 55.&e6 h4 56.& e5
h3 57.&e4 &g3 58.&e3 We have a
theoretical draw: 58...&g2! 59.Hg8f & f l !
White also fails to win after: 53.&e6? &g3
54.g g l t &f3 5 5 . H f l t &g4!! (55...&g3?
56.Hf8 loses as above.) 56.Hf8 Ha6f Black
gives checks until the white king goes
somewhere it does not want to, then gives
up the rook with a draw following.
53...5b8 54. e6!!
With the idea of H d l - d 8 .
54... g3
abcdefgh
47.. .6xf3?
The last chance to make the draw was
47...6f2!! 48.gg7 (or 48.gg5 &xf3 49.Hxh5
&g4 5O.Hh6 He8) 48...&xf3 49.e6 &f4 5O.e7
<
&e5! and Black holds.

48.e6 Se8?I

abcdefgh
55.Sglt!
Black is ideally placed for ...He8 and a draw
with king and pawn in a good position, so
White needs to be clever.
55...&f3
If the king goes to the h-file, 56.Hdl wins.
And if it goes to f2, White will collect the
h5-pawn after 56.Sh 1!.
56.Sflt &g4 57.§f8 Eb6t 58.&d5 S b 5 t
59.&c4 Se5 6O.e8=® Sxe8 61.3xe8 h4
62.&d3
White wins.

49.Sg6! &f4 50.&c6 &f5

abcdefgh
51.&d7!
An important tempo-winner.

51...Sa8 52.Sh6 Sa7t 53.&e8 &g4 54.e7

xh4 55.Sg6 &h3 56.&f8 Sa8| 57.e8=®
Sxe8| 58.&xe8 h4 59.&e7 &h2 60.£>fi6 h3
61. >f5 &hl 62.£g4 h2 63.&g3
1-0
Chapter 22

Active or Passive Defence

One of the big questions when we are defending a worse position in a rook endgame is: should
we defend actively or passively? In general, the answer is “active” as you will have gathered already
from reading the previous chapters.

But there are exceptions, and that will be the main factor in this chapter. Giving general advice
on when to go passive (with the default position being active) is hard. The main reason to look
at passive defence in the first place is if the active defence fails. But in general, I feel that with
things that work rarely, developing the intuition through seeing quality examples is the best idea.

In the first example on this topic, we shall see a modern style game, where Black achieves a
Stockfish-sanctioned equality, but still with practical problems in holding the draw (what we used
to call an equal position, before machines were able to calculate everything till the end). This is
the type of game where nine out of ten times, it will be a draw, with White winning the odd one
out. Which is this one.

I wanted to talk to Sam about the chess mistakes in the game, but he rightly brushed me off,
saying that it was a special case. He had felt foggy when he woke up and found it hard to
concentrate during the game. Sam gets a lot of his results from being more present and focused
than his opponent; thus, being less so was highly damaging to his chances. This does not change
the illustrative chess points I want to make. In this game Black at times should seek desperate
activity, and at times play more patiently. Sam did not do well in identifying these moments.

Emil Sutovsky - Sam Shankland

Batumi Olympiad 2018

l.e4 c6 2.d4 d5 3.£}c3 dxe4 4.£}xe4 Jtf5 5.£}g3 JLg6 6.h4 h6 7. f3 e6 8.®ie5 Jth7 9.JLd3
JLxd3 10.®xd3 d7 11.JU4 xe5 12.jLxe5 ®a5t 13.c3 14.0-0 £>g4 15.Sadl xe5
16.dxe5 Sd8 17.®xd8f ®xd8 18.Sxd8f &xd8 1 9 . S d l f &c7!?
We are still within Sam’s preparation. Obviously, this is a bit dry. Heading for an endgame with
the intention to draw it is usually not his style, but at the Olympiad, you hope your teammates
win with White, while you hold the fort with Black.
316 Conceptual Rook Endgames

19... e8!? 20.h5 JLe7 21. e4 was better for 24...&d7!?

White in Guseinov - Pantsulaia, Kigali 2017, I believe this is not only the right attitude,
even if Black later held the draw. but also that Sam was still within the margins
Honestly, I cannot tell if Sam’s move is equal of his preparation; at least in a conceptual
to that game, or slightly better. It is close. sense.

2O.Sd3 Objectively it may not be worse to play

Sutovsky goes for the active option, true to 24.. .h5, when after 25.&h2 &d7 26.Hf3 &e7
style. It was also possible to prepare for a minor 27.Hd3 White is a little better, as Black cannot
piece ending. contest the d-file. But I believe this position
is much harder to play for Black. That he lost
2O.h5 ie7 21.&e4 Hd8 22.gxd8 <&xd8 in the game does not challenge that opinion.
23. &fl would also give White an advantage, Rather, the fact that that he made so many
but is still within a drawing margin. small mistakes strengthens that impression.
Personally, this would be my choice, but to
Stockfish everything is 0.00... 25.h5 gxh5 26.Sxh6 &e7 27.Sxh5
White has to take the pawn immediately. If
2O...te7 2 1 . S B Sffi 22.£ih5 g6 23.&A6 he tried to play prophylactically with 27. &fl
jLxfib 24.Sxft> Hd8 28.&e2, Black would have 28...Hg8 with
active counterplay and an easy draw.

27.. .5d8 28.Sh4

The b-pawn has to be defended.

28... 5 d l f 29.&h2 Sd2 3O.Sb4 b5 31.f4 a5

32.Sb3

abcdefgh
We have arrived at the rook ending, finally.
White is better. The rook is active, Blacks rook
is passive, and there are real weaknesses in the
black position. Black already has to choose
between an active and a passive approach.

abcdefgh
 Chapter 22 — Active or Passive Defence 317

We have reached this position almost the king into the game too. The white rook is
by force from the opening. It is rare that so doing a bit better than it was, but the draw is
many pieces fall directly off the board into the still within reach.
woodchipper, and usually it is the death of the
variation as a try for White. This should have 34.bxa3 Sxa2 35.Sb3
been so too - but if you have a bad day, you 35.c4 leads to exchanges and a draw.
can lose almost any position.
35...Sc2 36.&g3 <$£8
Black has obvious compensation for the 36... f6! was still best.
pawn. The white rook has gone from active
to passive, and the black rook the other way. 37.&B
But White has also won a pawn, giving him a
healthy cushion for later. What is interesting
to me is that Black already is active, and his
positional compensation is so convincing that
he does not have to prove anything. It will not
disappear easily. Thus, we should look at this
as a decent position for Black, where he should
solve his main problem. As we learned earlier,
when the rook is as active as it can be, it is
a good moment to activate the king. Instead,
Sam tried to force matters.

32...a4?!
I do not like this decision. Although all abcdefgh
roads leads to a draw, there is no reason to go Sam was clearly having an off day. The black
through the swamp. king stands worse on f8 than e7, yet now Sam
32.. . f6! would either weaken and undermine chooses this moment to “strike”. If he had
the e5-pawn, or allow the Black king to become returned with the king to e7 and played ...f6,
active. It is easy to prove with variations that Black would still be fine, even though the time
this works. But honestly, this would be a lost on strange activities could have been better
misunderstanding of the decision-making used to activate the king.
process. Realising what problem needs to be
solved, as well as the best tool to do so, relies 37...c5? 38.&e4!
on either logic or a conceptual approach. Brute A pre-emptive move. There will be no check
force calculation is what leads to the need on c3.
to force a draw, when you go from drawing
any which way, to needing to be accurate, as 38...c4
you are not executing the counterplay from a Analysis reveals that 38...&e7 loses to
position of optimal strength. 39.&d3 Bf2 4O.Hxb5 §xf4 41.c4!, but in a
practical sense, this is not a surprise, even if
33.Sb4 a3!? you cannot see how in advance.
33.. . f6! was still the best move. After
34.exf6f &xf6 35.a3 e5 Black would bring 39.Sxb5 Sxc3 4O.a4 Sa3
318 Conceptual Rook Endgames

before acting. This is a common strategy for

defending such positions in the endgame.
Often the stronger side is in control, but has
no obvious way to improve his position. While
the weaker side should absolutely refrain from
doing anything, as it would be a risky gamble.
But there is a lot of psychological pressure
associated with being worse. The need to force
a draw can be strong, but should be resisted.
Whether this is exactly a perfect example of
this, or if that would be 32...a4?!, is open to
debate. To me it is more important that the
abcdefgh general idea is understood.
Move 4 1 , where the gods gave Sutovsky extra
45.Sb7!
time to think, which is ironic, as Sutovsky as
Winning a tempo to undo his mistake on
Director in FIDE has contributed to a further
move 4 1 .
reduction of thinking time, removing all
elements of thinking from the third phase of 45...6g6
the game. 45...Hxa5 46.Sxf7 would be winning for
Later in the book we will discuss the. 2nd White.
best placement of the rook. The best is behind
the passed pawn (a3). But is the rook better to 46.Sb5?
the side or in front of the pawn? Spoiler alert: Repeating moves. This is a bit hard to
if you need to advance the passed pawn, and understand. White had the chance to play
the king cannot help, it is in front of the pawn. 46.Ha7!> which he did not deserve to get again.
Which is the case here.
46...6h6?
41.a5? 46... < )g7! would have corrected the previous
This is therefore a conceptual mistake. error.

White was winning after 41.Ha5! with similar 47.Sb7! &g6 48.Sa7! Ea4 49.a6 Sxf4
lines to what we will see later in the game. 5O.Sa8 Sa4

41...c3 42.&d3 &g7 43.g4 &g6 44.&c2

&h6?
Leaving the pawn undefended, preparing for
f4-f5, which is entirely harmless. Black would
always be able to play ... h6!, if it happened.

44...6g7! would leave White without a way

to proceed, and could be argued to be a
passive defence. Or at least a sort of hedgehog
defence, where Black is waiting for White to
make the first move, lifting part of his armour,

abcdefgh
 Chapter 22 - Active or Passive Defence 319

51.a7! 62.Hg8f &f5 63.a8=® Sxa8 64.Sxa8 &xe5

Often we want to leave a space for the king on 65.Sa5t &e4 66.&d6 f5 67.Sa4f
a7, while at other times it is important to force 1-0
the rook to stay behind the pawn, so it does
not have time to snatch any pawns. Sutovsky Seeking counterplay
spent 1 1 minutes finding the winning method
before executing this move. The next game is a good example of a situation
where White, the weaker side, has to play with
51... g7 52.&xc3 Sal 53.g5 &h7 54.&c4 great activity in order to distract the black
Sa2 55.&c5 Sa6 56.&b5 Sal 57.&b6 S b l f pieces, while at the same time trying to control
58.&c6 Sal 59.&d6 &g7 the passed pawn. It was contested by two of
Ukraine’s greatest current stars, of which
Kovalenko at the time of writing is fighting to
defend his homeland, while Shevchenko has
transferred to Romania, to get the most out
of his talent while he is still young. I admire
both men for their character and resilience in a
difficult situation.

Igor Kovalenko - Kirill Shevchenko

Belgrade 2021

8
abcdefgh 7
Sutovsky had seen the winning method in
6
advance and now carried it out.
5
60. ±d7 4
60.g6I? also works. 6O...Sa6f 61. c7 S a l
(61...fxg6 62.& b 7 S a l 63.Sb8 wins.) 62.gxf7 3
&xf7 63.&d6 Sa6f 64. c5 g7 65. b5 S a l 2
66.&c6 Black has no defence against 67.Sd8,
with a transition to a winning pawn ending.
1
abcdefgh
60...5a2
It is difficult for White to deal with the h5-
6O...Sa6 61.&e7 S a l 62.g6! wins in the
pawn, which is why he needs to find active
same way.
counterplay. This means pushing the queenside
pawns, while at the same time trying to keep
61.g6 &xg6
the h-pawn under control.
Black has to try something...
65.&a2 h4 66.Se2 Sfl
61...fxg6 62.&xe6 and the e-pawn wins the
game easily.
320 Conceptual Rook Endgames

Black has to eliminate the f-pawn, partly in 72...±g6!

order to create a second passed pawn, but also
to create an entry point on g4 for the black
king.

67.a4l?
67. b4! Hxf3 68. b5 makes more sense to
me, as the white rook fits perfectly in the
newly created hole on b2. Still the variation
is non-obvious. 68. ..h3 69. b6 Sg3 7O.Se7t!
(70. b7 Sg8 would rely on finding 71.Sf2! to
make the draw.) Now 7O...Sg7 is forced, and
71.Sxg7t xg7 72. b7 is a trivial draw.

67...M 68.a5! Se3 abcdefgh

73.Sc3?
This attempt to set up a Vancura defence
does not work.

Bringing the king to the battle also fails: after

73.&c2? &g5, the rook needs to give checks,
which now cannot happen from the c-file. And
after 74.Sd3 Sc6f! the white king has to go to
a bad square.

White’s last chance to draw involved throwing

the b-pawn forward to distract the black rook.
73. b4! Eb6 (73...&g5 74.Ec3! gives White a
abcdefgh better version of the game, and Black cannot
69.S£2 win.) 74.&c2 Sxb4 75.&d2 Se4 76.Se3! The
69.Sh2!? was a bit more accurate in view of white king approaches the kingside in time.
69...h3 7O.Sf2! and the pawn will be hanging
on h3, which will be highly convenient. Often 73.. .5a5!
missing such small details is the first step Taking control of the 5 th rank.
towards losing control.
74. b4
69.. .f3 7O.a6 Se6 71.Sxf3 Sxa6f 72.±bl White has no choice but to seek counterplay.
72. b3l? &g6 73.Sc3! would also hold. Passivity after 74.&c2 Sh5! would be
A lot of checks are coming. 73...Sf6 74.&a4 devastating.
&g5 75. b4 &g4 76.Sc4f! and for example:
76...gf4 77.Exf4t &xf4 78.b5 &e5 79.&a5 74.. .5E5 75.Sc6f 76.Sc5t ±g4
&d6 80. b6 The queen ending is a draw - and 77.Sc4f
one of the ones you can hold...
 Chapter 22 - Active or Passive Defence 321

8O.Scl! h3 81.±c2 82.±d3 83.b5

±g4 84.±c4!
After 84.&e3? h2 85.S h i &g3 Black would
win.

84.. .E2 85.Shl! ±g3 86.b6!

The b-pawn is able to give enough counterplay
to secure a draw, if played accurately.

86.. .5B

abcdefgh
77...&B
Showing a lack of direction.

The winning idea was to keep the king active.

77...&f3! 78.Sc3t &e2 79.Sc2f &d3 The
black king is far closer to g3 here than it was on
f7 in the game. It also does a bit of shouldering.
8O.Sh2 h3 81.&cl &e3 Black wins.

78.Sc5t ±g6 79.Sc6t

abcdefgh
87.Sbl?
White could have made the draw with:
87.b7! Sf8 88.Sdl!!
The difficult move. The rook needs to be
able to move all the way up the board, but
also stay as far away from the black king as
possible.
88...5b8 89.Sd3t &g4 9O.Sd4f &g5 91.Sd5t
£g6 92.Sd6t &g7 93.Sd7t
93.Sdl? would be too soon! The black king
needs to be pushed all the way to the back
rank.
abcdefgh 93.. .6g8 94.Sdl!
White gets to the h-pawn first.
79...±f7?
79...&g5! and a return to the winning path
on move 77 was still possible. 87.. .5f7! 88.±c5
322 Conceptual Rook Endgames

Jacob Sylvan - Jakob Aabling-Thomsen

Ballerup 2020

abcdefgh
88...±g2?
Missing the chance to win the game.
abcdefgh
88...Eg7!! was the winning idea (see the
49.. .Bg4?
section on Rooks on the Adjacent File in the
49...6c7!! was the holding move. The idea
Lost Items Chapter). 89.&c6 &g2! 9O.Eb2f
is ... b6, after which Black can go for the
h3 91 .Bb3t g3 92.E b l Bgl Black wins.
g5-pawn without running into an e6f trick.
50.e6 is a decent try, but Black can play
89.Eb2f ±gl 9O.Eblf ±g2 91.Sb2f B£2
50... 6d6! and keep everything under control.
92.Sx£2t &x£2 93.b7 h l = ® 94.b8=® ®h5t
(But not 50...fxe6? 51.b5! when all the black
95.±d4
pawns will fall.)
Kovalenko held the game. The two
50. b5 is the other critical move, when
Ukrainians played till bare kings on move 164.
50... 6b6! is the idea. Black is holding after
either 5Lbxc6 c7!!, or 51.Exc6f xb5
Giving up material to become active
52.Ef6 Eg4.
The next game was played in the Danish
5O.±b3
League. Both players are friends of mine going
50.c4! followed by 5 L b 5 was winning. See
back decades, but in this situation I was hoping
the comments to move 5 1 .
for a draw, as Aabling-Thomsen and I play for
the same team.
50.. .Bgl?
50... c7! worked again.
It is difficult for Black to deal with the threat
of c3-c4 and b4-b5. The black king would be
51.Sa5?
cut off and there are additional targets on the
As stated above, White has a clear path
kingside, looking suspiciously as the second of
forward. One of the key reasons for this
the famous “two weaknesses”.
working is that the g-pawn is indirectly
defended. 51.c4! E b l t 52.<&c3 B e l t 53.&b2
Bgl 54. b5 cxb5 55.cxb5 E h l 56.b6 Eh8
57.&b3 Eb8 58.Eb5 &c6
 Chapter 22 - Active or Passive Defence 323

Pavel Ponkratov - Dmitry Kryakvin

Obninsk 2022

abcdefgh
5 9 . ± c 4 White is winning, as the pawn ending
is lost for Black after ±c4-d5 and the e6-break.
And 59...Sb7 6O.Sc5t! would end the game
instantly.
abcdefgh
51...Sg4 52.Sa7t ±d8 53.Sxf7 Sxg5 54.SK
58...E5? 59.±d5 h4 6O.±xe5 Sf3 61.Sh6
&d7 55.c4 gxe5 56.Sxg6 Se3t 57.±a4 Se4
h3 62.gh4
58.c5 Se7 59.±a5 ±c7 6O.±a6 Sh7 61.gg8
White is in control. The black king being
Sf7 62.Sa8 gf4 63.Sa7t ±b8 64.Sb7t ±c8
passive, while both the white king and rook
65.b5 cxb5 66.±b6 b4 67.Sh7 b3 68.Sh8f
are active, is too much to handle.
V2-V2

Different types of activity

62...Sg3 63.±f4 Sa3 64.gh5 ±g7 65.e5
Sb3 66.fi6f ±f7 67.±B SBt 68.±e4 Sa3
There are a few types of counterplay. A rook
69.Sh7t ±g6 7O.gg7t ±h6 71.Sg8 Sa7
72.gh8f ±g6 73.Sxh3 Sa4f 74.±d5 Sa5t
can harass the stronger side’s king, so he cannot
75.±d6 Sa6f 76.±c7 Se6 77.Sg3t ±f7
find solace. Both rook and king can attack the
opponent’s pawns, hoping to win material
78.Sg7t ±e8 79.Sg5 Sa6 8O.Sg8f ±f7
while passers are advanced, usually with the
81.Sg7t ±18 82.Se7 Sa7t 83.±d8 Sa6
hope of drawing with king and pawn vs rook
84.±d7 Sb6 85.e6 Sb7t 86.±d8 gb8f
and pawn, as so often in this book. Finally,
87.±c7 Sa8 88.Sd7
1-0
there is the pushing of a passed pawn. In this
position, this is where Black’s only chance lies.
The solution is to free the king and the h-pawn
But achieving a way to do this is difficult.
to become active, by preventing the white king
from taking the e5-pawn, and the white rook
from returning to e6.

58...ga3! 59.Sxe5 Sa6I!

The rook goes defensive, so the king goes
active. This is unusual, but is the only way for
Black to create active play.
324 Conceptual Rook Endgames

Black is holding even if pushed a row 62...Sa5t!

backwards. Obviously that makes no sense for Being no longer needed on the 6th rank, the
Black to allow, but it is curious nonetheless: rook harasses the white king.
59...£f6 6O.Ee6t &g5 61.f6 £g6 62.e5 Ea7!
and bizarrely, Black holds. 63.±d4 Sa4f 64.Sc4 Sa6
Despite the thin margin, Black makes the
draw.

Passive defence

Let’s turn our attention to passivity. Which is

not really total passivity in the examples I have
come across, but more a prophylactic patience.
The defender understands that he cannot
prevent the stronger side from advancing. He
also cannot do something active himself at
this very moment. But he can prepare for the
opponent’s advances, as in the following series
abcdefgh of examples.
6O.Sc5
Hans Askov Jensen - Leo Nielsen
White does not have a better way of
regrouping. Black’s next move is forced. Helsingor 2017

6O...±fi6! 6 1 . ± d 5 ±g5

abcdefgh
White lost after:
abcdefgh
62.Scl 58.Hc8? &xa2 59.g8=®t Hxg8 6O.Sxg8
The most ambitious try. ±b2 61.Sb8 b3 62.Sg8 c2 63.Sg2 ± b l
0-1
62.Sc6 Sxc6 63. xc6 h5 leads directly to a
draw.
 Chapter 22 — Active or Passive Defence 325

58.Hc4! would have made it impossible for The next example is not about winning back a
Black to make any progress. The a2-pawn pawn, but simply on preventing the opponent
cannot be saved, but White would be able to from advancing further.
win a pawn back and make the draw instead.
Alexander Baburin - Mark Hebden
The following example is quite similar. My
Birmingham 2014
friend Iain lost after White successfully gave
up his useless passed pawn on the 6th rank and
eliminated the black pawns that were blocking
the true candidate, the pawn not blocked by
the enemy king.

Daniel Howard Fernandez - Iain Gourlay

Crawley 2016

abcdefgh
Baburin lost on time, only to realise that the
position is a simple draw.

67.&g2! Sh3 68.Sa8 Sxh4 69.Sa3! and

Black cannot make progress.

The great Sethuraman won the next game

abcdefgh
using the umbrella technique, discussed on
53.±g7 Sgl 54.Sa5 S h i 55.Sa8f page 131. But before this, Black could have
±d7 56.Sh8 Sgl 57.Sxh7 ±xd6 set up a passive defence, this time combining
57...Sxg5 58.&h6t would be unfortunate, defence of the weaknesses with counterplay
but might also have cut the pain short. against the opponent’s weaknesses.

58.&xg6 e6 59.Sa7 Sg2 60.&h6 Sh2f

61.&g7 Sg2 62. g6
1-0

Iain could have defended with the passive

52...Shi! 53. &g7 Sh5, and White cannot
make progress. If the white king comes to
e6, Black will check him away. The distance
between e6 and g7 is exactly enough for Black
to oscillate between the two defensive positions
on e l and h5.
326 Conceptual Rook Endgames

J. Sai Agni Jeevitesh - S.R Sethuraman There are no checks from behind and the
king can support the e-pawn in peace.
Biel 2021

59.Sxh5 e3 6O.±fl S a i f 61.±g2 e2 62.Se5

e l = ® 63.Hxel Sxel 64.&f3 &xd4 65.&f4
Hflf 66.±g5 ±e5 67.g4 Sgl 68.±h5 ±£6
69.g5t ±g7 7O.g6 Sal
0-1

Black won the following game after mutual

mistakes in time trouble. I always feel a bit
guilty showing players I respect making
mistakes, but when it comes to endgame
positions, it is almost inevitable, as there are
rarely any events where the players are not in
time trouble in the endgame.
abcdefgh
57.Sg5? Miguel Angel Soto - R Iniyan
After the semi-passive defence 57. fl!,
avoiding mate, but otherwise just waiting, Internet 2020
White would have drawn. The idea is: 57... d3
58.Sd6 Sa5 59.&f2! and Black cannot win.

57...d4!
The umbrella.

57...&d3? would allow White to harass the

king from behind, leading to a draw.

58.exd4 ±d3

abcdefgh
In this example, White needs to keep Black
from either playing ...f3t> or answering f2-f3t
with ...&g3.

77.Sh6! g5 78.Sa6?
78.Sg6! was the only move. After 78. ..Sal
79.f3t &f4 80.&f2, White survives with
a passive defence, leading to a theoretical
position: 8O...Sa2f 8 1 . & g l ! Avoiding mating
abcdefgh threats. Now play could continue: 81... xf3
 Chapter 22 - Active or Passive Defence 327

82.Exg5 Eal 1 83.&h2 f4 84.Eg8 &f2 85.Eg2f excellent event. Like in the previous examples,
&e3 86.Sb2 with sufficient checking distance active play does not work. White has to wait
to secure the draw. patiently, for now.

78...Sbl? Alexander Shabalov - Leonid Sokolin

78. ..f4 was winning. After 79.Sa3 Sb 1 ,
St. Louis 2021
Black’s threats are too strong. The same is the
case after 79.f3t g3.

abcdefgh
abcdefgh White lost after:

79.Ha3?
49.Sd4f? ±c6 5O.Se4 Sd5!
This time White could play 79.Sa5!I, when
Cutting off the white king. It is also too late
after 79..T4 8O.f3t, Black again does not have
to prevent ...&b5 and ...a4.
80...&g3. Instead after 80...&h4, White needs
to play 81.Sf5!, to prevent the threat of ...g4,
51.Se8 ±b5 52.±e4 Sc5 53.±d3 a4
using the umbrella trick once again. 81...Sb2f
54.Sb8f ±a5 55.±d4 Sb5 56.Sc8 ±b4
82.&fl &g3 83.Sxg5t &xf3 84.&gl! leads to
57.±d3 Sd5t
the same theoretical position as above. White
Pushing the white king further away. Black
draws by checking from the side.
wins.

79...f4
58.&e4 Sc5
Black wins.
0-1
8O.Sa4 ±h3 81.Sa3t ±g2 82.Sa2 SfiL
The patient defence looks like this:
0-1
49.Sa4! &c5 5O.Sal! &b6 51.&e4
The next example is played in the resurgent
It would be impossible for Black to make
US Senior Championship. Rumour has it
progress. The following variation is not the
that Rex Sinquefield was told what the prize
only drawing line, but it is quite instructive in
fund was for the US Senior Championship
the way White combines patience with active
and was offended on behalf of the greats of
counterplay, the moment Black goes for it.
yesteryear, leading to a larger prize fund and
328 Conceptual Rook Endgames

51...5c5 5 2 . S b l t Tigran Nazaretyan - Bogdan Belyakov

52.&d4! is technically my preference.
Sochi 2020

52.. .±c6 53.Sal ±b5 54.Sblf ±a4

8
7
6
5
4
3
2
1
abcdefgh
abcdefgh 4O...Se2?I 41.±d3 S£2 42.±c4 SB?
The way to win was: 42...Sc2f! 43. &d3 S c l ,
55.±d4! when the rook is changing to g l . As in previous
Eventually White needs to harass the black examples, the passed pawn blocked by the
rook so it cannot find stability on the 5th rank. enemy king is not the candidate; the far away
pawns are. Black could win after something
55...Sc2 56.Salf ±b4 57.Sblf like: 44.&d2 S g l 45.&d3 h5 46.&c4 &f6
White draws, as after 57...&a3 he has 47. xc5 &e5 and the white pieces are entirely
58.Sb5. cut out of the action.

The last two examples of the chapter are more 43.g4 Sg5
complicated, but at their heart they are built on
ideas we have already seen. In the first, Black is 8
winning to start with, but the way to prove the
advantage is far from straightforward. 7
6
5
4
3
2
1
abcdefgh
This is the position where if not straight out
passive, then at least strikingly odd, defence
comes into play.
 Chapter 22 - Active or Passive Defence 329

44.S£2? 25... b8? is also met with 26. a5!, when the
Hoping for counterplay, but Black is too queenside is paralysed. This is not passive
well placed. defence, as there is no defensive element to
what Black is doing, beyond preventing the
The strange 44.Hh4!! would have held the short term loss of a pawn, at the cost of the
draw. After 44... h6 45.&d3 Hd5t 46.&c4 entire position.
Se5 47. &d3 it is difficult for Black to make
progress. The best try is 47...2e6!?, when after 25.. .6f8!? holds narrowly in a longer, boring,
48.Hh2 Hc6 49.Hf2! White is holding. variation.

44...§xg4f 45.&xc5 h5 Much more convincing is: 25...a5!l Preparing

Black is simply faster. to take on e4, now the a-pawn is no longer
vulnerable and 27.Hb4 has been prevented.
46.&d5 h4 47.e4 h3 48.e5 Sh4 49.Sh2 g5 After 26.Sxb7 Black can choose between two
5O.Shl g4 51.&e4 g3t 52.&B g2 53.§el equally good-looking options:
Sg4!? 54.Sgl h2 55.Sxg2 hl=W 26...g5P and the black king will head for e5
0-1 after White plays b2-b4 to get things going.
There is also the provocative 26...f5P, when
The final example sees Black under pressure, Black gets active play. For example: 27.exf5
but with lots of defensive resources. Some Se5 28. b4 axb4 29.Hxb4
active, some less so.

Marc Andria Maurizzi - Gerard Ayats Llobera

Barcelona 2021

abcdef gh
29...Sxc5 Play could continue 30.g4 h5!?
31.h3 hxg4 32.hxg4 Hc2! and Black will be
able to bring out the king, which is more than
can be said of the white position.

26.§xa7 bxc5 27.a5 g6?I

abcdefgh
For a while I believed that Black was holding
24.§xe8f Sxe8 25.§d7 b6? after 27...g5P. But eventually I found a
25...h5? (or any other pawn move on the winning line for White. 28.Ha6! Hxe4 29.Hxc6
kingside) is met with 26.a5!, with the idea of Sb4 3O.a6 Sa4 3 1 .&f2 f5 32.&e3 £f7 33.b3!
26...Hxe4 27.Hxb7, where Black is in serious Sa3 34. &d3 f4
trouble.
330 Conceptual Rook Endgames

abcdefgh
Preparing to create a passed pawn. Otherwise,
White plays < ?c4xc5. 35.h3! An important abcdefgh
move, coaxing the black h-pawn forward. 3O.a7?
35... h5 (Otherwise White continues 36. < ?c4 White missed the chance to aim for our
and 37.Hxc5, winning a pawn.) 36. < ?e4 Sa5 number one idea, to put the rook behind
37.Hc7t £f6 38. a7 White’s next two moves the passed pawn with: 3O.Hb3! He7 Here the
are 39.Hc6f and 4O.Hxc5, eventually leading accurate move is: 31. f2! (31.Ha3? would be
to a winning position. A long and not entirely premature. After 3 1 ...Ba7 32.&f2 &f6 33. &e3
relevant variation. < ?e5 Black arrives in time to keep the white
king out.) And White wins after 31...Ba7
28.a6? 32. Hb6 or 31... f6 32.Hb6!, with the idea
White would have been better off going for 33. Bb7 on the next move.
the c-pawn. The reason for this is not at all
obvious - not even later, where the players did 3O...Ba4 31.&E2 g5
not spot it. The idea on the next move was already
available here.
28.Ha6! Hxe4 29.Hxc6 Hb4 30.a6! Ha4
Otherwise White puts the rook behind the 32.Bc7 h5?
passer, but now the king comes to the queenside Black would have been able to set up a
via f2-e3-d3-c3-b3-c4xc5, and White wins. remarkable fortress (a rarity in rook endings),
unlike anything I have ever seen. 32...c4!!
28...Hxe4 29.Sb7 &g7? 33.&e3 c5!
Correct was 29...Sa4 30. a7 <&g7, with a
transposition to the game. Notice that 3O.Hb6
is not clever. After 30...f5 the king comes to
the queenside and White will lose his pawns
there.

abcdef gh
 There is only one way for White to make
progress. 34.<&d2 g4 35.&c3 &f6 36. b3, but
after 36...cxb3 37. xb3 Ha5! 38. &c4 <&e6
White cannot make progress.

33.&e3 &g6 34.Sxc6t fib 35.Sc7 g4

Another important variation goes like this:
35...f5 36.b3 Sa3 37.&d3 Sxb3t (37...h4
38.&c4 h3I? 39.gxh3 f4 4O.Sd7! f3 41.Sd3
White picks up the f-pawn and wins in the
long run.) 38. 4?c4 Sa3 39.S ?b5! Black cannot
prevent Sc6f followed by Sa6.

abcdefgh
36.b3 Sal 37.&f4?I
37.&e4! was cleaner.

37...Sa2?I
37... f5!? would have given more resistance,
although White still wins after: 38. s£?e5! h4
39.Sd7, where White keeps control.

38.&e4 Sa5 39.&d5 6 40.&c6 Sa3

4rl.&xc5 &g5 42.b4 f4 43.b5 £3 44.gxf3
gxf3 45.b6 Sc3t 46.&d4 Sxc7 47.a8=W
Active defence is almost always preferable;
but in some situations, a degree of patience is
better.
1-0
Chapter 23

Slow Play

As already talked about, there are positions where one side holds the advantage and the other
side can only wait. The technique I was taught by Dvoretsky and Yusupov is reliant on the idea
of first improving your position to the maximum, before you take active decisions that cannot be
reversed. This is specifically relating to this type of position (which is more frequent than you may
think). You will never know with certainty the scope of your advantage. Is the position winning
any which way, or only in the most accurate way? Why find out the hard way; aim to play as
patiently and accurately as you can each time. Not only to avoid being sloppy in the particular
game, but to create a habit of accuracy.

In this chapter we shall see some examples with slow, careful manoeuvres over many moves
achieving small gains.

I came across the first position in my analysis and found that it had great instructional value.
White’s advantage is obvious. He has an extra pawn and the black king is cut off from its defensive
duties of stopping the pawn. But this does not make the win automatic. There are some negatives
that have to be overcome. First of all, the white king is in check and will be harassed endlessly
with checks. There are only two ways to hide. On the 7th rank or on c l . Once White is not in
check he will need a plan for how to advance the pawn. There are additional problems. Sometimes
the pawn endgame is a draw. At the moment, if White goes to the e-line, Black will exchange the
rooks and play ...&e6, taking the opposition. And if White managed to somehow hide the king
and advance the pawn to c4, ...Se8! is often a critical resource. An important point to understand
about the annotations to this example is that White can go down a lot of false paths that do not
crack the defence, but without ruining the chance to go back and find the narrow path to victory.
There are also some alternative paths at certain points, making it unusable for study competitions.
I will in general ignore those here, but focus on the key instructional points.
334 Conceptual Rook Endgames

Aagaard An important moment. White obviously

want to advance the pawn, but good technique
Original 2020
is to collect as many advantages as possible
before committing. In this case, this is also the
only way to win.

4.&d7!
Preparing to push the black king further
away.

4.&b7 Sc4 5. b6 Hc8 does not make any

progress.

4.&d6 <&f6!5.&d7
5.Sf2t &g5 6.<&d5 Hd8t 7.&e5 does not
work as well as in the main line for two
abcdefgh
reasons. First of all, and simplest, Black can
l.&c4 play ...Se8t without White getting to hit
We can quickly see that 1 ,&c3 Sc8f 2.&b2 the rook with <&d7. Secondly, after 7...Sc8
Sb8f 3.&cl ? is way too slow. Black will play 8.Sg2f &h4! the black king has won a
3.. .6f6! 4.c3 &f5 5.&c2 5c8 and we . can tempo compared to our main lines.
see that the black king will quickly become 5.. .5c3 6.Se6t?
a strong actor. The white rook cannot both 6. d6 amusingly still wins. But after
prevent the black king from entering the game 6...5c8 White has to find the unfathomable
and help the white king, which would be done 7.Se6f! &f7 8.Se7t &f6 9.Se2!, where he is
by either defending the pawn or by acting as a exploiting a mutual zugzwang of sorts. He
shield for the checks. can also get there in other ways. But passing
the move is the key point. 9...Sd8f 10. c7
1...Ec8t 2.&d5 Sd8t 3.&c6 Ec8f Sd4 1 l . b6 White is winning.
6...6f5 7.Sc6 Sa3!
7.. .5h3? would be wrong due to 8. d6 &e4
9.Sc4f! and the black king is kicked back.
8. d6 &e4 9.c4 &d4 10.c5 Sh3
Black escapes with a draw, as after 1 1 .Sb6
Hh6f 12. c7 Hh7t 13. c6 &c4 he will
simply take White s extra pawn from him.

4.. .5c5
4...5c4 lasts a few more moves according to
the tablebase. But there is not really a different
idea needed for White. 5.Sf2f g7 6. d6
Hc8 is not progress. 7. d5! Sd8f 8. c6 Sc8f
abcdefgh 9. d6! Again we have this zugzwang position.
Here with the king on d6 instead of e6. Black
 Chapter 23 - Slow Play 335

has to play 9...&g6 (as after 9...Sd8f?! 10. c7! 7. £d6!

Hd4 1 l . b6 the king will manage to help the Again, we have this zugzwang position. Here
c-pawn advance on its own, leaving the black with the king on d6 instead of e6.
king stranded), when after 10. < ?d5, we have
reached move 8 of our main line. 7.&d5 Sd8t S.&c6 Sc8f 9.&b7 Sc3 1 0 . £ b 6
Sc8 and White has not made any progress.
5.§£2t &g7! The main point being: 1 l.Sd2? <&f612.Sd6f
Forcing White to show some additional &e5 13.Sc6 Sxc6f 14.&xc6 &d4 and Black
accuracy. makes the draw.

5... g6 6. d6! (6. e6? &g5! would allow the Black has to play:
black king to enter the game via g3.) 6...Sc8
7. d5! is a transposition to our main line. But 7...6g6!
easier to find. As after 7...Sd8f 8. c7 Sd4 9. b6! wins
rather trivially for White. The c-pawn will start
moving in a moment. The king will manage to
help the c-pawn advance on its own, leaving
the black king stranded.

8.&d5!
Threatening c2-c4.

8. e5? is an attempt to be subtle. But rather

than achieving it, it gifts a tempo that Black
will use to activate the king. After 8... g5!
9.Sg2f ?h4 Black will make the draw by
a tempo. The black king has won a tempo
abcdefgh compared to our main lines.
6.<&e6I!
6. d6 Sc8 is not progress. 7. d5 Sd8f 8...5d8t 9.&e6!
8. e6 Sc8 9.Sg2f &f8 with a draw. This is a key move. White is threatening
c2-c4, so Black will have to return to the c-line.
6...Sc8
The rook has to return immediately to the 9...5c8
8th rank, as Black has to make a move. The 9...Se8t 10.&d7 Se4 l l . £ d 6 ! and the
king would be checked if on g6 and be poorly white pawn will start rolling in a moment.
placed at the edge of the board.
10.Sg2t
The rook cannot find other spaces either, as At this point Black has two different
Black would have to be able to meet & d 5 with defensive ideas. The king can go to h5 with the
idea to go to h3. Or he can go to h6 (h7 is a
...Sd8f and especially not get hit with a tempo
that would allow White to advance the c-pawn worse version), which has its own point, as we
quickly. shall see below.
336 Conceptual Rook Endgames

abcdef gh
With the king on the 5th rank, White will
abcdefgh be able to follow up with a strong check on the
10...6116!? 4th, 5th or 6th rank.
The alternative is:
10... 6 h 5 l l . £ d 5 ! a) 15... g6 16.Sd6t
11.&f5 makes little sense. Black will play There is still one last trick ready for White
11...6h4 and circle his way around to the in this line.
scene of action, kicking the white rook away 16...&f5 17.Sc6 Sb8t
on the way.
l l . S d 2 also does not work. The black king
arrives way too early. 1 l...&g4 12.2d4f &f3
13.c4 < ?e3 with an imminent draw.
1 l...Sd8t 12<&c6
White could also rush to the c l -square, but
at this point it does not work, as the black
king will rush to h3 to kick the rook and
then run across the board to stop the pawn.
It does not even feel close. For this reason,
the white king is heading for the ideal
abcdef gh
b6-square. But in order to make this idea
work, it is important to gain time with a The important detail here is that the white
finesse. king has to follow the black king. It is an
12...5c8t 13.&b7! advanced example of shouldering. The white
13. b6? would lose an important tempo. king has to keep the black king at bay. So, if
1 3.. . h4! 14.Sd2 s£?g3! and White cannot the black king had gone to f7, White would
get in the check on the 4th rank he so badly now win with 18. c7! only. While in this
needs. position, it would be a mistake.
13...5c4 1 4 . £ b 6 Sc8 18. c5!
White has successfully improved his position 18.&C7? Sb2 would allow Black to escape.
as much as humanly possible. It is now time The black king can interfere after either
to press on. 19. d6 s£?e4! or 19.c4 &e5 20.c5 Sb5! and
15.Sd2! White loses the pawn.
 Chapter 23 - Slow Play 337

18...6e5 19.c4 black king cut off on the 5th rank, White
White wins. For example: will now advance the pawn slowly to the 8th
19...5a8 rank.
Waiting with 19...Sbl 2O.Sb6 2c 1 21.< ?b5 17.Sc5 Sb8t 18.&c7 Sb4
is not really different. White plays 2d6 and
c4-c5, or c4-c5 and Sd6, winning.
20.&b5!
And it is over. Notice that it is useful for the
king to defend the c-pawn, as 20...&d4 needs
to be answered with 21.Sd6f.

b) 15... g4?! 16.2d4t £f5 17.c4&e5 18.2h4

wins trivially. The black king may seem close,
but the coordination between the black pieces
is very unfortunate. 18...Sb8f loses to 19. c7
abcdefgh
and 18... d6 19.c5t Sxc5 2O.Sh6t wins on
the spot. White has to win an extra tempo in order
to make it.
c) 15...&g5 19.c3!
This is the more serious alternative. 19. c6? &e4 would be too slow. After 20.c3
16.Sd5t £f4 S b l Black is in time.
16... f6 also offers real resistance. White 19...5a4
has to squeeze everything out of every Now after 19...Sbl 20.c4 &e4 21.Sh5
move. 17.Sc5 Sb8f 18. c7 Winning an White is in time. Black is not.
important tempo. 18...2b4 19.c3 The most 20.&c6 &e4 21.&b5 Sa8 22.&b4
accurate, but also 19. d6 looks natural. 22.c4? would be a blunder due to 22...Sb8f
19...5a4 20.c4 £e6 21.Sc6f &e5 23.&c6 Sb4.
22...5b8t 23.Sb5 Sd8 24.c4 £d4 25.Sb7
White wins.

abcdefgh
It is time for a last accuracy. 22.s£?b6!
Threatening c4-c5 and 2d6 as always.
22...5b4f 23.&c5! Once again shouldering.
23...5a4 24.&b5 S a l 25.Sh6 White has abcdefgh
managed to untangle himself and with the We have reached the key point of the study.
338 Conceptual Rook Endgames

ll.£e5!
The fastest way to win. The logic will become
clear in a moment.

First of all, let’s look at the point of leaving

the king back on h6 (or h7, which is just an
inferior version). If White progresses in the
same way as before, he will be disappointed:
l l . £ d 5 ? ! Sd8t 12. c6 Sc8t 13.&b7 Sc4
14.<&b6 Sc8 15.Sd2 &g7!
This is the trick.

abcdefgh
16.Sg3 £h5 17.&b2
17.c3? &h4 18.Sd3 &g5 19.&c2 &f5 only
leads to a draw.

17.. .£h4 18.Sc3 Sb8f 19.§b3

19. a3? seems a bit aimless. Where does the
king think it is going? It is still nice to see that
19... g5! is the only move that holds, as after
abcdefgh (19...£g4 2O.Sc5! £f4 21.c4 £e4 22.Sd5 the
black king is cut off.)
16.Sd7t?
Technically this is the point where the 19...§c8 2O.Sb6
advantage is spent. White could still play
16.Sf2!, when he would still be able to
backtrack and return to the winning path
after move 5 in our main line.
16...6f6 17.Sc7 Sxc7 18.&xc7
The white king has been side-tracked and
Black will be able to play:
18...6e5
And simply pick up the pawn.

l l . . . S e 8 t 12.&d4 Sd8t 13.&e3 Sc8

14.&d2 Sd8t 15.&C1 Sc8
White has opted for plan no. 2. The king
is protecting the pawn and the rook is freed abcdefgh
to improve its position. This is only possible
This choice of square could look a bit random
because the black king is too far away on h6.
if you are unfamiliar with the themes we
have seen above. But when you recognise the
pattern of having the pawn on c5, protecting
the rook on d6, all becomes clear.
 Chapter 23 — Slow Play 339

20...&g5 21. &b3 ±6 22.c4 &e5 23.&b4

Sh8 24.c5 S h i 25.&b5 Sh2 26.Sd6
8
1-0 7
6
The previous example was deeply inspired by
the following endgame, where Ehlvest showed 5
excellent technique. I am not sure, but this 4
could easily have been after an adjournment.
3
Jaan Ehlvest - Ljubomir Ljubojevic 2
Rotterdam 1989 1
abcdefgh
49.&b3!
49. c4? would fail to 49...Se8! again.

49...Sb8t 50.&c4 Sc8f 51.&b5 Sb8f

52.&c6 Sc8f 53.&d7 Sc4 54.&d6 Sc8
55.Se7t £16 56.Sc7 Sd8t 57.&c6 &e6
58.c4 &e5 59.c5 &d4 60.&b7 Sd5 61.&b6
1-0

The next example has an interesting origin

story. Walton here resigned against the famous
abcdefgh endgame specialist GM Keith Arkell. Keith put
the position on Facebook, saying that he did
46.c3!!
not find the path to the win to be altogether
46.c4? would allow 46...Se8!, with a drawn
obvious.
pawn ending. The pawn ending is not drawn
now, as the king can get in front of the pawn.
Alan Walton -Keith Arkell

46...&fb 47.Se3 Sd8t 48.&c2 England 2019

Going for a trap.

48...5c8 was also possible. The winning

manoeuvre is the same as in the game.

abcdefgh
340 Conceptual Rook Endgames

I analysed the position a bit and found that

the engine was not much use, as it gave similar
scores to the moves that made progress, as it
did to those that were regressive. Eventually, I
realised what the winning method would be:
to queen the a-pawn. And the king needs to
assist. The winning line thus looks like this:

49. h4 g5!
Black needs to make use of the king. He
could go for the g2-pawn, but White would
be able to play Sxe2 in reply, when the pawn
ending is drawn. abcdefgh
6O.Sblt
50. hxg5t &xg5 51.Sg7t && 52.B£7t &e3 After 6O.Shl, Black can defend the h-pawn,
53.BF5 h4 but also win with logical play: 60...a4!
61.Sxh4f &b3 62.Sh8 Sc4!, when the white
king is cut off at the 4th rank.

60...&c5 61.Sal &b5 62.Sblt &a6 63.Sal

abcdefgh
54.Sh5
Also after 54.Sa5 &d3 55.S a l Sc6 56.Sa3t
Sc3 57. S a l &e3 Black wins. The a-pawn will
advance to a2, followed by ...Sb3-bl. abcdefgh
63...Sb2!
54...Sc6 Black wins.
Stockfish 1 5 now finds additional wins, but
63...Sc4? would however be a mistake, as
I like the one I found. after 64.g4!, White has time to recapture on
g3 and still make it back to the queenside by
55.Se5t &d3 56.Sd5t &c3 57.&xe2 <£>b4! teatime.
Helping the a-pawn.
The following difficult endgame, played
58.Sdl Sc2f 59.&B a5 between two top- 50 grandmasters, was a true
rollercoaster ride.
 Chapter 23 — Slow Play 341

Nils Grandelius - Le Quang Liem 64.Af7?

64.<±>f6! 65.&f7 &d6t 66.<±>g7!
Gibraltar 2019
67.&h6 and White wins in the long run.

8 64...&xf7t 65.gxf7 Hf3 66.&e6 &b6 67.He4

7 c2 68.Sc4 Se3t 69.&d7 Sd3t 70.&e7 Se3t
71. £6 S£3t 72.&g7 Sg3t 73.&£8 Sg2
6 74.Sc3
l
5 /2- l /2

4 White had a fantastic winning manoeuvre,

3 where slow play allows him to set up the
position in perfect detail, so he can win on a
2
small detail.
1
61.JLe4!
abcdefgh
Threatening to put the bishop on c2, forcing
61.JU7? Black to take.
Giving Black a chance to save the game.
61... xe4f 62.Hxe4 H£3t
61...&d7?
Missing 61...Sf3t! 62.&e6 63.Eg2
£>d4f 64.£d5 c2 6 5.Sgl 2>f5 66.§cl gf2 and
Black holds.

62.1e6t &d8 63.&e5

63. a5! was stronger, but the text move is still
winning.

63...&c7

abcdefgh
63.&g5!!
This is the difficult move to understand.

63. &g7? &c7! and Black draws similarly to

the game: 64.Sb4 &c6 65.a5 &c5 66.Sb8 Sfl
67. &h6 Sf6 and White cannot win.

63...Sg3t 64. &h6 Sh3t

64...&c7 65.Sb4! is similar.
abcdefgh
342 Conceptual Rook Endgames

65.&g7!
Now that the black rook is on h3 and not f3,
White is happy to step in front of the g-pawn.
All eyes are now on the queenside.

65...&C7
65...c2 66.Sc4 followed by an advance of
the a-pawn.

8
7
6 abcdefgh
5 71.a6!
White wins. The pawn queens and the
4 c-pawn is not offering counterplay. For
3 example:

2 71...c2 72.a7!
1
I have chosen to present the final example
abcdefgh
without analysis. It was played on my birthday
66.Sb4!! by my close friend Sagar, of ChessBase India
Cutting off the black king from dealing with fame. He published it on his website, stating
the a-pawn. Black is in trouble. that he was not at all sure if it was winning. I
put the endgame in the free software FinalGen
66...&c6 67.a5 &c5 68.Sb8 S h i (which does not appear to work as well on
Sflf 70.&e5! modern Windows machines, sadly), which
The king does not want to be on the 6th or found a win by slow play. I sent it to Sagar, who
7th rank when Black takes the g-pawn. later asked for permission for Karsten Muller
to write a large article about it {ChessBase
7O...Sgl Magazine 179). So, as this has already been
70...Self 71.&f5 Sflt 72.£e4 and the dissected in detail by Karsten, and as I have
king goes for the c-pawn. Black will not be in already taken you through other slow games,
time to stop both the white pawns. I just wanted to give the example here, so you
could get a feeling of how difficult it can be to
win such positions.
 Chapter 23 - Slow Play 343

Andreea Cristiana Navrotescu - Sagar Shah

Charleroi 2017

abcdefgh
The second goalpost has been achieved. The
king has crossed the middle of the board.
abcdefgh
64...Sxf4 83.Sa3t &c2 84.Sg3 &d2 85.Sg2 &d3
0-1 86.&B &d4 87.Sa2 &c3 88.Sa3t &b4
89.Sal Sd4
Sagar’s opponent did save herself a lot of
With the king kicking the white rook off
trouble by resigning here, presumably thinking
the 4th row, Black has the chance to take over.
that the two-pawn deficit was unconquerable.
However briefly. Making progress from here is
But as the most likely endgame to arise from
hard.
this is an endgame with f- and h-pawns, White
has real drawing chances. Only careful play
wins.
90.&g3 &c4 91.&h3 Sd6 92.&g2 &c3
93.Sa3t &d2 94.Sa2f &e3 95.Sa3t Sd3
96.Sa6 Sd4 97.Sa3t &e2 98.Sa2f &d3
65.Sa2 &d5 66.Sa5t &c6 67.&h3 Sb4
68. Sal Sb6 99.&g3 e4 100.Se2t &d5 101.Sa2 &e5
The first goalpost has been achieved. 102.Sa5t Sd5

69.&g3 &d5 70.&B &e6 71.Self &d7

72.Sal Sb3t 73.&F2 Sd3 74.Sgl Sd6
75.&B &e7 76.Sal Sd3t 77.&F2 &e6
78.Sa4 &e5 79.Sa5t Sd5 8O.Sa6 Sd6
81.Sa5t &e4 82.Sa4f &d3

abcdefgh
344 Conceptual Rook Endgames

Black has finally, after almost 40 moves,

managed to achieve his goal. Next up comes
...f4f.

103. Sa8 f4f 104.&f2 &f5 105.Sg8 Sd3

106.&g2 Sg3t 107.&h2 &g4 108.Sxg6f
&xh4
So we wound up in the rook vs these two
anyway; but in a situation that is favourable
for Black.

109.Sa6 &g4 110.Sd6 Se3 lll.£>g2 Se2f

112.&F1 abcdefgh
113.Sa6 h4 114.Sa3t Se3 115.Sa8 Sd3
116.&gl E d i t H7.£h2 &e2 118.Se8|
>£2 119.Se4 f3 120.SF4 Sal 121.S18 &e2
122.Se8t &fl
Black wins.
Chapter 24

Lost Items
In this penultimate chapter we shall discuss various topics that did not manage to get a full
chapter of their own, but deserve at least a few moments in the spotlight. Several of them have
already arisen earlier in the book, either in the games or in sidelines. Still, it is valuable to discuss
them separately, if only briefly.

Transition to a pawn ending

Many rook endings transpose into pawn or queen endings. The general theme for these is to
understand queen endings, to understand that the tablebases and books give the defending side
a draw in some theoretical endgames, which is virtually impossible to achieve for a human player
under time pressure. And otherwise, it is mainly a matter of calculation and imagination. As with
this example:

Ding Liren - Etienne Bacrot

Biel (var) 2 0 1 3

abcdefgh
If White had tried 42.h4, Black would have had to find this not-at-all-easy draw.
346 Conceptual Rook Endgames

35...&g7? 36.&B &F6 37.&e4 Sc7 38.h3

42...&d5? 43.Scl &d6 44.h5! wins for White. Sc8 39.&d5 h5 40.&e4 Se8f 41.&£3 Sc8
42.&e4 h4 43.&d5 Sd8f 44.&e4 Sd3
43.&g5 Sc8 44.h5 Hxc7 45.Sxc7 &xc7 45.Sxd3 cxd3 46.&xd3 &e5 47.&e3 ft>
46.h6 g6!! The white king has to retreat, but Black does
not get the opposition.

48.&B &d4 49.&f2! &e4

>/2-*/2

In order to win the pawn ending, Black needs

to make a weakness out of the g4-pawn. This
can be done by:

35...Sc5!! 36.&f3 h5! 37.h3 hxg4f 38.hxg4

&flB! 39.&e4 &e7 40.&d4 &d6 41. £e4
Sc7 42.&d4 Sc8

abcdefgh
The point.

47.&ft> &d7 48.&g7 &e7 49.£xh7 ±f7!

5O.g4 g5
With a fortress/stalemate.

Andrey Orlov - Uwe Boensch

Germany 2022

abcdefgh
Zugzwang. Black wins. For example:

43.&e4 &c5 44.&e5 Sc6 45.F6 Se6t

46.&B &d4 47.Scl Se5#

abcdefgh
Black has an extra pawn, but making
something out of it is far from easy.
 Chapter 24 — Lost Items 347

Mathias Womacka - Giso Jahncke Triangulation

Gmund am Tegernsee 2022

Triangulation is a common theme in all types
of minor piece endings - especially bishop
endings. It is also an important feature of
queen vs rook endings. But triangulation is
not too common in rook endings, and only a
few of my 1 200 examples actually featured this
theme. The following game starts with a little
detail, which is worth including, before we get
to the triangulation moment.

Gawain Jones - Uffe Vinter-Schou

Ulcinj 2019

abcdefgh
White won after:

45.Se3? f4? 46.Sxe5 fce5 47.&e2

1-0

45...f4 was a horrible mistake, easy to identify.

But finding the way to hold for Black is far less
obvious.

Black had to play 45...b5U, when after both

46.Sxe5 fxe5 47. c5 and 46.c5 Sd5 47.Se6
abcdefgh
Sd4 Black will hold.
43.&e4
Instead, White needed to activate the king. White had a nice winning method. 43.Sa4!
Prophylaxis against ...e4 ideas. 43... &f5
45.&el!! 44.&f3 &g5 45.g4! Black either has to take
With the idea &d2-d3 with further or will find it impossible to hold the line after
improvement. Only then comes f3-f4, with a White gets the king to e4. But after 45...hxg3
winning pawn ending. 46.&xg3, the h-pawn is a deadly distraction.
Black cannot hold two fronts. We have not
45...b5I? talked too much of the themes of distant
This is the way to “do something”, but after: pawns, as rooks can travel by the speed of
light, and not too much of the rule of two
46.Hxe5 fxe5 47.cxb5 &b6 48.&d2 &xb5 weaknesses. Mainly because I consider them
49.&c3 general endgame themes and covered them
White is winning due to the distant passed well in A Matter of Endgame Technique.
pawn.
348 Conceptual Rook Endgames

43...&£6 44. &d5 45.&c5 e4 After 47...Sa6 48.&c3! the black king has
lost a valuable tempo and White wins. The
king will get to e3 and decide.
48.a6 Sa7 49.&c5 e3 50.&b6 Sa8 51.&b7
Se8 52.a7 &e4 53.a8=® Sxa8 54.<&xa8 <&d3

abcdefgh
Black has managed to create counterplay abcdefgh
and it is now not easy for White to win.
55.g4!! hxg3 56.Sg2
46.&d4? White wins.
White also cannot win after: 46.&d5? &f4
46...&f4 47.Ha4
47.&d4 (47.Sa4 <±>f5!) 47...Sd6f 48.&c3 e3!
49.a6 e2 and Black holds. Now after 47. &c3 g3, there is no time for
48.&d2, and after 48. <&d4 $£?f4 Black is fine.
The winning method included a nice
triangulation manoeuvre. 47.. .e3 48.&d3t &g3 49.Sa2 e2! 50.&xe2
&xg2 51.Sa3 Se6t 52.&dl Sd6f 53.&c2
Sa6 54.£d2 Sd6t??
46.&b5! Sa8
Losing two tempos.

54... 6h2 was holding.

abcdefgh
47.&c4!!
47. a6 e3 48.' ’b6 <&e4! would give Black
enough counterplay.
47...&e5!?
Black has to try something. abcdefgh
 Chapter 24 — Lost Items 349

55.&e3! 59.. .5xf4?

White is surprisingly aiming for f4. Looking natural, but loses on the spot.

55...£xh3!? 6O.h7 Sh4 61.Sa6!

Desperation. White wins. His next move is 62.Sa8!,
55...5a6 56. &f4 and the h-pawn is lost. forcing Black to take on h7.
1-0
56.a6 Sd8 57.a7 &g2
57.. .Ea8 loses to 58.£>f3 £h2 59.£g4. Returning to the next game from page 180
with a sideline that was not important for
58.Sa2f £gl 59.a8=B Sxa8 6O.Sxa8 h3 the result in the game, but which was highly
61.£f3 instructive. The position could have arisen
1-0 if Black had deviated from the game with
56...Ha8 and White tried 57.Hh5?, which was
Typical tricks originally suggested by Christopher Yoo.

Let’s look at a few typical tricks. We could have Tania Sachdev - Vyanla Punsalan
looked at them up front, as they can appear
Batumi (variation) 2018
at any point. But I wanted a less tactical and
more conceptual narrative this time around.

In the first game Black is on the way to making

the draw.

Roman Grib - Mateusz Bartel

Suwalki 2017

abcdefgh
57...£fl!!
The only move! It looks odd for the king to
run away from the pawn, but threatening to
take it is what matters. The king can get back
on track as soon as the white pieces have gone
to inferior squares.

abcdefgh 57...&dl? loses to our favourite trick after:

59...Hh3 would have been fine. But I would 58.&b3 Hb8t 59.&c3 Sa8
strongly prefer 59...&e7!, to get the king over
to deal with the pawns.
350 Conceptual Rook Endgames

The following trick is pretty too.

Ante Brkic - Sofio Gvetadze

Vysoke Tatry 2 0 1 8

abcdef gh
6O.Sh4!! and the a-pawn will advance.
6O...Sa3t 61.&b2! is an important detail.

58.&b3 Sb8f 59.&a4

Usually in situations like the one after 59. c3,
we give the check. But chess is a details game
- and here the check is inferior to 59...Sa8!> abcdefgh
holding. (59...Sc8f? 60. b4 allows White to
advance the a-pawn without wasting a tempo.) 55.Sd7t
55.SH7? &c5 would allow Black to hold.

55...&c5
If the king goes to the e-file, White wins
with 56.Sh7!. Now this would be met with
...&xb5. But White has another trick!

56.Sdl!! Sxdl 57.c7

Black resigned. I would probably have played
on for a few miserable moves. You can always
blunder more pieces instead of resigning...
1-0

abcdefgh I first saw the following trick in an article by

Jan Timman New in Chess in the 1980s.
59...&e2!
59...Sa8f? makes no sense as long as White
has 6O.Sa5.

6O.Sb5 Sh8 61.&b4 &d3 62.a4 &d4 63.a5

S h i ! 64.Sb6 S b l f 65.&a4 Saif 66.&b5
&d5 67.Sb7 &d6 68.&b6 S b l f 69.&a7
Sal 7O.a6 &c6
White cannot make progress.
 Chapter 24 — Lost Items 351

Aleksander Mista - Shankar Gauri 50.. .5b3t 51.&c4 &e6!

Instead of saving the rook, Black uses the
Charlotte 2020
tempo to stop the pawns.

52.d7
52.c7 &d7 and the g-pawn will advance and
win the game.

52...5b8!?
52...&e7 also won, but it is a nice theme...

53.c7

abcdefgh
White is completely winning. For example,
49.He4 followed by &d4-d5 and then He4-e7
(or with He6f first, if the black king is on f6).

Instead, White tried to be flashy.

49.Sxb7?? Sxb7! 5O.c6

abcdefgh
Now in its purest form. The one I remember
seeing in the New in Chess article...

53...Sb4f! 54.&xb4 &xd7

0-1

abcdefgh
If the rook moves to safety, a pawn will
promote. This was obviously the idea. But
Black has an instructive way to deal with this
issue.
352 Conceptual Rook Endgames

Intermediate move 66...&d6

I was about to give the check on g6 again,
This is usually a middlegame theme, but is fully realising that I was lost, when to my
certainly also present in the endgame. The surprise my opponent offered a draw. Thank
following is my favourite example. you very much!

Playing the game was painful all the way and

ended in a strange way: I was never close to seeing 64.h7! c3 65.h8=®
Sxh8 66.Sxg3!! with an immediate draw.
Jacob Aagaard - Adam Ashton
Perpetual Check
Birmingham 2016

We have seen a few cases of perpetual check

on the 7th rank - and they all felt a bit odd.
The same is the case with the next example,
but I wanted to include it all the same. After
a reckless mistake by White, Hector suddenly
had a chance to force a draw.

Loek van Wely - Jonny Hector

Germany 2020

abcdefgh
64.Sg5t? &d6 65.Sg6t &e5?
I don’t have full control of all endgame
theory to the degree I would wish, but I do
know that for the rook to make a draw on the
1st rank, the pawns need to have four clear
ranks between them. Not three as here.

Black wins after: 65... <&d7! 66.Hg4 c3

67.Hd4t &e6 68.Hdl g2 69.h7 c2 7O.Eelt
&d5 71.h8=® Hxh8 72.&xh8 &d4 73.&g7
abcdefgh
&d3 74.Hgl &e3 Heading for the queen vs 6O.a5?
rook endgame. White had a lot of ways to win. For example,
6O.He8 Sf2t 61.&a3 Exh2 62.Hc8t £d4
66.Sg5t? 63.Hg8 and the rest is easy.
66. h7! draws - and if Black tries 66. ..c3?
67.Hh6! g2 68.h8=® g l =®t 69.&f7t White
would even win.
 Chapter 24 — Lost Items 353

60.. .5F5! 61.Se3t Rook on the adjacent file

61.Sxg4 Kxa5t 62.&bl Sh5! 63.H4 Sb5t
64.&C1 5a5 Black draws. There is no reason The following example is winning for Black,
for the white king to flee along the 1st rank, but as so often in rook endings, the drawing
...&f3 will come... tendency is great and the win relies on accuracy
and a single tempo, making it easy to botch
61.a6 Ha5t 62. b l Hxa6 works in the things up.
same way, the black rook is on the 6th
rank, which makes no difference at all. Filip Haring - Pierre Laurent-Paoli

Corti 2022
61.Ha4 leads to an immediate draw after
61...Hf2t 62.&a3 Hfl!.

abcdefgh
abcdefgh 55...Sd3?
The black king will have to come in via
61...&b4?
e4-d3 to help the pawn. Thus, the rook will
I think Johnny had no idea of the chance he
obstruct the king here.
missed until after the game, if at all...
55...Sa3? also fails, because it does not shield
The draw arises after: 61...&c2! 62.a6 Ha5t!
the king. 56.Hh8 c3 57.Sc8 £d4 58.h4 <±>d3
63.Ha3 Hb5 (The same position could arise
59. h5 c2 6O.h6 Sa8 61.Ec7 &d2 62.Sd7t
after 61...&d2 62.Ha3 c2 63. a6 Hb5!)
&c3 63.Hc7t &b2 64.Hb7t! With a draw.
64.Sa4 Hb2f with a perpetual.
56.Sh8 c3
62.a6 Sa5t
56. ..&d4 This manoeuvre is thematic, but
Black also loses after: 62...Hf2f 63. b l
also slow. White draws by a tempo. 57. h4 &c3
Hxh2 64.He6 & b 5 65.a7 Sh8 66.Sg6 Sa8
While we often are willing to step in front
67. Hg7 White remains in full control.
of the passed pawn, it can at times be slow.
This is one of those times. 58. h5 &d2 59. h6
63.&b2 Sxa6 64.Se4f &c5 65.Sxg4 &d5
c3 6O.Hc8 Hd7 61. f3 c2 62.&f4 White is in
66.h4 &e5 67.&c3 $5 68.Sg5t &e4 69.h5
time.
Sd6 7O.g4 Sd3t 71.&c4 Sh3 72.&c5
1-0
354 Conceptual Rook Endgames

57.Sc8 61.Sc7
We can now see clearly that there is no easy 61.Sc6?! Hb6 and the h-pawn falls.
path for the black king to control the c2-square.
61...&d2 62.Sd7t &c3 63.Sc7t &b2
57.. .6ft 58.Sc4t &g5 59.h4f &h5 60.&fl 64.&g3 cl=® 65.Hxcl &xcl 66.&f4
f4 61.&e2 Sd2f 62.&el Sa2 63.Sxc3
&xh4 64.&fl &g4 65.&g2 Sb2 66.Ha3
Sb7 67.Sa4 Sb3 68.Sc4 &£5 69.Sc5t &e4
7O.Sc4f &e5 71.Sa4
V'l—V'i

For Black to win this position, he needs to

meet three objectives. The first two are simple
to understand: the king needs to be able to
help the c-pawn advance; and the rook needs
to shield the king from checks from behind.
The third objective is to win time, which can
be done with a simple tactic.
abcdefgh
55...5b3! 66...Sf8!
This is the right square. The rook is not in The only winning move, but an easy one.
the way of the king, but it is on an adjacent file
to the passed pawn, ready to shield the king 67.h7 &d2 68.h8=® Sxh8 69.&xf5 Sf8f
from checks from behind. I trust you can take it from here...

56.Sh8 c3 57.Sc8 &d4 58.h4 &d3 59.h5 c2 The next example was winning by the same
60. h6 thinking.

Chitumbo Mwali - Jesper Thybo

Chennai Olympiad 2022

abcdefgh
6O...Sb8!
This is an important tempo-gain and a tactic
we have not seen for the last time in this book.
abcdefgh
 Chapter 24 — Lost Items 355

53.a6! Sanan Sjugirov - Krishnan Sasikiran

53.&e4 Ha4t 54.&d5 Ha3 55.&e4
Moscow 2016
secured the match victory and cannot be called
a mistake.

53...5xf5 54.Sb8!
The key move. The rook is needed on b6,
where it shields the king, defends the pawn
and cuts off the black king along the 6th rank.

54.. .5a3 55.Sb6 Sxg3

Black has to try something.

56.&c6 £6 57.a7 Sa3 58.&b7

Threatening 59.Ha6.
abcdefgh
58.. .5xa7t 59.&xa7 g5
6O.a6? Sx£2 61.Sd7 Sa2 62.a7 Sa5!
The rook is perfectly placed.

63.&c3 &£6 64.&b4 Sal 65.&c5 &e5

66.Sxd5t &e4 67.Sd7 68.&b4
68.Sd3t < ’g2 69. bb Sb I f is also a draw.

68...&g2 69.Sc7 &h3

69. .. <&xg3?? 7O.Sc3t is all Black has to
avoid.

7O.Se7 &g2

abcdefgh
The winning line is astonishing - if you have
60.&b7!
not seen this theme before.
Same same.
6O.Sa6f! &f7
6O...gxh4 61.&c6 &g6 62.&d5
The first point of the check is that if the
White is in good time.
black king goes to the 5th rank, White can
throw the rook to the back rank and advance
The next game ended in a draw in what seemed
the a-pawn. It will take Black a tempo to get
to be an entirely natural way.
the king back close to the 7th rank (where it
has to be if White plays a6-a7), a tempo he can
ill afford to lose. Therefore, it goes to the 7th
rank now.
356 Conceptual Rook Endgames

61.Sb6 Sd2f 65...Sx£2

White also wins after 61...Hxf2 62.a6 Hg2 The point of the 64th move is that White can
63.&xd5 Hxg3 64.&c6 Ha3 65.&b7, as the respond to 65...Ha2 with 66.Hb6f!, winning a
black king is cut off along the 6th rank. And tempo and the d-pawn.
after 65...g3 66. a7 g2 67.Hbl the king is too
far away to help the g-pawn, which thus has 66.&xd3 67.&c2 Sxg3 68.&b2 Sg2f
to be sacrificed. 67. ..gl=® 68.Hxgl Hb3t 69.&a3 Sgl 70.&a2 Sg2f 71.Sb2
We are heading for a well-known theory White wins.
position. 69.&c6 Sc3t 70.&b5 Ha3 71 . S h i !
The position is winning without this trick, but Domination
it certainly helps. 71...&g7 72. &b6 White
Domination is generally a theme from
wins by putting the king in the corner and
transferring the rook to b8. endgames with minor pieces, or endgames
with uneven material. It is rare we see it in
rook endings, as the pieces are at the same time
62.&e5 d4 63.a6 d3
agile, able to travel the board quickly, as well
as blunt instruments, which do not have the
ability to do other types of domination than
cutting pieces off; a theme we have already
covered. But this shall not prevent us from
seeing a few examples of domination in the
rook ending as a finish to this chapter. White
was winning, but not as he played the position:

Levon Babujian - Hovik Hayrapetyan

Yerevan 2 0 2 1

abcdefgh
64.Sb7t!I
There are several reasons why this is the only
winning move.
One of them is that 64.Hd6? can be met
by 64...He2f 65. <&d4 Se6!! with a draw, on
account of: 66.Hxe6? d2! and White probably
has a fortress, but has certainly made things
difficult for himself.

64...&g6 65.&e4
65.Hd7!? also works. 65...Ba2 66.Hxd3 Ha5t abcdefgh
67.Hd5 Sxa6 68.Hd6f! with a winning pawn 43.d8=£)t?
ending. Black would also have held after: 43. h4?
5d2f 44.d?xc4 Sxd7 45.Kxg 6 t &f5 46.Kg5t
£e4 47.Hxh5 Sg7!
 Chapter 24 - Lost Items 357

White also has no advantage after 43.Hxg6f? The winning method included putting the
&xd7 44.h4 Hb8 45.&xc4 &e7 46.gg5 Sh8, black rook in its place.
when White cannot attack the h5-pawn in an
effective way. Black holds with passive play, as 43.&c3!
long as the white king is cut off The rook has to retreat.

43.. .6B
Running for the kingside pawns, seeking the
rook and knight vs rook draw.

44.§f7t
The king cannot get there in time either.
Once the h2-pawn falls, the g3-pawn does not
have long for this world.

44.. .6g4 45.Sf4f &h3 46.Sh4f &g2

47. e6

abcdefgh
43...Sb8 44.&xc4
Only now.

44...Sd8 45.Sxg6t &xd7 46.Sh6

White wins the h-pawn too.

The next game is yet another example from

these horrible hybrid tournaments played in
2021, as a qualification to the World Cup.
Black was fortunate to draw after:

abcdefgh
47...&hl!
Avoiding 47... gl?, where White would
gain a serious tempo with 48. g5!, and win
the game.

48.&xc4 gxh2 49.£ff4 &gl ! 5O. xg6 &g2

The g3-pawn is toast.

51.g4 hxg4 52.gxg4f

The game was drawn on move 104.
358 Conceptual Rook Endgames

Alexander Moiseenko - Frederik Svane 71. &b4 Sxg5 72.Sc6 Sg6 73.Sc7t
74.Sc6t
Hybrid 202 1

Moiseenko missed a beautiful idea, which is

the domination part. After:

55.Sa7t!!
The black king has to go to h6, where it will
be dominated entirely.

55...&h6 56.Sf7 6 57.a7 Sa3 58.&e2I?

Black is in zugzwang and will have to either
let the king advance, or play the desperate
...g5. Thus, the king stays near the centre.
White wins.
abcdefgh
54...Sxg3? 55.Sc8? Sa3 56.Sc7t &h6 Black should have played actively to start with.
57.Sc6 &g7 58.&c2 Simplest was:

abcdefgh
58...g5! 54...g5! 55.61?
The simplest. 55.fxg5 fxg5 56.hxg5 Hxg3 with a draw.

59.fitg5 ficg5 6O.hxg5 h4 61.&b2 Sa5 55...gxh4

62.&b3 h3 63.&b4 h2 64.Sh6 Sxa6 55...g4 also holds.
Why a draw had not been agreed earlier, or
at least now, I cannot say. 56.gxh4 Sa4!
The h4-pawn is eliminated and Black
65.Sxh2 Sc6 66.Sh4 Sc8 67.Sh6 Sc7 makes the draw. Especially as the king gets
68.&b5 Scl 69.&b4 Sc2 70.&b3 Sc5 the g5-square. And with that, we are ready for
something spectacular...
Chapter 25
Unbelievable
Final chapter of the book! Let’s make it a good one. On the following pages, we shall look at
some remarkable endgames, where we will see lots of the ideas we have been discussing in this
book, but also where the aesthetics and the complexities will go beyond concepts, and become
conceptual art. A lot of the examples in this chapter are extracted from long and deep analysis,
but I have tried to cut away as much detail as possible, to leave you with the core of the story each
game tries to tell.

The first game is one of my own. Not because I played well - I did not - but because we can learn
from my mistakes too...

Steffen Pedersen - Jacob Aagaard

London 1997

abcdefgh
42...Sc5?!
We know what the role of the king is in a position where we are better: to help the passed pawn.
Here this meant 42... d4! 43. &F4 c4 44.Sb6 b4, when Black is winning. 45.Sxf6 b3 46.Sb6
Sb5 And the pawn queens.
360 Conceptual Rook Endgames

43.Sd8 b4? 56...&f2 57.h7 b l = ® 58.h8=® Sb6t 59.&g7

A completely unnecessary pawn sacrifice. ® g l t 60.&f8
At this time, I tried to solve all problems with
brute force. And giving up the h-pawn to force
the b-pawn forward is certainly the brute force
option. Long lines to calculate. All entirely
senseless; but you can calculate them...
43...Sd5! would still return to the previous
note, but I was on a different path.

44.Sb8 Sc4 45.Sb5t &d4 46.Sxh5 b3

47.Sb5 &c3 48.h5 b2 49.h6 &c2
abcdef gh
A mesmerising position. With the kings in
the open, we are definitely in the 4th phase.
Black needs to defend the king and give the
queen a path to attack the white king too.
60...®g3!!
There is a sneaky checking opportunity on
a3.
61.®h5
Or 61.Se8 ®xf4 and Black wins.
61...5b8t 62.Se8 Sxe8t 63.£xe8 ®xf4
This queen ending is hopeless for White.
abcdefgh
It should be said that the challenges in the
5O.h7?!
It turned out that White could have offered game were sufficient for me...
more resistance with:
5O...Sc8 51.&g4 Sh8 52. 6 gxh7 53. >xf6
5O.f4Sc3t 5 1 . £ g 4 f 5 t ! !
Check! The only winning move. 8
52.&g5
7
The pawn is immune, or the b-pawn would
queen with check. 6
52...Sg3t 53.&f6 Sb3 54.Sc5t &d2 55.Sd5t 5
&e2 56.Se5t
White is also lost after: 56. h7 Sb6f! 57.&f7 4
(57.&g5 b l = ® 58.h8=® ® g l t leads to a 3
winning attack.) 57...Sb8 58.h8=® Sxh8
59.Sb5 &f3 6O.Sb4 S h i ! Preventing a
2
defence from the front. Black wins the 1
f4-pawn and the game.
abcdefgh
 Chapter 2 5 — Unbelievable 361

53...Sh3? 54.Sxb2f!
53...HH4? looks natural too. Especially as it Confession: I had not seen this move at all...
cuts off the white pawn. But after 54.Hxb2f
£xb2 55.£e5!!> we have shouldering 101 in 54.. .£xb2 55.f4 £c3 56.6
effect. I of course understood that this was a draw.
So, I tried to shuffle around to see if I could
The winning move was the stunning: confuse him. I still cannot believe that it
53...Sh6f!! worked...

56...5e3!? 57.£g7 Sg3t 58.£f7 £<14

59.£e6
59.f6! would have drawn easily.

59...5e3t 60.&d6 61.&e6 Be3t

62.&d6 Ba3 63.16 Ba7I? 64.&e6 Ba6f

abcdefgh
Check! The white king is forced to go to a
worse square and Black wins.
54.£ g 5
Forced. If the king goes to the 7th rank,
Black will play 54...Sh4!, this time simply
winning the f3-pawn.
54...5h3!
The point. The white king is now on the
wrong side of the pawn. abcdefgh
54...Sa6? 55.f4 Sa5 does “win”, but feels
65.&15??
really shaky in practice. A blunder.
55.Sc5t &b3 56.Sb5t &a3!
Black has to keep the option of ...Sb3 alive. White holds after: 65.&e7 &e5 66. f7 Sa7t
57.&g4
67.&e8 &e6 68.f8= t in traditional fashion.
The pawn cannot advance, so needs
defending.
65...&d5 66.f7 Sa8
57.. . 5 h i
Lots of instructive mistakes and missed
Black wins. The white king is misplaced. opportunities.
0-1
362 Conceptual Rook Endgames

Pontus Carlsson - Evgeny Tomashevsky 61.. .e2

This would allow White to force the black
Doha 2014
king to block the e-pawn and give White
time to roll his own pawn up the board.
8 62.Sd7t &e3 63.Se7t &d2 64.Sd7t £ e l
64...& c l 65.Be7 & d l 66. b4! and the pawns
7
ensure the draw. The black king is too far
6 away.
5 65.a6!
As so often, we need to make the most out of
4 our greatest assets.
3 65...5c8 66. a7 Sf8!
The dangerous winning attempt. The black
2
e-pawn will queen.
1
abcdefgh
White is in obvious difficulties, and lost
quickly.

61.Eh4? Bel 62.Bh3 &d2 63.&b4 e2

64.Eh2 &d3 65.Sxe2 &xe2 66. a6 Bal
67.&b5 &d3 68.b4 &c3 69.&c5 Sxa6
7O.b5 Sa5 71.&c6 &b4 72.b6 Bxh5
0-1
abcdef gh
I would have been tempted to say that it is 67.SI7!!
nearly impossible to hold this position for a A small tactical nuance, which would be
human, if I had not seen one do it. possible to miss, is 67. b4 &£2 68.Be7 Ba8!
when Black wins.
61.Bb7! 67.. .5xf7 68.a8=® £f2
Putting the rook on the 7th rank is essential. White cannot prevent Black from making
It is ready both to defend the a-pawn and to a queen, but there is one defensive resource
eliminate the g-pawn. left.
69.®e4!
61...Bc5! The simplest of many drawing moves.
The dangerous try. Pushing the pawn too 69...el=® 70.®h4f!
early is not without venom, but less of a Followed by 71.Wc4f, winning the black
danger. Not because it does not look close; but rook.
because White’s moves are all without serious
alternatives. 62.b4!I
The first difficult move, as it seems brute
force would hold.
 Chapter 2 5 - Unbelievable 363

The majority of training games I have had 75.&b3 &d3 76.&b4 £d4 77.&a5 &c5
from this position ended with Black winning 78. b6 &c6
because of a specific detail. The pawns are under control.

62.Sd7t? &e4 63.b4 62.. .e2

63. &b4 Se5! and Black wins. Another serious try would be:
63...Sd5! 62...5e5 63.Sd7t! &c4
Winning an important tempo. 63. ..&c2 would win the rook, but after
64.Se7t Se5 65.Sxg7 e2 66.Sgl &d3 67. S e i 64.Sc7t & b l 65.Sxg7!, the black king
is at the wrong end of the board and the
connected passers are enough to secure
White a draw. Notice that this line would
not work for White if he had put the rook
on the 8th rank. Black would simply win.
64.Sdl e2 65.S e i Se4 66.a6 Se6

abcdefgh
It appears that White is doing OK. The
connected passed pawns are powerful and
Black will have to get the king to the 2nd
rank to win the white rook. But there is an
idea we have seen before (page 52), which
wins the day. abcdef gh
67.. .5e3!! 68.&a2! But White draws with:
The best try. 67.b5! &xb5 68.a7
After 68. a4 the rook will find a sweet spot White will make the draw. Black cannot win
on a2. the h5-pawn without losing the g7-pawn in
68. <t&b2 is a worse variation of the main line, return.
as the b-pawn would fall with check, making
Black’s task easier. 63.Sd7t
68. . .£d2 69.Sbl &c2 7O.Sel White has to win this tempo, as usual.
70. a6 Sa3t and White may be able to queen
the pawn, but Black will give two checks with 63...&c2 64.Se7 Sc3t
his, the second being ...Walt, skewering the
white king and queen.
70.. .6d2 71.Sbl e l = ® 72.Sxel Sxel 73.a6
Se8! 74.b5 Sa8
A strong GM resigned as White in a training
game here. Black wins after:
364 Conceptual Rook Endgames

David Navara - Mateusz Bartel

8
Katowice (rapid) 2022
7
6
5
4
3
2
1
abcdefgh
65.&a2!!
I would have loved it if the point here was
shouldering, but it is all about preventing abcdefgh
the black rook from getting behind the white
44.£h2?
pawns.
This is a mistake, as the king does not escape
the corner this easily.
65.&a4? &d2 66.Sxe2f (After 66. a6 Se3 there
is also the queen check on a l to consider.)
44.a5! g5
66...<t&xe2 67. a6 S c l and Black wins.
The attempt to set up a Vancura-like defence
with 44...Sc5 45.&f2 Se5 is too slow to
65...&d2 66. a6 Be3
succeed. If Black had time to play both
The only try. White is already comfortably
...g5 and ...h5, it would resemble a practical
holding.
chance, and White would need to take care
to play h3-h4 at the right moment. But as
67.Sxe3 &xe3 68.a7 el=® 69.a8=® ®xb4
things stand, after 46.&f3 Black cannot
It may appear that Black has practical
stop the king from marching forwards, with
chances here, but the fabulous GM Adhiban
a clear win. Interestingly though, if White
held easily in a training game in 2019.
moved the g-pawn either one or two squares
forward, he would lose a vital square for the
The following endgame was played in the 2022
king, and the fortress would hold.
European Rapid Championship. It contains
44...Sa2 45.a6 g5 46. &fl h5 loses to 47.g3!,
many basic ideas, and at the same time is
as seen in the main line. (47.& e l h4 48.& d l
immensely complicated.
g4! followed by ...Sxg2 creates a strong
passed pawn that will secure the draw.)
45. f l !
The greatest shock to the system is that 45.a6?
turns out to be a draw with an advanced
version of a basic theme, the Vancura
Defence. 45...Sc6! 46. &fl Se6! and the
white king cannot get out of the cage. White
 Chapter 2 5 - Unbelievable 365

can try a few tricks and traps, but they don’t 44.. .5c3
succeed past the blitz test... The moment the The most natural-looking move, and enough
white king comes to g4, Black plays ...Sf6, for a draw.
making the cage even smaller.
45...Sa2 But it is worth noticing another idea: 44...2f2!?
Since 45...Sc6? allows 46. e2 before the with the idea 45.&g3 Sfl!, when the white
rook can cut the king off from e6, Black king is caught in the cage and there is no way
must give up the idea of the Vancura defence to play for a win. If he tries to get the rook
here. manoeuvred to the 4th rank, the black rook
46. a6 h5 will find an ideal square on a2, ready to strike
at the kingside, before the a-pawn will be able
to become dangerous.

45.a5 Sa3 46.h4 h5?

A move that feels right in the hand, yet it is a
crucial mistake, opening up the idea of &f4-h6
to White, ruining the Vancura-ideas for good.

46...5b3!
This would have held the draw. There are a
few important reasons why this works.
abcdefgh
47.&gl
It is important to understand that White 47. a6 would once again allow Black to
does not just need to untangle himself rely on the advanced Vancura setup after
on the queenside to win such an ending, 47...Sb6! 48.&g3 Sf6 49.&g4 h6, and
he also needs to create a scenario on the White cannot make meaningful progress.
kingside where Black cannot make the 47. g4 creates weaknesses and Black can hold
draw after losing the rook. The key idea is in many ways. Simplest is 47... h5 48.g5
not to exchange pawns. Currently Black is Sb4, when Black is both ready to take on h4,
threatening ...h4, followed by g4 and later and to set up the Vancura defence, if White
...Sxg2, creating a passed pawn. advances the a-pawn.
47.g3! g4 48. h4! For this reason, 47.&F1 seems the most
White is threatening to win by slowly testing. But Black can hold by bringing
bringing the king to the queenside. Black can the king to the queenside to look after the
only try to follow along with his king, but after passed pawn, just as it says in the textbooks!
48...&f8 49.Sh7! White will win the h5-pawn 47...5b4!
and slowly manoeuvre into position to win the Creating a weakness, relying on 48.a6 Sb6!
g4-pawn as well. Notice that if there were no still being a draw, as the rook can make it to
h-pawns, White would only win the g4-pawn, e6 in time.
which would be insufficient to win the game. 48.g3
366 Conceptual Rook Endgames

to him winning both pawns on the kingside.

53Ac2 &d8 54.Sa8t &c7 55.a7! was
the most natural win. White will take the
g6-pawn and his king will be faster to get to
the kingside, defending the g3-pawn, after
which Black will not be able to defend the
h5-pawn in the long term.

53.. AfZ 54Ac2 &f6 55 Ab2 Sa5 56Ab3

A£5 57Ab4 Sal 58Ab5 &g4 59.Sd8
abcdef gh Sblt!
48...5b2! 49Afl Sa2 5 0 A e l £f8 5 1 A d i Navara clearly relied on 59.. Axg3? 6O.Sd4!
It is tempting to try to get behind the black for the last half-dozen moves.
king after advancing the a-pawn, but there is
a strong defensive idea: 51.a6 e8 52.Sa8t 60Ac6
&f7! 5 3 A d i &f6 and the draw is not too
far away.
51.. Ae8 5 2 A c l &d8 5 3 A b l Sa3 54.a6
&c8
Black makes the draw after another few
moves.

47.g3 &h8 48 Ag2 &g8 49 Af2 &h8 5O.a6

&g8 51 Ae2 52 A d2 &e8

abcdefgh
60.. .Sal?
This is a common mistake. Black returns
with the rook to its ideal place without getting
the most possible out of the checking. I am
not sure why this mistake is so common
(I am sure I have fallen prey to it myself at
some point), but it could be that the “rook
behind the passed pawn” is the only concept
abcdefgh most players (including GMs) know about,
53.Sa8t? while the knowledge of checks in rook endings
Quite a horrific decision, based on an is essentially nil.
understandable miscalculation. The problem
for Navara in this position is that he is planning 60...Sc I f 61 Ab7 Sb I t drew without a fuss.
to win on the queenside, rather than using the White has to put the king on the odd square
a-pawn as a powerful distraction that can lead in front of the pawn, giving Black a chance to
 Chapter 2 5 - Unbelievable 367

take on g3, put his rook on the 5th rank and Bartel could have held the game with:
play ...g5, before White will get anywhere on 65...Sf6f! either kicking the white king out
advancing the a-pawn. of the winning zone, or forcing the white rook
to get involved in the queening of the a-pawn.
61.Bd4f &xg3 62.&b5! A draw is achieved after 66.&c7 2f7f! 67.Sd7
Navara is allowed to execute his plan of Ef8 68.Sd8 Sf7t 69.&b8 Sxa7 70.&xa7
bringing the rook to the a-file after all. &xh4 and the white king is as far away as can
be, giving Black an extra tempo to the draw, as
62...Sblt 63.&a5 Sfl long as he is focusing on advancing the h-pawn,
shouldering with the king on the g-file.

66.Sa4 Sffit 67: &c5 SfB

abcdefgh
64.a7?
This should have failed. Those unfamiliar abcdefgh
with the power of checks in rook endings,
68.Sb4
as well as the importance of a significant
Navara is keen to keep control, although
checking distance, might find this surprising.
queening the pawn at this point would
But by now hopefully you find this somewhat
obviously work too.
elementary.
68...Sf5t 69.&b6 70.&b7 Bf7t 71. &a6
64.Sa4! was the only winning move and one
Sfbf 72.&a5 Ef8 73.Sb8 Sfl
that does not need variations. The black rook
can give a few checks, but will eventually have to
go to a8. White can then win by either bringing
in the king to help promoting the a-pawn, or
by bringing it to the kingside to eliminate the
black pawns. The latter plan may look risky, as
Black has the ...g5-break, but if the white king
has made it to e6, he still wins by a tempo.

64...Sf5t 65.&b6 SflB?

Never miss a check in a rook ending
(although don’t always play them either...).

abcdefgh
368 Conceptual Rook Endgames

74.Sb3t! 39...Sa8?
The black king is cut off on the 3rd rank, A bad move for basic reasons. The rook
as otherwise a check on the 4th would help should not be passive.
queen the a-pawn. After 39...Sc8! 4O.Sxa6 Sxc5 4 1 . £ d 2 Sc3,
the black rook is destined for g3.
74...<±>h2 75.&a6 SflB 76.Sa3 SfiGf 77.&b5
Sf8 78.a8=® 4O.g4
1-0 I like 4O.c4f! bxc3 41. xc3, when White
has improved his position and should hold
Eric Hansen - Ju Wenjun easily.
Internet 2022
4O...3a741.$e3?
4 1 ,c4f! was still holding.

abcdefgh
Another internet rapid game, where two
abcdefgh
strong players show why good moves require
good thinking time. Chess history is not a list 41...6C4?
of results, but a series of highlights we like to The computer comes up with the fancy
think of as a game. Classical chess is considered 41...h61, with the idea 42.&d3 Sc7, when
a mixture of science, sport and art. If you Black wins. Activating the rook is the core idea.
reduce the time control too much, you remove
the deep thinking that makes chess accurate 42.c6 Bc7 43.Sxa6 &b5 44.Ba8 Bxc6
(scientific) and creative (art) and are left with a 45.Sb8t &c4 46.&d2 Sc7 47.Sd8 Sc5
competitive sport without the depth and soul
that makes chess the best game in the world.
By analysing this game, we shall try to inject
some intelligence into the game, as well as
try to make art out of a muddy mess... All of
this is not a criticism of the players, nor of the
format. It is just whining, in hope that classical
chess will continue to exist as long as I am able
to recognise the world around me.
 Chapter 2 5 - Unbelievable 369

57.Sfl the black rook is suddenly out of play

on g4. 57. ..b3!? is a funny try, but after 58.Sf3!,
White holds the balance. And after 57... &d4
58.<&b3 &e3 59.&xb4 f3t 6O.c4 f2 61.<±>b5
White is in time.

56.Sd3 SB 57.Sd8 g3 58.Sg8 &d4 59.Sg4f

8
7
6
abcdefgh
5
48.&cl?
A passive move, while a pawn down. This 4
should feel bad. 3
It was necessary to drag the black rook away 2
from the 5th rank and prevent the ...h5! idea: 1
48.Eh8 h6 49.Sg8! Ec7 5O.Eb8 Black is a
pawn up, but without a clear plan to exploit it. abcdefgh
59...&c5?
48.. .h5! The type of mistake you see a lot of in rapid.
Getting rid of a weakness and creating The b4-pawn is hanging, so Black defends it
weaknesses in the white position. without really thinking.

49. gxh5 Sxh5 5O.Sd3 SB!? After 59 ...&e3! Black would have won on the
Reorganising the rook is natural, although spot.
50... f5 followed by ...g5 was strong too.
60.&a2?
51.Sg3 g5 52.&b2 Sf4 53.Sd3 f5 54.Sg3 White missed the idea to play 60. & c l ! <&b5
g4 55.hxg4 fxg4 61.&d2, placing the king on a much better
55...Sxg4? would be a serious mistake. After: square. Black cannot make progress. For
56.Sf3! f4 example, 61...<&a4 62.Sg8 <&a3 63.Sg7 &b2
64.Sg4! and White holds.

While a lot of this game is about basic ideas

we have discussed a lot in this book, we
suddenly enter a phase in which everything
becomes incredibly deep, with themes such as
mutual zugzwang, corresponding squares and
triangulation.

abcdefgh
370 Conceptual Rook Endgames

abcdefgh abcdefgh
60...&b5? 61...5.3?
Black could have exploited the poor 61...gc3I!
placement of the king on a2. This would be a position of mutual
zugzwang.
60...£d5! 61.5xb4 £e5 62.gg4 &f5 63.§g8 62.£ b l <&c5 63.&b2 <&d5
&f4 Black is winning. For example: 64.c4 The next mutual zugzwang - and the one
gc3!; or 64. b2 Sf2 65. b3 g2 66. c3 s£?f3 that hurts.
67.&c4 gfl 68.<&d5 g l = ® 69.Sxgl gxgl 64.gxb4
70.c4 &f4 and Black wins with the usual White has to take the pawn, as only Black
single-tempo margin. can improve her position.
And after 64.<&bl <&e5 65.Sxb4, the rook
61.&b2? on c3 would not be hanging!
61.<&bl! (or even 61.<&al) would have been 64...gc4! 65.Sb5t &e4 66.&b3 Sd4 67.c4 g2
correct. 61...&c5 62.<&b2 gc3 63. & b l <&d5 68.gg5 &f3
64.<&b2 This is the key position. White does Black wins, as after 69.Sxg2 <&xg2 the white
not want to play <&bl, and Black does not king is misplaced.
want to play 64...<&e5, when White draws
with 65.Sxb4 gf3 66.gg4 &f5 67.Sg8 and the 62.&a2
c-pawn is ready to fly. White could also try to bring out the
king with 62.<&cl Sf3 63. &d2 gf2f, when
After the text move, White should end up in he would have to play accurately to hold
zugzwang. the draw: 64. c l ! (64. d3? would lose to
64...g2, when the black king is threatening
both to go to a4-a3-b2, to attack the c2-pawn,
and to come to the kingside quicker than
White can eliminate the b4-pawn.) 64...g2
65.&b2 <&c5 66.<&b3 &d5
 Chapter 2 5 - Unbelievable 371

65.Sg4 &d5 66.Exb4 &e5 67.Sg4 &f5

68.Sg8 &f4 69.c3?
White was holding after 69.c4!, creating the
needed counterplay with the c-pawn.

69...S£2t?
Gifting a tempo away.

Black was winning after 69...&e3 70. <&b3 &f2

71.&c4 g2 and White is a tempo short.
abcdefgh
67. c3 White is in time to make the draw: 70.&b3 g2
67. ..bxc3 68.<&xc3 &e5 69.&d3 &f5 70.<&e3!

62...Sc3
62...Sa3t 63. & b l ! and White holds.

63.&b2 &c5 64.Sh4?

The window of opportunity for greatness
died for White with this peculiar move.

White’s defence is still 64.<&bl! <&d5 65.&b2!

<&e5 66.Exb4, drawing.

abcdefgh
71.st?b4??
A terrible way to lose the game, but fitting
for the exhaustion that must have kicked in
around here.

The standard shouldering manoeuvre 71.&c4!

&f3 72. <&d5 holds the draw effortlessly.

71...&B 72.<&c5 Efl 73.c4 gl=®t 74.gxgl

Exgl
abcdefgh Black is in good time. White resigned.
64.. .5.? 0-1
And here it ended for Black.
Black wins after the exceptionally elegant The next game is a rapid play-ofF game from
64...Ec4!! 65.Sh5t &d4 66.Eg5 (66.Eh4f is the 2017 World Cup, where Ding Liren
met with 66...<&d5! 67.Eh5t &e6, winning) performed excellently and reached the final.
66.. .Ec3 67.Eg4f &d5! 68.Exb4 Ec4! and we In my opinion this was his greatest result, up
have returned to known territory. until he become World Champion in 2023.
372 Conceptual Rook Endgames

Kravtsiv comes poorly out of the opening and 15-JLb51? ®c7 16.jlb2 might have been the
will have to defend a worse rook endgame. right way to play. Chances are equal.
A lot of the themes we have discussed in this
book are present in the game. 15... e4!

Martyn Kravtsiv - Ding Liren

Tbilisi 2017

l.c4 2.£lc3 e6 3.d4 tb4 4.e3 0-0 5.td3

d5 dxc4 7.Jtxc4 c5 8.0-0 cxd4 9.exd4
b6 10.®e2 tb7

8
7
6
5 abcdefgh
4 16.B?
This is a sad positional admission.
3
2 White has many ways to be uncomfortable. I
am sure that lengthy analysis will reveal a way
1 to defend his position, but in a rapid game,
abcdefgh those are not available. 16.jlel Hac8 17.Sacl
®c6 18. d3 c5! 19.f3 xd3 20.®xd3 ®a4
ll.Sdl?!
is another way for White to face prolonged
Probably trying to make things up on the
suffering. 21.c4 JLa6! is a key point.
spot, but not a likely replacement of 1 l.JLg5,
which is the theoretical move.
16.jld3 xd2 17.®xd2 Hac8 18.c4 would
11...JLxc3 12.bxc3 ®c7 13. e5 give White different problems. The pawns in
1 3.JLd2 looks ugly, but is not bad, although the centre are soft spots.
Black can force a draw with an amusing
combination: 13...jlxf3!? (13... e4 is a decent 16...£lxd2 17.®xd2 Sac8 18.1b3 JU5
alternative.) 14.®xf3 ®xc4 15.®xa8 5k6 Black has taken control of the c4-square and
16.®b7 d5 17.Hacl Hb8 18.®d7Hd8. is perpetually better.

13...£lbd7 14.£lxd7®xd7 15.td2?! 19.®d3 ®c6 20. Had JLc4 21.ILxc4 ®xc4
Already White’s task has become one of 22.Hc2
answering questions. 22.®xc4 Hxc4 23. d5 e5! is just worse, like
every other possibility for White. Finding a
1 5-JLd3? ®c6! would drop a pawn. draw in a position like this is hard, when you
have no activity.
 Chapter 2 5 - Unbelievable 373

22.. .5fd8 23.&E2 h5 24.®xc4 Sxc4 25.&e2 38.E4 &g6 39.Bc6f &f7 4O.Bc7+
b5 26.f4 f6 27.&d3?! 41.Bc6|
This walks straight into a pin, but in poor
positions, there are no easy choices.

27.Sb2! would still give reasonable drawing

chances, although White is still worse.

27.. .e5! 28.fee5 fee5

abcdefgh
Black has played with a smooth efficiency
and won a pawn. But from this point onwards
we see a lack of intuition from the Chinese
number one, regarding the rook endgame,
based on sensible thinking for other types
of positions, but a thinking that shows lack
abcdefgh of understanding in rook endings. This is
29.Sbl? somewhat surprising, as with many Chinese
White had a last chance to make the draw: players, excellent technique has traditionally
29.Eb2! exd4 (29...a6 30.&e4! is one of the key been a big strength. But the new generation
points.) 3O.Sxb5 Sxc3t 31.&e2 Se3t 32. &f2 seems to be more tactically adept and less
Sf8f 33. & g l Se2 34. h4 and his rooks will stay technical.
active and he should make a draw. This book
does not cover double rook endgames; mainly 41.. .6e5 42.Sc5t
due to space concerns. But a key point made Ding Liren calculates quickly and accurately.
by Gelfand is that double rook endings have a There would be no problem seeing the
higher drawing tendency than rook endings, winning continuation if he looked in the right
which are all drawn anyway (according to Dr direction. I should maybe mention that he
Tarrasch). The main reason is that it is difficult spent 38 seconds on the previous move and
to involve the king, as two rooks will be able to just over a minute on this move. It would be
harass it immensely. The second reason is that very interesting to know what his thoughts
a passed pawn can be blocked with one rook, were and why he chose to do what he did. All
while attacked by an active rook from the side, we can do is to look at the logic of the decisions
tying up both of the stronger side’s rooks. and analyse the position to a clear conclusion.

29.. .a6 3O.Sbb2 exd4 31.cxd4 Sdxd4f 42.. .<±>fi6I?

32.d?e3 Be4f 33.&B Sf4f 34.&e3 &h7 The right decision based on general grounds
35.g3 Sfe4f 36.&f3 Bxc2 37.Sxc2 Ba4 should be the first thing that springs to mind:
374 Conceptual Rook Endgames

to bring the king down to assist the passed In two training games (played simultaneously),
pawns. But why should it not be used to fight Indian GMs Aravindh and Karthikeyan Murali
the white passed pawns; you may ask. There found a way to win the game without giving
are a few reasons, mainly that the black pawns up the kingside pawns. 46...<&h6 47.Sc6f g6
are farther advanced and Black is playing to 48.Ec2 a5 49.<&e3 Sg4 50.<&f3 Ec4 5 1 . S d 2
win. But we should be careful not to create b4 Black is in control and Karthikeyan won
overly complicated strategic concepts that have the training game in Chennai Training Camp
limited scope. The basic idea of having the 2019.
king dealing with the passed pawns is simple
and useful and would have been enough here, 47.&e3 &g7 48.Sc7t 49.8c6f &£5
if understood through the lens of multiple 5O.Sc5+ &e6 51.Ec6+
examples. With our few examples, we already Ten moves later and Ding Liren is still facing
have a good chance of understanding that the the same dilemma. Where should the king go?
king belongs in front of the black pawns on
the queenside.

42.. .<&d4!43 .Sxh5 Sa3 1 ! A strong intermediate

move, although Black would also be winning
after taking the a-pawn directly. 44.<&f4 Sxa2

abcdefgh
51... e5?
The same mistake as ten moves earlier.
abcdef g h The black king belongs on the queenside, as
does the white king. With this move Black
4 5 . S h 7 b4 Black is quite clearly a few tempos wins another pawn, but ends up in a drawn
ahead. A visual inspection cannot be trusted
endgame.
entirely, but it weighs heavily here.

5 1 ...&d5! was the right choice (similar to the

43.Sc6f &£7 44.Sc7t &g6 45.§c6f &h7 42... <&d4! line noted earlier). After 52.Sxg6 the
46.Sc2 g6
“best” win is: 52...Sxa2 5 3 . S g 5 t &c4 54.Sxh5
Black is playing his concept. There were
b4! Pushing the passed pawns. 55.Sh8 a5
other ways to try to convert the extra pawn,
(Keeping the pawns together, although 55...b3
but let us stay with Ding Liren’s concept, as it
also looked good.) 56. h 5 a4 57.g4 a3 It is clear
is the one many of the players I have shown the
that Black is arriving first.
position have gone for.
 Chapter 2 5 - Unbelievable 375

52.Sxg6 Sa3t 53.&d2 Sxa2f

53...&d4 54.<&c2!and White is within the
drawing margin.

54.&c3 st?e4 55.Sg5 56.Sxh5 st?xg3

abcdefgh
We have reached a deeply interesting
endgame with rook and two pawns vs rook,
where the objective evaluation is that White
should make a draw, due to his well-placed
king. Below we shall see various reasons why
abcdefgh one position is a draw and another lost. It is all
White has managed to eliminate the about White restricting Black’s possibility for
h 5 -pawn and Black’s passed pawns are untangling himself.
currently under control. The endgame can be
held for White, but it is far from easy to play. 68.Sdl &e3 69.&c5?
An innocent-looking move, but actually a
57.Sh8 Sa4 58.E5 &g4 59.&b3 &g5 mistake.
60.&b2 Sh4 61.Sa8 Sb4f 62.&c3 Sc4f
63.&b3 Sc6 64.SB8 Sh6 69.S a l ! would be a natural move. Black will
With nice play Ding Liren has managed not be able to bring the king to b7, as there
to win a second pawn, but the white rook is would always be Sxa6 at some point.
active, the white king is dealing excellently
with the passed pawns and the black pieces are 69...&e2?
entirely cut ofF. 69...Sh4! would have cut off the white king
and ensured a win. For example, after 70.S a l
65.Sd8 &xh5 66.Sd5t &g4 67.&b4 &f4 <&d3! Black is winning. The b-pawn will run
off soon.

7O.Sd4?
After 70.S a l ! <&d3 71.&b4 Black cannot
improve his position in a meaningful way, as
above.

7O...Shl?
376 Conceptual Rook Endgames

70.. .6e3! 7 1 . S d l Sh4! still wins.

71.Sd6 Sal 72.Sh6?
72.<&b4!Sa4f 73. &b3 with a draw.

72...&d3?
72...5a4! would cut off the white king once
again.

73.&b4 S b l f 74.&a5 S a i f 75.&b4 <&d4

76.Sh4f &e5

abcdefgh
78...&g5?
If Black had played 78...<&g7!, he would have
been able to remove the white king and bring
his own king to the queenside. For example:
79.Sb6 (After 79.Sh2 Sa4f 80.<&b3 there is a
hole on a5 for the black king for later. After the
black king comes to b7, White cannot prevent
...&b6-a5 without allowing ...Sc4-c6, which
will untangle the black forces and get the
pawns going.) 79...Sa4f 80.<&b3 &f7 Again
abcdefgh the king goes to b7 and eventually hides on a5.
Black is threatening 77...Sa4f, so White has
to abandon the 4th rank. 79.Sd6 Sa4f 80.&b3
The black king is on the wrong side of the
77.Sh5t? 6th rank. It cannot reach the b7-square, which
77. &c5? would lose, but the path is long. means that the black rook is tied down to the
77.. .<&f5 78.Sb4 S e l f 79.<&d5 S e i 80.<&c5 a6-pawn.
Se4 81.Sbl Sa4 82.Sb2 £e6 83.Sb3 &d7
Black is slowly making progress; but there is a 80.. .6B 81.Sh6 <&e5 82.Sg6 <±>d5 83.SF6
long way to go. &c5 84.Sh6 Sal 85.&b2
Not giving Black any breathing room.
77.Sh6!! would set up the fortress that arises
in the game. 85...5a4 86.&b3 Sb4f

77.. .£f6 78.Sh6f

 Chapter 2 5 - Unbelievable 377

should not start giving checks until Black

would be unable to get to g4 in response to a
check on h5.

So, after 91.Sc5 &e6 92.Sg5 &d6 93.Sh5

<&c6 White can give checks: 94.Sh6f <&d5
95.Sh5t with a draw.

91...6g5 92.Sa6 Sa4f 93.&b3 94.Sh6

Sg4 95.Sb6 a4f 96.&a2
96.<&a3 Sg3t 97.&a2 Sb3 and wins.

abcdefgh 96.. .5b4 97.Sh6 Sb3

Black is in complete control, but the end is
87.&a3!
worth including.
87. &c3? would allow 87...Sc4f! 88.<&b3 a5
when Black is making real progress.
98.Sh4 &e5 99.Sg4 <±>d5 100.Sh4 <&c5
87.. .a5 88.Sh5t &d6 89.Sh6f &e5 101.Sh5t &b4 102.Sh4t &a5 103.Sh5
SB 104.Sg5 a3 105.Sg4 b4 106.Sh4 S£2f
9O.Sh5t!
Again, 9O.Sa6? would lose to 9O...Sa4t 107.&b3 Sb2f 108.st?c4 a2 109.Sh8
91. b3 &d4 followed by ...Sc4. Black will
eventually be able to put the king on b6 and
block checks with ...Sc6, when he is winning
slowly.

90.. .6ft>

abcdefgh
109...Sc2f 110.&b3 al= #!
0-1

The following game, also given in its entirety,

is not as deep as some of the others. I could not
abcdefgh locate the right chapter to put it in; but on the
91.Sh6f? other hand, I did not want it not to be in the
The final mistake. If the white rook had kept book. Mostly because of the cool zugzwang on
the king cut ofF along the 5 th rank, the draw move 33 - but also because it’s a great game
was still there. The key point is that White overall (from the Swedish Championship).
378 Conceptual Rook Endgames

Isak Storme - Tiger Hillarp Persson

Helsingborg 2021

l.d4 d5 2.c4 e6 3.£if3 c6 4.e3 f5 5.£ic3

6.£ie5 bd7 7.f4 JLb4 8.Jtd3 £ixe5 9.6ce5
&e4

abcdefgh
17.®f4!
White is back to being OK.

17. jbcc5? would lose immediately to 17...b6!

18. txf8 tb7.

abcdefgh 17.. .®xf4 18.exf4 cxd4?

Tiger is one of the most creative players I
10.1,xe4?!
know; and I know him well, as we often sit
10.0—0! with a slight edge was stronger.
next to each other in the Danish League. But
The key strategic point is that if Black starts
this exchange sac is just nuts. I asked Tiger
grabbing material with 10...®xc3?! Il.bxc3
about it and he said he was trying to bring the
JLxc3, then 1 2.®h5t! g6 13.®h6 gives White
game out of balance. Taking these chances was
a strong initiative, the tactical point being that
not sensible and could have lost the game.
13..Jlxal? 14.®g7 Sf8 15Jla3 wins on the
spot.
18...jld7 with even play was prudent.
10...®h4f 11.g3 JLxc3t 12.bxc3 ®xe4
19.txf8 &xf8 2O.cxd4 td7 21.Sfcl tb5!?
Black is surprisingly a bit better. The
Practically a good decision.
weaknesses of the light squares is substantial.

21...b5 22.a4 would split the black pawns.

13.0-0 dxc4 14.®h5t g6 15.®g5 0-0
16.ta3 c5!
22.Sabl ta6 23.&f2 &e7 24.<&e3 &d7
It is more important to open the long
diagonal for the bishop than to count pawns.
 Chapter 2 5 — Unbelievable 379

29.Sa4!! was White’s last chance to win the

8
game. The rook will end up on a3, tying down
7 the black pieces, making it possible for White
6 to open a second front on the kingside.

5 29...axb6 3O.Bxb6f &d7 31.Sxb7t &c6

4
3
2
1
abcdefgh
25.d5!?
Not the only idea, but activating the king
makes good sense.

25.. .exd5 26.&d4 &e6 27.Sb4?

27.Sc3! with the idea of 5=£a3-a5-c5 would
have won the game quickly, I think.
abcdefgh
32.Sb2?
27.. .b6 28.Scbl lb7 32.§f7! was the only move. The key line is:
32...§xa2 33.Sf6t &d7 34.&xd5 c3 3 5 W t ,
when White either keeps giving checks or is
allowed to come in behind the c-pawn.

32.. .5a3!
Threatening mate and forcing White into
passivity.

33.Sd2
White also loses after 33. e6 Sd3t 34. &e5
Se3to followed either by 35.S$?d4 Se4f and
36.. . 6 c 5 , or by 35. f6 c3, when the pawns
are too strong.
abcdefgh
29.Sxb6t?
Returning the exchange and in practice
going from playing for a win to playing for a
draw.
380 Conceptual Rook Endgames

8
7
6
5
4
3
2
1
abcdefgh abcdefgh
33...SB? 38.&d2?
33... h5! would have put White into White was still drawing after 38.&b2
zugzwang. The key variation is not too deep: S b 3 t 39.&c2 Sb8 40.&c3 &c5 41.Se5! Se8
34. e6 S$?d6 35.e7 xe7, with the idea 36.S$?xd5 42.e7 and Black cannot make progress. He
Sd3t and the pawn promotes. has to move the king soon and White gets in
&c3-d4, drawing.
34.Sdl?
White misses a chance to make the draw 38...Sa8?
after 34. e6! S$?d6 35.a4! S$?xe6 (or 35...Sa3 38. ..d4! just won. 39. e7 Sa2f! is the key
36. e7!) 36.&c5! Sd3 37.Sa2 with sufficient point. It may seem elementary by now, but
counterplay. this is a grandmaster missing “checks in rook
endings”. Its a common mistake. 40.& d l
34.. .5E2? S a i f 41.&e2 S x e l f 42.&xel &d7 White can
34...Sa3! 35.Sd2 h5 was winning again. resign.

35.Sei!? 39.&c3 &d6

35.a4! was more principled, to create the
most possible counterplay, using the assets.
35...5a2 36. S e i Sa3 37.Sdl Sxa4 38.e6! £d6
39.e7 &xe7 40.&xd5 White draws.

35...5d2t 36.&c3 Sxa2 37.e6 Sa3t

abcdefgh
 Chapter 2 5 - Unbelievable 381

4O.e7? 34.e4 ®xe4f!

An unfortunate “move 40 mistake” from The rook ending is more promising than the
White, blundering away the e-pawn. queen ending.

40.§e5! would have held. The e-pawn will go 34.. .®xg4? 35.®xc2 ®xe4f 36.®xe4 dxe4
forward. The key idea is 40...§e8 41.&d4! and 37. b4!
the draw is near. 37. &fl ? would give Black the chance to play
37... b4! winning. The idea is that Black has
4O...Be8 41.Sal Sxe7 42.Sa6f &c5 43.Sa5t an extra pawn, which is doubled, but will
&c6 44.Sa8 Se3t 45.£d4 Sd3t 46.£e5 c3 take a tempo to capture, as well as an extra
0-1 tempo with ...a6-a5 at the time of mutual
zugzwang.
The following game was not well played, but 37.. .6f6 38.&fl <±>e6 39.&e2 &d5 40.&e3
has proven to be a fantastic resource for me S$?c4 41.&xe4 S$?xb4 42.S$?xe5 S$?a3 43. f4
to teach the principles of rook endings over S$?xa2 44.g4 b4 45.f5 gxf5 46.gxf5 b3 47. f6 b2
the years. While completely winning for 4 8 . f 7 b l = ® 49.f8=®
Black, many players have failed to overcome This queen ending was reached in three
the decent defensive effort I have been able training games with kids playing each other
to present them with in training games, with in teams in Kolkata 2017. They all missed the
only a handful managing to win. I have deep accurate move order on move 37 - and Black
analysis going in many directions, all of them won the queen ending in all three cases. As
interesting, but we shall focus on the main repeated throughout the book, this ending is
points, to keep the percentage of pure gold difficult to hold in rapid, but should be held
instruction high. in classical.

Viktor Laznicka - Arkadij Naiditsch 35.Sxe4 Sxd2 36.Sxe5 37.Se8 Bxa2

38.Sd8 &e5 39.Sg8 $6?! 4O.Sd8 &e4?
Linares 2017
41.Sg8

abcdefgh
abcdefgh
33...Sc2!?
The most natural move.
382 Conceptual Rook Endgames

We have passed the time control and Black Only after 34.§h4 would Black play 34...Sc2!.
has a completely winning rook ending, despite One of the main reasons is that after 35.e4
a few inaccuracies. With his next two moves, Sxd2 36.exf5 the g6-pawn is no longer pinned
which lose a heavy amount of time that could and can recapture.
be spent focusing on what is good about the
Black position, the win slipped away. 34...®h5!
Pinning the rook on g4 and preparing an
41.. .b4? 42.Sxg6 a5? 43.g4! invasion along the h-file.
Creating counterplay and making the draw.
35.®dl Sh8 36.&fl Bh3t 37.&el Sg2
Black has a winning attack.
43.. .d4 44.g5 a4 45.bxa4 Hxa4 46.Hb6
d3 47.g6 d2 48. d6 Ba6 49.Sxd2 Bxg6t
The first mistake Black made was to play:
5O. fl b3 51.&e2 Sb6 52.Sdl b2 5 3 . S b l
&d4 54.&d2 55.Sxb2 Bx£2f 56.&cl
Sxb2 57.&xb2
Vi-Vi

To start off with. There was nothing objectively

wrong with the starting move, but Black had a
surprising alternative that would have won the
game quickly.

33...Sc8!!

abcdefgh
39...&B?!
We have discussed at length the core
principles of rook endgames, the most central
being the ubiquitous endgame theme of
promoting pawns. This can only happen if
we push them. The main reason Naiditsch
squandered his advantage is that he never
pushed the d-pawn.

39...d4!
abcdefgh
This may not be the most natural move to
The rook gets out of harm’s way, making the those using middlegame intuition; but for
threat of ...®xg4 real. those with a developed sense of the endgame,
it is.
34.®e2 40.§xg6
It is difficult finding ideas for White. The only critical move.
40...d3
We can see, compared to the game, that the
384 Conceptual Rook Endgames

49.Sb6 £d4 44.%d6? has been tried a number of times

Black remains in control and wins. too. Danish GM Jesper Thybo came up with
a stronger win than I did in 44...a5! 45.g5 Sc2
The final point is that although it was not 46. g6 Sc7 and White cannot defend against
winning, Black should definitely have played: two passed pawns from behind.

42...d4! 44.. .d2 45.Sd6 &e5!

An important tempo-gainer that almost
wins the game.

46.Sd8
46.§d7 transposes.

46.. .a5

abcdefgh
The drawing line from here is fabulous and
I have given this position as a playing position
to many grandmasters (always with an hour on
the clock). Only two of them managed to hold
the draw.

43. g4! abcdefgh

White desperately needs counterplay. Black’s pawns are coming quick. It is
Some training games went: 43.He6f &d3 surprising that White can make the draw.
44.g4 §e2! 45.Sc6 §e8 46.f4 a5 47.f5 a4
48.bxa4 Hb8!! and Black wins. 47.g6!
White needs to be hyperactive in order to
43...d3 44.g5! hold.
The most flexible approach is to delay
decisions to follow circumstances. 47. &g3? is too slow. After 47...a4 48.f4f &f5!
49.&f3 Sc2! Black is on his way to winning.
Checking the black king would mainly force
it to go to a better square. 44.§e6f? &d4 47...&F6 48.Sd6f &e7!?
45.Sd6f (45.g5 Se2! wins an important Worth trying.
tempo for Black.) 45...&c3 46.g5 Se2 Black
is winning. For example: 47.Sxa6 d2 48.§d6
&c2 and either the pawn queens, or the
b3-pawn goes. Or both.
 Chapter 2 5 - Unbelievable 385

54.Sd7t!!
Those beautiful checks.

54...&f8
54...&f6?? 55.&f4 and Black is mated.

55.ft
Threatening mate also.

55.. .5a8 56.Sf7t

56.Hh7 eS 57.He7t works in the same way.

abcdefgh 56.. .6g8 57.Sg7f &f8

49.Sd8!!
An important move from the tricks
department.

49...&F6
OK, that did not work out, so we return to
the main line.

49...&xd8 50.g7 d l = ® 51.g8=®t allows

White to make a draw by perpetual. The black
heavy pieces are not able to help the black king
find a safe haven.

5O.Sd6t &g7 51.£g3 a4 52.f4!

abcdefgh
White has no time to waste. 58.Sf7f &e8 59.Se7t
With a draw, achieved by Aravindh in 2019
and Sam Shankland in 2020. Black has to
return to the kingside not to see the g-pawn
promote with mate.

abcdefgh
386 Conceptual Rook Endgames

The following endgame is one of my favourite 50.&d4 <$g7?! 51.§b7f &g6 52.Sb8 &g7?!
pieces of chess art ever. Analysing this endgame Black is happy to repeat, but should perhaps
for close to 1 00 hours, alone and with students, have pondered if the white rook really is so
I have found so many interesting points and glorious behind the h-pawn.
variations that presenting them here over
just a few pages is intensely unsatisfactory. A better idea was to advance his king with
The game was played in a rapid play-off at 52...&g5!, making the draw with ease. As
the earlier stages of the World Cup and was we shall see below, Black will always have to
most likely ignored by all but me. The players advance the king anyway, so he cannot feasibly
made a lot of the important decisions in this prevent White from playing Sh8, should he
game within 1-2 seconds and perhaps never want to.
looked back at them either. When we enter the
game, it is clear that White has the advantage,
as his king is better placed and the connected
passed pawns are stronger than the split
pawns. But the position should still be within
a drawing margin for Black, as the h-pawn is
well advanced. As elsewhere in this book, the
annotations do not correspond perfectly with
computer evaluations, but presumes that chess
is played between two humans without access
to assistance and thus the difficulty of the
decisions each player will have to make weighs
in heavily on the evaluation of the quality of
the moves. abcdefgh
53.Sc8?!
Hrant Melkomyan - Boris Grachev Logic will tell you that the white rook
Tbilisi 2017 does not belong in the c-file, where it will
lose flexibility between going to c l and c2, a
decision that has to be made in advance, so
he can continue with &c5-b6. The rook also
does not belong in the e-file, as the advancing
f-pawn will win a tempo, vital in any race. So,
the rook belongs in the d-file, meaning that
White should play the move that will always
be the same first.

We will analyse 53. c5! below. First we will

see some of the amazing variations that could
have arisen in the game.

abcdef gh 53...h3!
The correct move. Advance the best passed
pawn, as long as you stay in control of it.
 Chapter 2 5 - Unbelievable 387

53...F51? Black draws by chasing the rook along the

This is less natural, but still within the 1st rank.
drawing range. 63.Sxh3I? can be tried, but as we shall see in
54.§c2 &f6 55.&c5 f4 56.&b6 &g5 57.b5 f3 several places in the analysis to this game, this
58.a6 does not win. 63...fl=® 64.§xh4 ®f2 White
cannot make progress.

White is now prematurely forced to choose

between the 1st and the 2nd rank. Melkumyan
felt it was natural to stop the black pawns as
early as possible.

54.Sc2
54.Sell?
A tricky move, as 54...h2? would lose (see
53.&c5! analysed below on page 389; the
abcdefgh
points are the same). But Black has other
Black’s only drawing idea is something from possibilities.
out of space. 54...f5!?
58...&f4!! 54...&g6 also draws, but a lot less
58...&g4?! looks natural and loses. 59.&a7 entertainingly.
&g3 60. b6 f2 61.§cl! Delaying giving up 55.&c5 f4 56.&b6 f3
the rook until the last moment. (61.b7?
Sxa6f! does not win for Black, but with
such a gross blunder, White does deserve to
lose. After 62.S$?xa6 fl=Wf 63. S$?a7 W a l t
64.£b6 W b l t 65.&a7 ®xc2 66.b8=®t he
would still have to redeem himself in this
not entirely trivial, but still theoretically
drawn, ending.) 61...S$?g2 62. b7 White wins
by a tempo.
59.&a7 &e3 6O.b6 f2 61.Scl <±>d2!
Winning an important tempo.
6 2 . § h l §h3!! abcdefgh
White has two strong tries:

a) 57 .Sc2
Blocking the pawns from advancing, forcing
Black to find a few accurate moves.
57...5e3I!
Activating the rook and allowing the a-pawn
to run is the only way for Black to proceed.
58.a6
Another difficult moment.
58...§e6t!
abcdefgh
388 Conceptual Rook Endgames

Check the checks! The white king cannot game for White. And if the king goes to g6,
hide from them further up the board and White will play 60.&b7! and after 6O...Sh3
will have to retreat. This tempo is highly 61.a8=® h l = ® 62.Sxhl S x h l , he will have
useful for Black. 63.®g8f, winning. So, seemingly Black is
58...Se2? 59.a7 Sxc2 6O.a8=® f2 61.®f3 lost. Whatever he does, White plays &b7
h2 62. b5 &g6 (62.. .Bel? 63.®g2f! wins on the next move, promoting the pawn. But
both pawns) 63.®d3t &g5 64.®xc2 fl=® it is now time to talk about the downsides
65.®xh2 This endgame is objectively drawn, of 60.&b7. After this move, the queen on
but practically almost impossible to hold. a8 is temporarily out of play. The diagonal
In a classical game, it may be possible to a8-hl(g2) is blocked and there are no useful
draw against the a-pawn, but in rapid even checks. This downside means that if Black
this is difficult, as a recent high-level game could only keep status quo, he would make
showed. One mistake and you are lost. With the draw. And he can:
the b-pawn, people don’t draw in practical 59...5a4!! 60.&b7 Sh4!
games. One example is Deac - Shankland, The rook can also go to c4 first, but not f4,
Douglas 2019. as there is a dangerous check on a l . With the
59.&a5 Be2 6O.Scl h2 61. b5 £g6 62.b6 white rook on h l , this is not dangerous, as
Sa2f 63.&b5 Bb2f Black would have ...®xal .
Black makes the draw. White can try to 61.S h i
“sacrifice” by allowing Black to play ...Bc2t, 61.a8=® h l = ® t 62.Sxhl Sxhl draws.
but it is silly. White should still make the draw, 61...5f4 62.Sfl Sh4
but the risk is not worth it. Black holds.
More natural is:
54...&g6 55.&c5 &g5 56.&b6 &g4
b) 57. b5! f2 58.a6 h2 59.a7

abcdefgh
Black is seemingly lost. The attempt to be
creative with 59...Sc3 loses to 6O.a8=® abcdefgh
Sxcl 61.®g2f!. But we can quickly see that 57.a6?!
the king is perfectly placed in all variations. After this Black draws effortlessly.
If it goes to h7, White will eventually play
S h i and threaten take the h2-pawn with White could have caused more practical
check. 59...&h7? 60.&b7! Sc3 61.S h i ! . The problems for Black with: 57. b5! f5 58.a6 &g3
same would be the case if the king goes to 59.&a7 h2 60.Bel &g2 61.b6 Sc3! 62.Sxc3
f7, only now it will be Sxf2f that wins the h l = ® 63. b7
 Chapter 2 5 - Unbelievable 389

For a long time, I had this as a playing

exercise, which was solved only by the amazing
Adhiban Baskaran from Chennai. It was
around this time that I came to understand
that despite his hellraiser personality, Adhiban’s
greatest strength is perhaps his endgame play.
Quite an unusual thing for such a brilliantly
original and dynamic player.

53...113
abcdefgh This appears to be the most logical move.
63...®dl!l 64.§c7 ®d4f 65.&a8 ®d5 Black
draws. White will never be able to get the king 54.Sd8! h2
out and promote the pawns. There are too 54...f5 also fails. 55.&b6 h2 (55...f4 56.§d2!
many checks. shows the value of keeping the flexibility.
White wins with the normal moves 57. b5,
57...&g3 58.a7 h2 59.8cl &g2 60.&b7 58.a6 and 59.&a7!.) 56.§dl f4 57.Ehl! §a2
hl=® bl.Hxhl ‘i’xhl 62.b5 f5 63.b6 f4 58.a6 f3 59.a7! A little unusual. (The usual
64. &c8 13 65.b7 Sxa7 66.b8=W f2 67.®xa7 method, 59. b5? f2 60.&a7? loses half a point
*/2-*/2 per move, as after 6O...Se2! Black will queen a
pawn first.) 59...f2 60.&b7 &g6
As entertaining as this may be, the most
instructive part is looking at the best play.

53.&c5!
53.Hd8! h3 54. &c5 is a transposition, but
also really weird.

abcdefgh
61.b5!I This is the brilliant move. White needs
to queen the right pawn in the end; and in this
game, it is always the b-pawn. 61...&g5 62. b6
&g4 63.&c6 &g3 64. b7 &g2 65.Sxh2f!
&xh2 66.b8=®t &g2 67.®g8f White wins.

55.Sdl ±g6 56.Shl! Sa2

abcdefgh
390 Conceptual Rook Endgames

8
7
6
5
4
3
2
1
abcdefgh abcdefgh
57.£c4!! 54...±g6!!
This is the surprising move. White has not The king can also go to h6, but it looks
advanced his pawns too far. It is still possible weird. I should add that this is an excellent
to kick the black rook away from its perfect example of pushing the pawn too far being a
place o n a2. downside. Here it does not lose the pawn, but
it does lose a tempo, and thus the game.
57.' ’b6? would allow Black to draw by a
tempo. 57...&g5 58.a6 &g4 59. b5 &g3 55.±b6
60.&a7 &g2 61.Sxh2f &xh2 62. b6 f5 63. b7 What Black is looking for is the following
Sb2 64.&a8 f4 65.a7 Exb7! 66.&xb7 f3 variation: 55.Sh8 &g5 56. &b6 &g4 57. b5
67.a8=® f2 With a draw. &g3 58.a6 &g2 Yes, the pawn could have been
pushed too, but this beautifully illustrates the
57...&B 58.±b3 Sg2 59.a6 Sg3t 6O.±c4 point. 59.a7 h2 60. b7 h l = ® 61.Sxhl &xhl
Sa3 61.b5 Sa2 62.Sxh2! Sxh2 63.b6 62.a8=® Sxa8 63.&xa8 f5 64. b6 f4 65. b7 f3
I resigned in the training game against 66.b8=® f2 With a theoretical draw based on
Adhiban. stalemate.

It would be years before I realised there was 55.. .6g5 56.a6 h2!?
a serious mistake in my analysis. In the lines A little unnecessary, as 56...sl?g4 draws too.
given above, White wins an important tempo But I have a point to make.
when he plays 56. Sh 1!, as defending it with the
rook serves no additional purpose for Black. If 57.Sdl
Black was able to defend the pawn with the 57.Sh8 looks strong, but Black has
king, he would advance the king and defend 57.. .5a4!!, making 58.&B5 the only sensible
the pawn at the same time, making up the one move. After 58...Sa2 Black is by no means
tempo needed to secure the draw. This may worse. You could even imagine 59.&b6 being
all sound simple, when explained, but no one a repetition, although, having won a tempo,
ever came up with this logic before Stockfish Black can also draw with 59...&g4.
15 suddenly reversed previous evaluations of
the position with a glaring 0.00 evaluation, 57...±g4 58.b5 ±g3 59.±a7 ±g2 6O.b6
which I immediately investigated. White is arriving first it seems. But Black has
one trick left up his sleeve.
 Chapter 2 5 — Unbelievable 391

6O...Sd3! 61.Sbl Sb3! 62.Sxb3!? hl=® We join the next game at the moment where
63.b7 the players received an additional 30 minutes
towards the rest of the game, with the custom
30 seconds per move. White’s advantage is
obvious, but it is still a moment too early
to give up on the former Women’s World
Champion.

Florent Samoun - Antoaneta Stefanova

Sitges 2 0 2 1

abcdefgh
Black draws these lines. White will not be
able to escape the pin after:

63...®h7!! 64.&a8 ®e4! 65.a7 ®d5

With a draw.

abcdefgh
41.b4?
Organising the pieces was a convincing way
to win. 41.Scl ! White is ready to advance the
pawns. And after 41...Se3t 42.&c4 §xh3
43. a5 the win is trivial.

41...5e3t 42.<£d4 Sxh3 43.a5 Sa3 44.Sc3!?

It looked promising to try:
44.a6
Play continues:
44...f3 45. b5 £2 46.§cl h4
White’s position looks tremendous. The
curious thing is that after:
47.Sfl Ea2 48.&c5 h3 49.&b6 h2
392 Conceptual Rook Endgames

Clever. Black does not want to walk into

b8=®t.

51.b7 Sbl

abcdefgh
We have gone through a portal in time
and space, and entered the analysis to
Melkumyan - Grachev above.
50. a7!
5O.sl?a7? still loses to 5O...Se21.
5O...Sa4l! 51.&b7 Sh4 5 2 . g h l Sf4
Black draws. abcdefgh
52.<&a7?
52.§c2! was winning. Black had to walk
8
into the check with 52...&g3 to make
7 progress. Now White wins another tempo
6 with: 53.Scl !!, when he either wins a tempo
to queen the pawns, or forces Black to go for:
5 53. ..Excl 54.b8=®t &g2 5 5 > e 5 ! White is
4 winning, although there may be a few difficult
problems to solve before we get that far.
3
2 52...E2 53.Sf4 ±g2?
1 53...&g3! 54.Ef7 h4 was drawing. The
hoover variation is near.
abcdefgh
44...Sa4?
Black missed a simple but elegant drawing
variation.

44...Sxc3 45.&xc3 h4! The advance of the

pawns has to be accurately timed. 46. a6 f3!
47. ' >d3 h3 With the f-pawn blocking the long
diagonal, Black will get a queen too and draw.

45.&c5 £g6 46.&b5 Sal 47.&a6 <&g5

48. b5 &g4 49.b6 B 5O.Sc4f &h3!

abcdefgh
 Chapter 2 5 - Unbelievable 393

54.S16?? 61.b7 ®f7 62.®glt ±h3

This is a truly confusing mistake. I think
White was expecting 53...&g3 and in time
trouble was unsure about the difference, which
is that the white rook was hanging. Now it was
not and it did not have to be moved.

54.b8=® just won.

54.. .h4 55.Sb6 Sxb6 56.axb6 fl=®

57.b8=® ±h3 58.®c8t &g3 59.®c3t &g4?
Black was holding after something like:
59...6g2! 6O.b7 W and one or two accurate
extra moves.
abcdefgh
60.®d4t? White had a last chance to win the game.
White was winning after pushing the
pawn: 60. b7 ®f2f 61.&a6 Starting a nice
63.®e3t?
circle manoeuvre around the pawn to end The silent winning move was:
up on a8. 61...®flf 62.&b6 ® b l f 63. &c7 63. ®g51!
®h7t 64.& b 8 Black is out of checks. 64...h3 The queen is dominating a lot of important
65.®d4f &f3 66. &a8 White wins. squares. White wins in a way typical for
queen endings, where the king goes down
the board, to make it possible to block a
check with the queen, giving check back.
63...Sa2t
White is also winning after 63...Wf2f
64.&a6!, or 63...®h7 64.Se3t and 65.&a8.
There is an interesting trap with 63...®e6!?,
where White should not promote to a queen
due to stalemate tricks. Instead 64.®c5 is
the shortest win (38 moves).
64.£b6 ®b3t 65.&c6 ®c4f 66.®c5 ®e6f

abcdefgh
60.. .±g3?
Black missed a fantastic defensive resource:
60... 6h3!! 61.b7 ®gl!! with stalemate after
62.® x g l , and a draw also after 62.®b6 S a l t
63. ®a6 Sd4f 64. &a8 We4 and it is hard for
White to do anything quickly, while Black’s
h-pawn is not to be disregarded either.

abcdefgh
394 Conceptual Rook Endgames

67.£b5! ®b3t 68.Wb4 Sd5t 69.£a4 Wc6f 43.a5 Sa6 44.±B

70>b5 ® e 4 f 7 1 . £ a 3 A poorer try would be 44...&g7 45.&e4 &f6,
White wins. even though it is not without merit. White is
of course headed for b5 with the king, but the
63...£g2 64.®d2f £g3 65.£a8 ®f3 best path to get there looks like this: 46.&d3!
66. S e l f £h3 67.®e6f £g2 68.®a2f £g3 &f5 47.Sa4!, when White is winning as both
69.£a7 @e3t 7O.£a8 ®f3 71>g8f £112 g3-g4f and &c4-b5(t) are on the horizon.
72>g5 h3 73.£a7 ®a3t 74.£b6 ®b3t
75.£a7 ®a3t 76.£b6 ®b3f 77.£a7 45.£e4 £e7?
Vi-Vi This allowed White to put Black in zugzwang
(even though he missed it).
The following game from the Spanish Team
Championship offered the players a chance For this reason, we can go back and find that
to play incredibly deep chess, based on fine Black could have saved the game with:
nuances that had to be worked out in the 45...<±>e8!!
narrowest detail. If this is too much for human If White plays 46. &d4, Black will play
beings to do, we shall ponder another day. On 46... &d7 with a likely transposition to the
this day it was... analysis below. And after:
46. &e5
David Anton Guijarro - Alexei Shirov

Linares 2020

abcdefgh
46...&e7!
It is White who is in zugzwang! This may
seem like an extreme statement, but it is quite
clear that any White move would worsen his
abcdefgh
position. The same is the case for Black, as
41.Bxb2 Bxc3 42.Ba2 Bc6! already established. If White was to move
The rook has to block the pawn as early as the pawns on the kingside, Black would, as a
possible, if it will ever have a chance to come minimum, win time by attacking them.
out again. When the rook moves away to allow The same is the case after:
the king to take over the responsibility, the 47.Ea3 f6t 48.&d4 £d6 49.&c4 &c7!
pawn has to be a few steps from the 8th rank When both 50. b5 and 50.&b4 can be met
to make the transition possible. with 5O...Sd6!, where the rook will win an
396 Conceptual Rook Endgames

49.&b4! A horrible move, leaving the e6-pawn

The importance of this move has already unnecessarily frail.
been explained. The black rook should not
be allowed to occupy the 5 th rank with a 47...f6!? would hold; but the best move would
tempo. have been 47...&d7l, making no concessions.
49...Ed6
49...6b7 50.§d2 &c7 5 l . & b 5 and White 48. f4!
breaks through. 48.&c4 &c6 49.f4!, would transpose.
5O.a6 &b8 But an important little trick is revealed after
After 5O...Sd4f 5 1 .&c3! is important. Now 49.Ea3? e5 50.§al
the rook has to go back to a8 and Black loses.
51.&c5 Sd3
51... 5 d 5 t 52.&c6 is not an improvement
for Black.
52.§e2!
Making use of the extra tempo.
52.§b2f &a8 53.Sb7 is a more complicated
path to victory.
52...e5
52.. .5f3 53. &d6 will see the king aim for
the h6-pawn. Once again we see the damage
...g5 did to the black position. abcdefgh
53.Se3! 5O...f4!! 51.g4 &d6 Black holds - but only
just. After 52.& b 5 Sa8 53. a6 e4! 54.fxe4
&e5 Black has managed to create enough
counterplay on the kingside.

48...gxf4 49.gxf4 d7!

49...&c6 50.&e5 &d7 51.&f6 &d6
52.§d2f &c7 53. &g7 Exa5 54.£xh6 and
White wins.

50. ±c5?
abcdefgh A big step in the wrong direction.

White wins. The pawn ending is hopeless for The same can be said of 50.&e5, although
Black and if White is allowed to play Sf3, he after 50...&e7 it is easier to play 51.&d4, as
will win as well. 51.Ea4 §d6 52.§d4 §a6 53.Ea4 §d6 would
be an immediate draw.
46.. .±d6 47.B!
47.&c4?! &c7! 48.&b4 §d6! would allow
As we shall see below, the right move was
Black to activate the rook and leave the king to
50.&c4!.
block the a-pawn.
5O...±c7
47.. .B?
 Chapter 2 5 - Unbelievable 397

Back to where we were two moves ago.

52.<±>c4l!

abcdefgh
abcdefgh
51.±b5?
We need to check out three different
This very natural move is a surprising
defensive tries: a) 52...&c6, b) 52...&c8 and
mistake. Black is now able to save the game.
c) 52...&c7.
The check on d5, which will either push the
white king backwards or to the ugly a6-square,
a) 52...&c6
is the main issue.
This is rather cooperative, but is important
to understand first.
White has many other moves he could try, but
53.&b4 &c7 54.Ec2f! &b7 55.Ed2
only one worked. Whereas most squandered
the chances, it is important to have a quick
look at 51.Ea4?!, when Black has a great
defensive try in 5 1 ...&c8! and now:
52.&b5? makes no sense, as Black replies
with the stable ...Ed6.
52.Eb4?! is close to working, but Black
survives by a hair’s breadth after 52...Exa5t
53.&d6 g a l 54.&xe6 E h l 55.&xf5 Exh5t
56. &g6 Eh4!! and there is just enough time to
bring the king over.
White actually needs to wait with something abcdef gh
like 52.S a l l , when after 52...&c7 53. &d4 55...&c6
&d7 54.<&c4! we are on the way back to the 55...&c7 56.&b5 Ec6 does not put up a
winning path. lot of resistance. The key difference from
other lines where Black activates the rook
The right path was the road back to where we is that the king is not dislocated from b5.
were before. 57. a6 is most straightforward. The following
51. d4! <&d7 line is long, but not difficult. 57...Sb6f
51...6d6 52.Ea3 makes no negative 58.&a5 Ec6 59.Eb2! Ec5t 6O.Eb5 E c l
difference for White this time around. 61.Eb7t &c6 62.Eb6f &d5 (62...&c5
51...Ed6f 52.&e5 Ea6 53. &f6 also offers 63.Eb5t &c6 64.Ee5! leaves Black lacking
White no troubles. good options.) 63.Eb4 &c6 64.a7 S a l t
398 Conceptual Rook Endgames

65.Sa4 §xa4f 66. xa4 &b7 67. & b 5 &xa7 needed, and hopefully it is not so difficult
68. &c6 And White wins. for most.
56.§d8 §a7 57.Eb8 §d7 62...6xf4 63. a7 Sa8 64.&a6
Trying to activate the rook is the sensible try. This is why the king went to a5 rather than
After 57...&d5 58.&b5 &e4 59.&b6 Black c5. Had it gone to b6, it would have blocked
will soon have to give up the rook. The white the rook from reaching b8.
king arrives in good time to deal with the 64...e5
black pawns. Black’s problem is that he also
has to avoid losing the h6-pawn under bad 8
circumstances. 7
6
5
4
3
2
1
abcdefgh
65.Sb8 Sxa7t 66.&xa7 e4 67.&b6
White wins.
abcdefgh
b) 52...&c8 53.<±>b5 Ed6 54.Sg2! S d 5 t
58.§b6t!
An important check. Without it, White
would not be able to win.
58.. .6d5 59.a6 &e4 60.§b7
6O.Sxe6f &xf4 61.Sg6! &e4 62.Sgl! f4
63. S a l wins by a tempo and is rather flashy,
but also appears somewhat unnecessary.
60... 5 d l

abcdefgh
55:&c&.
It is always nice to threaten mate.
55...&b8 56.gg8t &a7 57.Sh8 Sd4 58.Sxh6
£a6
58...Sxf4 59. b5! Sh4 6O.Sh7t &b8 61.h6
and Black cannot deal effectively with the
abcdefgh two wing pawns.
59.Sxe6 Sxf4 6O.&d5t &xa5 61.Sf6 &b5
61.Eb5!Sd8 62.&a5!
62.&e5
This line also only wins by a single tempo,
White wins. The black king is far too far
but it is only here that any kind of finesse is
away.
 Chapter 2 5 - Unbelievable 399

c) 52...&c7
This is the trickiest try.

abcdefgh
59.Be3!
Surprisingly, only this move wins. The idea is
abcdefgh
simple. Either White is allowed to take on e6
53.Sa3! (black checks to the white king matter not), or
Necessary prophylaxis. Now after 53... Sd6, 59... d7 is met with 6O.Ha3, when Black will
White has 54.Sd31. have to use the rook to block the pawn, losing
53.. .6c6 on the spot.
53.. .6c8 54.&b5 Sd6 55.Sg3 Sd4 56.Sg6
Sxf4 57.Sxh6! This is the only winning
move, but it should be an easy one to find.
If White was to take the pawn on e6 first,
the rook would then block the h-pawn
afterwards. Not very meaningful. 57...Sfl
58.Sxe6 f4 59.h6 S h i 6O.a6 & b 8 61.Se8f
<&a7 62.Se7t b8 63. h7 White wins.
54.<&b4 &c7 55.Sc3t &b7 56.Sd3
The most natural is to not allow the black
rook to get activity through the d-file,
although 56.Sg3 Sd6 57.&c4! also wins.
56...6c7
56...5c6 57.Sd7t a6 58.Sd8 leaves Black abcdefgh
without a decent move. 51...Bd6!
57.&b5 Sc6 58.a6 Black manages to activate the rook and
58.Sg3!? Bel 59.Hg6 also wins. And maybe create enough counterplay to hold the game.
by a bigger margin.
58. . .Bel 52.a6
52.Hg2 Hd5t 53. <&a6 e5 and Black draws.

52...6.8 53.a7t &a8 54.Ba4

54.Hg2 <&xa7 55.Hg6 Hd4 also holds.

54.. .Bdl 55.&c4 S e i 56.&d3 S h i 57.Sa6

S e i 58.Sa4 S h i 59.Sa6 S e i 6O.Sa4
Vl-1/!
400 Conceptual Rook Endgames

The next game is from the Chinese

Championship, played in the period where
China looked as if they had escaped the
worst of the pandemic, but were living under
severe restrictions. At this point, playing such
a tournament must have been a relief to the
players.

Yu Yangyi- Liu Yan

Xinghua Jiangsu 2021 abcdefgh

Now this cannot be prevented. For this
reason, Black may as well take the pawn.
37.. .gxf5 38. h4 Sb7t Tricky. 39.&a4! The best
square. White wants the king to go to a5, but
he also does not want Black to play ...Sb3.
39.. . 5 b l 4O.Sa7t White wins after 40...&e8
41.<&a5, or 40...<&e6 41.Sc71, where we can
see another advantage of giving up the f-pawn.

37...Sb7 38.Sa8
38.f5! was still strong. This was the last
chance without Black accepting a repetition.
abcdefgh White is still winning without it, but only
because Black is trying to avoid it.
What is characteristic for this game is that
Yu Yangyi missed or misevaluated a key idea
38.. .5b5t!?
for quite a number of moves (including the last
A check meant to create counterplay.
few moves), where he could have played it at
any moment. In the beginning it makes sense
38.. .6e6! was tempting too, when White
not to play it too early, but then it becomes
would have to find 39.Sa6 <&d7 40.f5!.
late, and finally too late. Only then does Yu
play it!
39.&a6!
In this phase of the game Yu Yangyi plays
35.£c3 £e7 36.&B4 &d7 37.&a5 fantastic chess.
To me this was the perfect moment to play
37.f5!» as it is not possible to create a passed 39.&a4? S b 1! would give Black counterplay.
pawn on the kingside by advancing the g- or
h-pawn.
39...Sb4 4O.Sa7t! &e6 41.Sc7 Sxd4
 Chapter 2 5 - Unbelievable 401

8
7
6
5
4
3
2
1
abcdefgh abcdefgh
42.&b7! 43.&xc6?
Excellent play by Yu Yangyi. The king needs The natural follow-up, but also a significant
to enter the game in order to keeping the Black mistake.
king at bay.
What was difficult to see was that 43.Hxc6f!
42.Hxc6f? d7! 43.Sxf6 looks natural, but was stronger, as it forces the black king away.
43...Hd2 gives Black enough counterplay. The 43. . .6f5 44.Sc7 d4 45.c6 The white pieces are
critical variation is quite instructive. 44. h4 perfectly coordinated. White wins. 45...Hb4f
Sa2t 45.&b5 Sb2f 46.&a5 46.&a6 d3 47.Sd7 &e4 48.c7 Hc4 49.&b7
Hb4f 50.<&c8 for example...

43.. .d4 44.Sc8 &e7 45.Sc7t &e6 46.f5 1 1?

Finally, Yu plays this idea. But at this point it
is little more than a tricky try.

abcdefgh
46... <&c7! Prophylaxis, avoiding a check on d6.
(46...d4? loses to the surprising 47.Sd6f &c7
48.Hb6!, when the white king gets out of the
corner, or where White wins after 48...Hxb6
49.cxb6f b7 50. b4! and the extra pawn
decides the game.) 47.Sd6 Hg2! Black is on his
way to making a draw. abcdefgh
46...gxf5?
42...Sc4
 Chapter 2 5 - Unbelievable 403

38...g3?
A terrible move. Black is seeking a way to
bring the rook back, but it’s not the right way
to exchange the most advanced pawn; the asset
Black should seek to win with later.

38...Hxh2! was the move Black would want to

play. And should have played. White will not
be able to promote the b-pawn without Black
being able to give up the rook along the way.
abcdefgh For example: 39. b8 g3 40. b7 Hc2, winning.
White is ahead in the race, which eventually
will lead to a draw with a comfortable margin 39.hxg3 Sxg3
in both directions.

35...Sx£2 36.b6 Sa2f

A sensible check in time trouble, which
mainly seeks to repeat moves.

36...Hxg2! was simpler.

37.&b7
I am confident that after 37. b5 Black
would have played: 37...2b2f 38. <&a6 Hxg2!
39. b7 Ha2f 40.<&b6 Hb2f followed by
advancing the pawns on the kingside.
abcdefgh
37...Sxg2 38.Sb4 40.&a6?
The more natural and stronger option was:
40.&a7!
If Black gives a check on a3, the white king
goes to b8, when Black would have to waste
a tempo to play ...Hg3-g8 to stop the b-pawn
eventually.
40...e5
40...&c5 41.b7 is also fine for White. A
small point is that after 41...Ha3t 42.<&b8
< ?xb4?! 43.&c8 White will queen, although
Black still draws with the plan of ...Se3-e4.
41. b7 Sg8 42.b8=®t Hxb8 43.Sxb8 e4
Black can draw in quite a few different ways.
abcdefgh To focus on one over the others makes little
sense.
406 Conceptual Rook Endgames

37.. .5a5!! 38...Sa4!!

Mission accomplished. The rook is returning Black holds. The rook maintains position on
to the ideal post on a4. the a-file, and sooner or later White has to try:

37.. .Hc5 would lose to 38.Hd4!, as the Black 39.Sc8 f5t 40.&h4 Sxa3 41.Sc6f &g7
rook needs to take the a-pawn on the third
rank and the black rook would get stuck on a5
in a sort of zugzwang.

38.Sc3
38.Hd8 f5t goes the same way, but would
deprive us of seeing the rook returning to a4.

abcdefgh
Finally, we can all see why h2-h3 had to be
provoked: the g3-pawn is loose, and Black
would be able to capture it if the white king
were to advance.

abcdefgh
Epilogue

Concepts to Live by

All of these recurring concepts are presented to you in short and sharp wording. This is done
with intelligence implied. I presume you are intelligent, and you will show me the courtesy of
returning the favour. These concepts are not rules to mindlessly follow, but recurring themes of
powerful value, if understood properly and implemented intelligently. In other words, without
reading the book first, this will seem like a senseless list of platitudes.

This is an (incomplete) list of the most important general observations (concepts):

Always have at least two serious options on every move (not endgame-specific).
Don’t make mechanical “normal-looking” moves without thinking. Chess is a rich game with
stunning surprises possible on every more.
When you can, keep your options open.
Learn the theoretical positions (read Theoretical Rook Endgames by Sam Shankland).
Endgames are mainly about promoting passed pawns.

Rook endings have a reasonably large drawing margin - but are not all drawn...
Passed pawns become more important as they advance up the board.
Passed pawns must be pushed.
In rook endings we should seek to activate the rook first.
The rook is strong when placed behind passed pawns.
The king naturally belongs on the side of the pawn where the opponent’s king is approaching
(shouldering).
The rook vs pawn endgame is, at its core, a race; and for this reason we need to use every
finesse to win time.
Often we need to delay giving up the rook for the opponent’s passer to draw the opponent’s
king as far as possible away from the finish line.
If your opponent is waiting for you to initiate the race, strengthen your position to the
maximum before you fire the starting gun.
Usually, when the weaker side gives up the rook for a passed/promoted pawn, we should
recapture with the rook, to get the rook behind other passed pawns, and to keep the king
closer to the other end of the board.
412 Conceptual Rook Endgames

H K
Hamitevici 296 Kalod 74
Hammer 54 Kanarek 234
Hansen, E. 368 Karasev 311
Hansen, M.S. 402 Karpov 8
Haring 353 Kasimdzhanov 32
Haubro 190 Kasparov 15, 112
Hauge 236 Kelly 49
Hausrath 52 Keso 405
Hayrapetyan 356 Keymer 29, 47, 54, 55, 309, 310
Hebden 16, 229, 325 Khademalsharieh 75
Hector 352 Khismatullin 265, 266
Hernandez 20 Kirk 185
Hillarp Persson 378 Kizatbay 119
Horvath 197 Klepek 208
Horwitz 111 Kling 111
Hovhannisyan 239 Kollars 123
Howell 172 Kopaev 101
Hracek 222 Korchnoi 8
Huschenbeth 278 Korley 101, 168
Hvenekilde 293 Korneev 165
Korpa 154
I Kovalenko 319, 322
Inarkiev 225, 226, 265, 266, 267 Koykka 173
Iniyan 285, 326 Kozak 183
loannidis 285 Koziorowicz 105
Ipatov 208 Kramnik 61,219
Ivanchuk 137, 157, 159 Krasenkow 129
Krasikov 287
J Krastev 108
Jacobson 33 Kravtsiv 109, 304, 372
Jahncke 347 Krikheli 39,217
Jaksland 153 Krivenko 201
Jakubowski 105 Kryakvin 323
Jani 49 Kuljasevic 220
Jankovic 45 Kupervaser 27
Jeevitesh 326 Kushko 223
Jensen 324
Jianu 303 L
Jogstad 40 L’Ami 122
Jones 347 Lasker 135
Jumabayev 200, 302 Laurent- Paoli 353
Ju Wenjun 368 Laznicka 109, 381
414 Conceptual Rook Endgames

Popov 178, 179 Sebag 176

Postny 126 Sekar 207
Powierski 167 Sethuraman 325, 326
Praggnanandhaa 148 Sgircea 52
Priehoda 74 Shabalov 327
Psyk 36 Shah 343
Punsalan 180, 181, 197, 349 Shankland 4, 6, 2 0 , 2 1 , 53, 87, 109,
118, 126, 128, 165, 202, 226,
Q 227, 289, 290, 300, 315, 385, 388
Quang Liem 47, 198, 341 Shaw 20, 189
Quparadze 257 Shearsby 166
Shepley 244
R Shevchenko 319
Rabinovich 241 Shirokov 236
Radjabov 175 Shirov 394
Rafiee 280 Sindarov 154, 231
Rakhmanov 34, 242, 383 Sinquefield 327
Rapport 145 Sivuk 233
Real De Azua 233 Sjugirov 259, 355
Recabarren Gajardo 82 Smyslov 121
Reis 307 So 56, 283
Reti 82, 84 Socko 43
Rewitz 402, 404 Sodoma 110
Roy 50 Sokolin 327
Rubin 236 Solozhenkin 178
Sosa 184
S Soto 326
Sachdev 180, 197, 304, 349, 383 Sousa 307
Sadzikowski 43 Sprenger 60
Salgado 383 Stefanova 391
Samoun 391 Stefansson 198
Samunenkov 195 Storme 378
Sandipan 198 Stukopin 165
Santos Latasa 191,235 Sutovsky 122, 123, 315,
Santos Ruiz 308 316,318,319
Sarana 11,12, 22 Svane 357
Sargsyan 200 Sylvan 322
Saric 45, 194 Szpar 48
Sasikiran 355
Schekachikhin 269
Schroeder 253
 Name Index 415

T w
Tai 215 Walton 339
Tan 192 Wang Hao 134
Tancsa 86 WeiYi 116, 120, 402
Tang 123, 191 Womacka 347
Tari 235
Tarrasch 39, 44, 373 X
Terbe 94, 95 Xiong 56
Thybo 61,354, 384 Xu Yinglun 205
Tiviakov 211
Tomashevsky 362 Y
Topalov 110, 111, 112 Yaniv 288
Torres 65 Yankelevich 191
Triapishko 103 Yogit 117
Trumann 106 Yoo 33, 149, 349
Tsay 230 Yusupov 15, 309, 333
Tyurin 254 Yu Yangyi 400, 401,402

U Z
Urh 86 Zeng Chongsheng 221
Urkedal 50 Zhao Jun 221
Usmanov 280 Zhou Jianchao 116
Zhukov 67
V Zwirs 290
Vachier- Lagrave 32, 127, 171, 283
Vajda 208
Vakhidov 48, 53, 154
Vallejo Pons 177
Van der Heijden 250
Van Foreest, J. 46, 89
Van Foreest, L. 139
Van Kampen 211
Van Wely 263, 352
Vasiukov 213
Vestby-Ellingsen 295
Vidit 145
Vila Gazquez 132
Vinter-Schou 347
Vitiugov 240
Vlasak 149
I hope you enjoyed this book. I put my heart and soul into it. I strongly recommend
that you look at the other side of the coin, Sam Shankland’s Theoretical Rook Endgames.
The books were written as a pair. We hope you like them both.