Positional

Sacrifices

Neil McDonald

CADOGAN CHESS
LONDON, NEW YORK
CADOGAN BOOKS DISTRIBUTION

UK/EUROPE/AUSTRALASINASINAFRICA
Distribution: Grantham Book Services Ltd, Isaac Newton Way,
Alma Park Industrial Estate, Grantham, Lines NG31 9SD
Tel: (01476) 67421 Fax: (01476) 590223

USNCANADA/LATIN AMERICNJAPAN
Distribution: Paramount Distribution Center, Front and Brown
Streets, Riverside, New Jersey 08075, USA
Tel: (609) 461 6500 Fax: (609) 764 9122

First published 1994 by Cadogan Books pIc, London House,

Parkgate Road, London SW11 4NQ

Copyright © 1994 Neil McDonald

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, mechanical,
photocopying, recording or otherwise, without prior permission in
writing from the publishers.

British Library Cataloguing in Publication Data

A CIP catalogue record for this book is available from the British
Library.

ISBN 1 85744 1109

Cover design by Brian Robins

Typeset by B. B. Enterprises
Printed in Great Britain by BPC Wheatons Ltd, Exeter
Chief Advisor: Garry Kasparov
Editor: Andrew Kinsman
Russian Series Editor: Ken Neat

For a complete catalogue of CADOGAN CHESS books (which

includes the Pergamon Chess and Maxwell Macmillan Chess lists)
please write to:
Cadogan Books pIc, London House, Parkgate Road,
London SWll 4NO Tel: (0171) 738 1961 Fax: (0171) 924 5491
Contents

1 Introduction: The Psychology of Sacrifices 7

2 Theoretical and Standard Sacrifices 17

3 Sacrifices to Create a Passed Pawn 27

4 Sacrifices to Destroy the Opponent's Centre 42

5 Sacrifices to Open Lines 57

6 The Indian Bishop 81

7 The 'Karpovian' Exchange Sacrifice 97

8 Queen for Rook and Bishop Sacrifices 106

9 It's Your Turn 119

 1 Introducti o n: The Psychology of
Sacrifi ces

It has never been more difficult to This brings us to the subject of

win a game of chess. Even at club our book: the most difficult posi­
level, many players have a pro­ tions to judge in chess are those
found knowledge of opening the­ with a material imbalance and dy­
ory, which often reaches well into namic chances for both sides. A
the middlegame. Informator is sacrifice disrupts the equilibrium
universally available, and there of the position, and can disturb
are now a host of specialist and upset the opponent. This is
monographs and magazine arti­ especially true if it sets new and
cles. The gap between the En cy­ unexpected problems. The oppo­
clopaedia of Chess Openings and nent may be in a state of shock
the Encyclopaedia of Chess and react in an inappropriate way.
Endings is narrowing at an This is what Rudolf Spielmann, a
alarming rate. Viennese grandmaster famed for
So we have a problem to solve. his attacking prowess, meant when
Assuming there is no great differ­ he remarked that a sacrifice should
ence in playing strength, how can not be judged according to its
we beat our equally knowledge­ soundness, but rather according to
able and well-informed opponent how dangerous it is.
in the next club game or round of Here are some examples of
the tournament? world-class players being be­
Winning depends on at least mused by a sacrifice.
one mistake by the opponent, and
probably more. (Jon Speelman
K i r.Georg i ev-J.Pol g a r
assesses it as two medium-sized
Budapest (zonal) 1 993
mistakes before Black can expect
to lose, and three for White; such (see diagram overleaf)
is the advantage of the first
move.) So the best method of White played the spectacular 22
play is the one most likely to lhf7! �xf7 23 ii'h5+ and here of
make your opponent falter. course Black played 23 . �e7 . .
8 Positional Sacrifices

moving the king from the open you are one of the best players in
file and keeping the rook on e8 the world. It is even more diffi­
defended. There followed 24 cult to think calmly and objec­
'it'xh7 �d8 (the bishop is inde­ tively when your opponent has
fensible) 25 'ii'xg7 and Black had just sacrificed a rook. One should
no good answer to the advance of not underestimate the demoralis­
the White g-pawn since her ing effect of a sacrifice on the
pieces are cut off from the king­ opponent. The following rein­
side and her king is stuck in the forces this:
middle. White won on move 47.
Dolmatov-Lutz
Germany 1993

Judit Polgar is a fierce attacking

player herself with a fme feel for
the initiative. Therefore it is sur­
prising at move 23 that she did not White, who has already sacri­
find 23 . . . �g8 ! returning the rook. ficed a knight and a pawn, con­
Then after 24 'it'xe8+ It:)f8 25 c5 tinued with 29 :xh7+ �xh7
.tb7 (as given by Georgiev) 26 (what else?) 30 It:)e7+ g6 (30 . . . d3
'ii'h5 �4! Black has play for the 3 1 'iWh4 or 31 'iWh3 mate) 3 1
sacrificed pawn. In any case, the �4+ �g7 32 'it'xf6+ �h6 33
game continuation is so obviously 'ilh4+ (a little repetition to gain
hopeless that Black would surely time on the clock and clarify his
have chosen 23 . . . 'itg8 if she had thoughts) 33 . . . �g7 34 'ii' f6+ �h6
considered it at all. 35lt:)xg6! (D) .
This is a clear case of sacrifi­ Now Black cannot prevent a
cial shock and stereotyped quick mate. Note the enormous
thinking. It is difficult to con­ power of the bishop on b 1 . If
vince yourself that it can be cor­ 35 . . . d3 then 36 It:)xe5+ �h7 37
rect to allow your opponent to 'iWg6+ and 38 It:)t7+ wins the
capture a rook with check, even if queen with mate to follow.
 Introduction: The Psychology of Sacrifices 9

Black's pieces drop one by one as mate) 36 lLlfS+ �g8 37 lLlxd4

he attempts to stave off mate: ibxd4 38 'i'h7+ (not 3 8 'fIxd4?
3S . . . �xg2+ 36 �xg2 ibg4 37 l::Id 8 39 'fIg l ':'d2) 38 . . . �f8 39
'i'h4+ �g7 38 'i'xg4 ibgS 39 'ilxc7 ':c8 40 'ii'd6+ �f7 4 1
'i'xd4+ ! .i.eS 40 'ii'x eS+ :xeS 4 1 'ii'x d4 ! ? ':c l+ 42 'ifg l lbg l +
�xeS+ and Black resigned, since with a level endgame.
he is a piece and two pawns down Over the board, with the clock
after 4 1 . . . 'i'xeS . ticking, few players would be
willing to embark on such a risky
adventure as 29 .. . c�g8. Yet, as
with the previous game, we can
point out that the chosen con­
tinuation led to a clearly hopeless
position within a couple of
moves. Black had to bravely
control his nerves and resist the
temptation to capture the rook. A
little calm objective analysis
would have shown where the true
chances of safety lay.
Among modem players, Alexei
Let us return to the position Shirov of Latvia is renowned for
after 29 ':'xh7+. Is it more natural his fearless attacking play. He
to capture the opponent's rook always seems able to tum a
than give up your queen? Maybe, placid position into a dynamic
but Black should nevertheless try fight, where material or posi­
29 ... c.tg8 ! . Then 30 ibh6+ c.tf8 3 1 tional deficiencies are of secon­
':'xf6+ �e7 ! (not 3 1 . . .gxf6 32 dary importance compared with
'i'g8 mate) 32 ':'xg7+ (one can king safety or tactical factors.
sympathise with Black not want­ Sometimes he is punished for his
ing to have his king buffeted by gambling style, but more often
white pieces, but Black has one than not it is his opponent who
consolation - his king has avoided collapses under the pressure. We
the deadly power of the bl shall now examine some exam­
bishop) 32 . . . �xf6 33 :Xc7 �xc7 ples of Shirov's enterprising play.
34 'i'h4+ �g7 35 ibfS+ rJ;f7.
Here Dolmatov's analysis stops
Sh lrov-Hjartarson
in Informator 57, with the helpful
L ucerne 1993
conclusion that it is unclear. The
aggressive 3 5 �xd4 ! ? also leads (see diagram overleaf)
to equality after 35 . . . ':'xd4 (not
35 . . . ibxd4 3 6 'i'gS+ with a Quick There is only one open file on the
 10 Positional Sacrifices

board, and although it is in moves in a blocked position. In­

White's possession, the entry stead, he finds he has a lot of hard
points into Black' s position, h7 thinking ahead of· him as he
and h8, are well defended. Black judges how dangerous White's
has just played 32 . . . ttJg6, attack­ threats to his king really are. Per­
ing the rook. One imagines that haps objectively he has nothing to
White will move the rook, say to fear, but the practical difficulties
h6, when in view of the blocked are enormous.
nature of the position and lack of 35 :tb7!
aggressive pawn advances for Black begins well enough. He
White, a draw seems likely. In­ brings his inactive rook to the
stead White played: defence of his king. But how
much of Black's precious remain­
ing time did it take to find this
manoeuvre?
36 lbg3 1:[17
37 lbh5 'iib2+
38 '1f.;f3

33 l:txg4! fxg4
34 'ikh5 lbf8
35 'iixg4
Black's king suddenly feels a
shade draughty. The white bishop
on d3 may strike a blow along the
newly opened diagonal; the ad­ 38 'ii h 2?
vance of the g-pawn may be dan­ In time pressure, Black not
gerous; and the knight manoeuvre surprisingly attempts to check
lbg3 -h5-f6 may be strong. Sud­ White's king. 'If White is in
denly Black faces a new set of check, he cannot land a devastat­
problems which have no easy ing blow, ' thinks Black, ' and
solution. In time pressure, this is once I reach move 40, I will have
very disagreeable. He was time to work out just how dan­
probably expecting to be able to gerous White's threats really are.'
reach move 40 (the time control) 38 . . . .tc3 was better, returning
with a series of nondescript the bishop to active play. Then 39
 Introduction: The Psychology of Sacrifices 11

lDf6+ .ixf6 40 gxf6+ �h8 4 1 the white bishop moves from e3

'ii'hS + �g8 42 "'g4+ would then 29 . . ..tf4 wins. A more ag­
.

draw. It is difficult to see how gresive non-sacrificial continua­

White can continue playing for tion is 27 "'c5 when 27 . .. lDxe3
advantage after 38 . . . ..tc3. 28 l:th5 g5 ! ? 29 "'xe3 .if4 30
39 g6 "'hl+ "'d4 (not 30 "'cS "'xcS+ 3 1
40 � 'ii'h2+ l:xcS ..te3+) 30 . . . 0-0! 3 1 l:tc5
Black has reached move 40 - 'ii'g 6 32 bxa6 l:fd8 again leaves
but he lost on time as he did so. White with menacing passed
After 41 �f1 (4 1 �f3 "'h l + is a pawns, though once again his
draw) 4 1 . . :i'h l +? 42 ..tg l l:e7 king is open to attack. Therefore,
43 .ie4 ! ! wins Black's queen Shirov decides to sacrifice his
because of the fork on f6. A bet­ queen in order to gain the initia­
ter fighting chance is 41. . .l:e7 42 tive. Note that in the variations
g7 (with the threat of 43 lDf6+ above, White's king proves a tar­
queening and if 42 . . . lDd7 43 get while Black's is perfectly
.ih7+ ! ) 42 . . . ..tc3 ! ? 43 gxf8�)+ safe. In what follows, the situa­
�xf8. White must still tread care­ tion is reversed.
fully, e.g. 44 f5 'ii'hl+ with an
obscure position, but the two
pieces should ultimately prove
superior to the rook after 44 .i f2.
However, Black's flag fell and
that automatically finished the
game. It was the problems set by
33 lbg4 ! which pushed Black
over the precipice.

Shl rov-Stohl
Germany 1 994
27 "'xdS! ..th2+
A very interesting position. 28 l:txh2 'ii' xdS
White's queen is attacked and if 29 :hS!
he retreats then play could con­ Black must not be allowed to
tinue 27 'ii'd 2 0-0 (27 . . . lDxe3 28 castle, which would safeguard his
'ii'xe3 ..tf4 29 'ii'f2 "'gS 30 lth5 ! ) king and co-ordinate his rooks.
28 bxa6 l:fdS. White's connected For example, if 29 bxa6 0-0 and
passed pawns may look impres­ Black can play . . . 'ii'e5 followed
sive, but Black has very danger­ by advancing his f-pawn to attack
ous threats. The immediate one is White's king.
29 . . . lDxe3 and 30 . . . ..td4, and if 29 fS
 12 Positional Sacrifices

Not a move Black wanted to by 3S . . . �b8, since White simply

play. However, he would still plays 36 a4, defending the
have been deprived of the right to bishop, when he threatens both 37
castle after 29 . . . 'iVe6 30 :c6 'iVa2 lId6+ and the further advance of
3 1 1::te S+, etc. the a6 pawn. Black's best defence
30 bxa6 is 3s . . .'i'd l + 36 �h2 1!Vg4! as
Now of course 3 0 . . . 0-0? loses pointed out by Eising. Now 37
the queen to 3 1 .li.c4. If 30 . . . 'iVe5, i.b6+ cJi;e7 (any other square
White can play 3 1 .li.c5 (stopping leads to a discovered check at­
castling and threatening Jt bS+) tacking the queen) 38 l:tc7+ \t>f8
and if here 3 1 . . .�f7 then 32 39 .li.cS+ �g7 40 .li.d4+ �f8
1::txfS+! 'iVxfS 33 1::t£ 1 �x£1 + 34 (forced) and it's difficult to see
�x£1 and the passed pawns and how White makes progress. If
bishops defeat the two rooks. White plays slowly (i.e. without
This variation demonstrates the giving check) then Black will be
enormous power of White's able to play .. . 'iVf4+ or . . . f4-f3
queenside pawns. starting his own attack. In any
30 g6 case, the game came to an abrupt
Safeguarding the f-pawn and end after
preparing 3 1 . . hS followed by
.

. . . kth7. Black must get his king's

rook into the game somehow, but
these pawn advances weaken
Black's second rank and his dark
squares in general.
31 kIh3 h5
32 l:tc7
The familiar rook on the sev­
enth rank.
32 :f8
33 Jtc5 :f7
Black has achieved his aim, but
White's connected passed pawns 35 llc8?
and dominant bishops are a lethal 36 lld6+ 1-0
force. A simple oversight, but the ter­
34 .li.b5+ �d8 rible difficulties Black has faced in
35 l':tc6! (D) this game no doubt contributed to
White is not interested in ex­ the blunder. It is a frequently re­
changing rooks. The sly retreat curring theme that the psychologi­
hits the g-pawn and also intro­ cal pressure inflicted by a sacrifice
duces the idea of i.b6+. This leads to a collapse later on in the
cannot be successfullv countered game.
 Introduction: The Psychology of Sacrifices 13

Connected with the theme of 24 .. .'i'd7 25 exd5 'ii'xd5 26 :e4

setting the opponent problems is followed by 27 ':fe l . Of course
that of finding the path of maxi­ 24 . . . dxe4? 25 fxe4 would be
mum resistance in a bad or lost complete positional capitulation
position. A sacrifice can be an by Black. White would gain the f­
ideal positional trap, since the file and a strong centre. Almasi,
resulting positions can be irra­ playing Black, is fully aware of
tional and hard to judge. It is the danger he is in. He does not
easier to be 'lucky' in such a fancy a slow death after
position. 25 . . ....d7. So what should he do?
Black has two sources of hope
in this position:
Yusupov-Al masl
i) White has built his centre
Altensteig 1 993
at the cost of weakening his
king's pawn cover. As long as
White is in control, this weakness
is irrelevant, but should Black
gain the initiative, White may
have cause to regret his liberal
use of pawns.
ii) White's centre is not set in
steel but can be undermined by
.. . b5-b4.
In such positions, it is neces­
sary to be objective - even coura­
geous - and admit that the posi­
tional build-up has gone wrong.
White began to conquer the The only hope of safety is in
centre with counterattack and sacrifice.
23 g5 lDh5 Therefore, Almasi played:
24 e4! 24 b4!
The two bishops should prove 25 axb4 axb4
significantly stronger than the 26 cxb4
black knights as the game opens Or 26 exd5 bxc3 27 .i.xc3 lLJf4!
up. Yusupov had foreseen that (threatening . %3+) followed by
..

24 . ..lLJf4 25 'ifg3 lLJd3 26 'ifxc7 ...lLJxd5 with a good game. Note

':xc7 27 �xd3 cxd3 28 exd5 is that in this variation, the knight on
winning for White. Meanwhile, h5 suddenly fmds itself on an ex­
White threatens 25 exd5 . It seems cellent centre square.
he will acquire a protected passed 26 "'b6
pawn on d4 and the e4 square as 27 �a4
an outpost for his pieces, e.g. If 27 exd5 lLJxd4 threatens
14 Positional Sacrifices

. . . lDxf3+. Yusupov's move forces looking piece, and White is gen­

Black to sacrifice another pawn erally weak on the white squares.
to keep the momentum of his He sorely misses the bishop on
counterattack going. d3. And because Black's pieces
27 :'a8 are so much better co-ordinated,
27 . . . dxe4 28 fxe4 lIxe4 29 Black's c4 pawn is much more
'iWxf7+ �h7 30 .li.c2! 'ii'xd4+ 3 1 menacing than any of White's
�h l I:txe l 32 'iWxg6+ �h8 33 own passed pawns. So we can
'iWh7 mate is a variation that conclude that Black has excellent
demonstrates White's attacking compensation for his pawns.
potential if Black is careless. What prevents him from having a
28 .ltxc6 'iWxc6 substantial advantage is the off­
29 exd5 'iWd7 side state of his knight on h5.
30 J:lxe7 'ilxe7 31 'iWe3 'i'id7!
Eyeing the h3 square. Of
course, 3 1 . . . 'ilxe3? would be a
massive positional blunder. Black
needs to keep the queens on if he
is to generate attacking chances
against White's king.
32 :e1 �h7
Black's compensation for the
pawns is based on solid, perma­
nent features of the position. So it
will not disappear if he spends a
move safeguarding his king.
33 'iic3
Black is now two pawns down. Yusupov defends well. He sets
However, White's once proud up a blockade on the dark
pawn centre is in ruins. The black squares. At the same time he
rook is more actively placed than makes sure that the black knight
its white counterpart, and is pre­ on h5 doesn 't get back into the
pared to swoop down to ' seventh game: f4 is always kept guarded.
heaven' ( .. J::ta 2). The absence of 33 'ii' h3
pawns on White's second rank, With the plan of 34 . . . l:ta2,
save the h-pawn, means that threatening 35 . . . lDf4! and mate
Black's queen and rook can hope on g2 whether or not the knight is
to infiltrate into the core of captured.
White's fortress through the gaps. 34 %:te2! �a2
Meanwhile, Black's king is per­ 35 lIf2
fectly sheltered behind his pawns. A notable defensive manoeu­
The hishon on d2 is a feeble vre by White.
 Introduction: The Psychology of Sacrifices 15

falls apart.
36 'tixc4??
White's fine defensive play
certainly deserved a better fate.
We can surmise that White was
in time trouble, and was dream­
ing of the initiative when he was
hit by 35 .. :1'114 ! ! . With only sec­
onds to think, Yusupov snatched
the c-pawn, but how should he
have met the threat of 36 .. . lLlf4
here? In the cold light of study,
he notes that White should play
Has Black's attack come to an 36 1i'e3 l:a l + 37 �fl �xfl+ 3 8
end? The bishop on d2 fends off 'itxfl 'tixh2 leading t o unclear
the knight on h5; the queen on c3 play, or 36 :g2 'tih3 ! 37 1:1f2
keeps the rook out of a 1 ; and the 'ii'h4 with a draw by repetition.
rook on f2 secures the kingside. However, even in his post­
White is now ready to push his b­ game analysis Yusupov is still
pawn. Once he has assumed the fighting the ghost of the queen on
initiative, he can compel Black's h3 . After 36 'ii'e 3 l:1a l +, as he
pieces to retreat to less aggressive suggests, White can play 37 'itg2 !
squares. This would be the be­ and Black has no convincing way
ginning of the end for Black. In­ to continue his attack, e.g.
stead, Black found a way to keep 37 . . . lDg3 ! ? 38 'ii'f4 (but not 3 8 d6
his attack going: 1:1h l ! and White will be mated)
3S �h4!! 38 . . . lLlf5 39 d6 and White's
You can only play deep posi­ passed pawn is dominant.
tional chess if you are alert to So Black must try 36 'ili'e3 c3 ! ?
tactical nuances. Without this 3 7 J.xc3 (3 7 'iWxc3 lLlf4, with the
move, all Black's fine strategical threat of . . . lDh3 +, is very danger-
build-up would have been ous) 37 . . . l:.xf2 38 'iWxf2 'ii'x g5+
wasted. The queen retreats one 39 �hl 'ili'xd5 . White remains a
square, leading to an almost im­ pawn up, but Black's queen and
perceptible change in the posi­ knight will co-operate splendidly
tion. Yet now there is a threat of to exploit White's ragged pawn
36 . . lDf4 followed by 37 . .. lDh3 +
. structure. For example, 40 'ij'fl
winning, since 3 7 i.xf4? i s met lDf4 4 1 b5? 'ii'b 3 ! 42 i..d2 lDd3
by 37 .. :Vifxf2+ and mates. Now and White's b-pawn is lost.
we see the drawback in White's White' s passed pawns hem in his
apparently solid defensive line bishop, but if he advances them
up: if any piece is dislodged, it they will fall victim to Black' s
16 Positional Sacrifices

control of the white squares. move like 35 . . . 'i'h4!! in his time

Therefore, after 4o . . . lLl f4, White pressure, he will blunder.
would have to avoid any weaken­ 38 'i'n 'i'xd2
ing pawn moves such as 4 1 b5? 39 'ii'f2 'ii'xgS+
and limit himself to manoeuvres 0-1
with the queen and bishop. A Evidently White was too short
draw would then be inevitable. of time to resign at move 37.
The game actually came to an Here however he manages it. A
abrupt end: great counter-attacking effort by
36 llxd2! Almasi.
37 lhd2 "'e1+ It is time to summarise. To
The reader may wonder why maximise our chess success, we
world-class players never blunder need an 'intimidating' style that
away pieces against them. Well, tests the calculation, imagination
if you can keep a top grandmaster and endurance of our opponent.
under intense pressure during the Only then will our friends marvel
middlegame, and then find a at our 'luck'.
2 Theoretical and Standard Sacrifi ces

A good combination is a se­ huge number of variations. And

quence of moves, perhaps involv­ every tactician makes his deci­
ing a sacrifice, that leads to an sions partly based on intuition.
improvement in the player's po­
sition. There are no dark alleys,
Ku pper-Tal
no obscure side variations left to
Zurich 1 959
chance. Everything is analysed
conclusively. The player sees the
combination.
A positional sacrifice, on the
other hand, tests not only the cal­
culating ability of the player. The
player' s experience and knowl­
edge is needed to fill in the 'gaps'
left by the human inability to
analyse everything.
Some players have a greater
capacity to calculate variations
than do others. So, for example, a
player such as Smyslov who has a
deep feel for positional chess may Tal played 16 lbxb2 and later
...

play an exchange sacrifice since wrote 'in sacrificing the knight, I

it ' looks right', while a Tal makes did not calculate variations. It
the same sacrifice only after an would be strange if after the sac­
intense analysis of variations. rifice, White were able to find a
Smyslov is making a positional defence against Black's over­
sacrifice; Tal is playing a combi­ whelming attack' (The Life and
nation. Games of Mikhail Tal, RHM
Of course, this is an over­ 1976). So not a single variation
simplification. Even the most was calculated by the master tac­
positional of players calculates a tician! How then does he know
18 Positional Sacrifices

the position is overwhelming? gous positions and had it down to

Was he just gambling? a science: pry open the KR [h- ]
The game continuation proved file, sac, sac . . . mate! '
him right: 1 7 'it>xb2 bxc3+ 1 8 1 6 . . . ':'c7 1 7 h4 1i'bs 1 8 h5
'it>xc3 0-0 1 9 ':'b l 'ifaS+ 20 <i;d3 lUc8 1 9 bxg6 bxg6 20 g4 as 2 1
':'ac8 ! 2 1 .f2 .i.a8 ! 22 1:.b3 e5 23 g5 liJh5 22 ':xhS ! (of a similar
gS exd4 24 liJxd4 (24 gxf6 11xc2 ! sacrifice against Gligoric, Fischer
2S 'it>xc2 'ii'x a2+ 26 ':'b2 ltc8+ quips 'I've played this sacrifice
wins) 24 . . . .i.xd4 and White re­ so often, I feel like applying for a
signed. 2S 'ifxd4 'ii' xe l or 2S patent! ') 22 . . . gxhS 23 g6 e5 24
�xd4 1:.xc2 is hopeless. gxf7+ �f8 2S .ll e3 dS 26 exdS
No, Tal was not gambling. He :lxf7 27 d6 ':f6 28 .i.gS 'ii'b7 29
did not need to calculate varia­ .llxf6 .llxf6 30 d7 ':d8 3 1 "i'd6+
tions since years of experience of and Black resigned.
making sacrifices had taught him To the experienced player,
that the knight offer must be sacrifices like Fischer's 22 .:txh5 !
good. And no doubt, he had seen against the Dragon are so well
many similar sacrifices by other known, even hackneyed, that it is
masters. The sacrifice was purely easy to forget they were only dis­
standard. Here is another exam­ covered in the 1 950s. Here is an­
ple: other common sacrifice: Black's
. . . ':'xc3 known by all experienced
Sicilian players.
Fischer-Larsen
Portoroz 1 958
Wel l s-Du ncan
Hastings Challengers 1 993/94

Fischer played 1 6 .ll b 3 and

remarks in My 60 Memorable
Games (Faber 1969) : ' I ' d won This game is just one of any
dozens of skittle games in analo- number of similar examples.
 Theoretical and Standard Sacrifices 19

Black played in 1 2 ! However, Peter Wells only

16 lbxe4 had one minute left to reach the
17 'iWe3 time control at move 40. As a
This is still known theory; the rule, a combination is something
experience of years of master that computers find easy, while a
games have taught us that Black positional sacrifice often baffles
doesn't get mated after the natu­ the machine. I'm sure a computer
ral 17 lbxe4, or rather that Black would have found the mating se­
gets in first: 17 . . . .txd4 1 8 .txf8 quence in one minute. But would
'jib6 ! ? threatening both it be able to play positional sacri­
19 . 'ifxb2 mate and 1 9 . . . .te3 .
. . fices such as 17 . . . lhc3?
17 ':'xc3 27 :xh7+! rl;xh7
Here it is. Black gives up the 28 'i'h3+ rJ;g7
exchange to weaken White's 29 'i'h6+ cj;n
queenside pawn structure and lay 30 lIn + rJ;e6
open his king to attack by Black's 31 .h3+ rJ;d5
queen. An intuitive sacrifice? 32 lbb4+ rl;c5
When it was first played in a 33 'iWe3+ rl;b5
similar position 40 or so years 34 rJ;b2?
ago, yes; nowadays it is known Here 34 a4+! rJ;xa4 35 lbxc6
hy everyone! bxc6 36 'ife4+ rJ;b5 37 'i'b4+
18 bxc3 lbf6 �a6 38 rJ;b2! and mate by 39
19 .txg7 rl; xg7 !:ta l , as pointed out by Black af­
20 l1h4 ter the game.
Future generations may estab­ 34 a6
lish whether this position is good 35 a4+ <ifi>a5
for White or Black, but in the 36 <ifi>a3 'i'c7
present dark ages the verdict is 37 :bl 'ifb6
'unclear' . The usual theoretical Black just survives and wins
moves are 20 lbe2, 20 ':'h2 or 20 with his extra piece.
'iWe6+ according to Chris Duncan, 38 •d2 'if'c5
so 20 .l:!h4 may be an attempt to 39 <ifi>a2 'i'c4+
solve the mystery. 40 <ifi>a3 'itb6!
20 :g8 41 'iff2+ 'ifc5
21 llJe2 .tc6 42 'fin ll g7
22 llJf4 e5 0-1
23 g5! lbe8
24 llJd3 f6 Such standard sacrifices are
25 f4 fxg5 discovered by the trial and error
26 fxg5 �e7 of master players. A player sees
White has built up a ferocious an interesting sacrifice and, after
attack. and in fact it is now mate doin.!! some analvsis over the
20 Positional Sacrifices

board, decides to risk it. The sac­ 9 �xe6 bxe6

rifice proves effective, and he 10 'i'a6 'i'd7
wins the game. Then a second More or less forced as 1O . c5
. .

player sees this game (perhaps in 1 1 'ii'c6+ �d7 12 i.g5 gives

print, or in the flesh, or from a White a powerful attack.
friend) and thinks the sacrifice 11 'ili'b7 lieS
looks like a good idea. He plays it 12 'fIxa7
at the next opportunity. And then Now White is a pawn up and
a third and fourth p layer become apparently has a safe position. In
interested . . . his book Bobby Fischer and his
I n this process, many unsound Predecessors (Bell 1 976), Euwe
ideas are of course eventually remarks at move 8 that 'it was not
discarded, but others stand the so difficult to see the loss of a
test of time and become part of pawn by force in two or three
every player's repertory. Hence moves, but Capablanca appar­
Chris Duncan can play 17 . :Xc3 . . ently did not imagine that such a
and know it must be quite good, thing could happen in the solid
whereas Alekhine could only Steinitz defence of the Ruy
'risk' . . Jhc3 after a great deal of Lopez. Capablanca's mistakes are
analytical work. After all, he just as clear as his good moves. '
never saw one . . . ':'xc3 Sicilian Neither Nimzowitsch nor Euwe
sacrifice in his lifetime! could comprehend that Capa­
The Benko Gambit 1 d4 �f6 2 blanca had deliberately sacrificed
c4 c5 3 d5 b5 !? 4 cxb5 a6 5 bxa6 the pawn. Euwe calls it an
is well established these days as a 'accident with a happy ending' .
respectable if somewhat double­ Capablanca himself saw things
edged opening system. And yet in somewhat differently: 'I believe
the 1 9 1 0s there was probably [Nirnzowitsch] has been unjustly
only one player in the world who criticised for losing the game . . .
understood the value of such a [the critics] have all suggested
sacrifice. moves here and there; but the
games of the great masters are
not played by single moves, but
Nimzowltsch-Capablanca
must be played by connected
St Petersburg 1 9 1 4
plans of attack and defence, and
these they have not given' (My
1 e4 eS 2 �f3 �e6 3 �c3 �f6 4 Chess Career, Macmillan 1 920).
�bS d6 5 d4 i.d7 6 �xe6 i.xe6 A comment that shows a deep
7 'il'd3 exd4 8 �xd4 g6 appreciation of chess strategy.
Has Black blundered a pawn? 12 i.g7
Nimzowitsch certainly thought so 13 0-0 �O
!lOci f'!!l?:l"rlv nlavl"ci Black is able to improve his
 Theoretical and Standard Sacrifices 21

position with a step-by-step plan. on the queenside and White

This could consist of the follow­ doesn't have a ghost of counter­
ing: play.
i) Place the rooks on the a- and 23 'i'e3 l:[b4
b-files. 24 .gS
ii) Manoeuvre the knight to c4 All White can do is respond to
via g4 and e5, or perhaps d7 and one move threats. There is no
e5. harmony among his pieces.
iii) Win one or more of 24 J.d4+
White's queenside pawns by the 25 �hl :abS
intense pressure of all Black's The pressure intensifies. Now
pieces. The bishop on g7 exerts White has no choice but to sacri­
enormous pressure on White's fice the exchange, since
position. 26 .i.xc3 is threatened and the
...

Meanwhile, what should White knight has no good moves.

do? He has no obvious plan, ex­ 26 :Xd4 'i'xd4
cept the vague notion that he 27 l%dl 'ifc4
should defend and win with his
• Now b2 is dropping and
extra pawn' . This shows one of White's game collapses. The fmal
the key features of a good posi­ moves were: 2S h4 l%xb2 29 'i'd2
tional sacrifice - it deprives the 'ifcs 30 l:[e1 'i'hS 31 :al 'i'xh4+
opponent of a good plan, while 32 �gl .hS 33 as :as 34 a6
the sacrificer still has means to .cS+ 25 �hl .c4 36 a7 .cS
strengthen his position. The game 0-1
continued:
14 'ifa6 lUeS This game, one of Capa­
15 'ifd3 'ife6 blanca's best, is not included in
16 f3 lDd7 Golombek's book Capablanca 's
17 .i.d2 lDes 100 Best Games. Like Euwe and
IS 'ife2 lDc4 Nimzowitsch, Golombek, it
19 .:tab!, :as seems, believed Capablanca had
Black's plan gradually unfolds simply blundered a pawn! Almost
while White flounders. any strong modern player would
20 a4 lDxd2! appreciate the value of Capa­
Removing the one piece that blanca ' s concept. Why then were
could challenge the supremacy of Euwe and Nimzowitsch, two of
Black' s dominant g7 bishop. the strongest players in the his­
21 'ii x d2 "'c4 tory of chess, blind to the merits
22 lIed 1 J:ebS of the sacrifice? The answer is
Black's pieces are now beauti­ that they had never seen such a
fully co-ordinated. Every piece concept before. They had no in­
contributes to the plan of attack ternal model against which to
22 Positional Sacrifices

judge the sacrifice. Therefore it �f1 1i'bl+ 7 �e2 ti'c2+ 8 :d2

was easy to believe that the pawn 'ii'e4+ and 9 . . g3 exchanges
.

offer was simply a blunder. But pawns, thereby breaking the

the question remains: if Euwe and blockade, as does 5 'iti?d2 .n 6
Nimzowitsch could not under­ 1:te2 g3) 5 . . . . b5+ 6 'iti?el 'ile5+ 7
stand the sacrifice, how could :e3 (7 �f1 'i'xg3! 8 fxg3+ �xg3
Capablanca conceive it? Well , 9 <ii;>g l <ii;>h3 10 �h l g3 and wins)
Capablanca was a genius! 7 . .xe3+! 8 fxe3 <itth3 and the
..

There are also many standard king shepherds home the passed
sacrificial ideas in the endgame. pawn long before White's own
We shall look at several interest­ pawn gets moving.
ing examples. A knowledge of such block­
ades can save a lot of heartache. I
remember Kupreichik, the Rus­
sian grandmaster, giving a simul­
taneous exhibition against the top
British juniors some years ago. In
one game his opponent had an
easily winning position, a piece
up for nothing, but unfortunately
he followed a well-known piece
of advice if you are material up:
exchange off at every opportu­
nity. One by one the pieces dis­
appeared, then Kupreichik
Black cannot win, since White's 'blundered' his queen . . . and fi­
rook oscillates between e3 and g3 nally reached a drawn position
where it is defended by the f2 very similar to the diagram
pawn. Meanwhile, the white king above. Here is another example:
cannot be driven from the g 1 or g2
square where it defends the f2
Klr.Georglev-Anand
pawn. Black's king cannot ap­
Las Palmas 1993
proach, so the game is drawn.
Note that this is White's ideal (see following diagram)
defensive set up. A slight change
could spell defeat. For example, Things look desperate for Black,
put White's king on el. Now if since he must lose a piece. But
White can get his king to g l , we remembering our blockade
have the book draw above. How­ above, 32 . . . :g2 ! ? suggests itself.
ever, 1 'iti?n 'i'h l + frustrates him. Then 33 'ii'xe3 :xe2 34 'ii'xe2 a6!
Then 2 �e2 'iti?h4 3 :g3 'iWc 1 4 (ruling out any disruptive a6 by
:�::\ 'ii'h 1! 5 %:tg3 (5 :d3 ti'b2+ 6 White) and White cannot win.
 Theoretical and Standard Sacrifices 23

leads to a curious material bal­

ance. There are no weaknesses in
Black's position and he can
hardly lose. Nei ther has he any
winning tries, so a draw would be
inevitable.
33 ':a4+
Black would have to resign
after 33 . . . :e4? or 33 . .. b6? 34
'i't7! .
34 r;.tb3 ':xa6
3S .ixa6

Even if White wins the e5 and

h7 pawns, Black could establish a
hlockade like the following:

3S lLldS!
If 35 . . . bxa6? 36 'i!Vxe3 and
Black wouldn ' t be able to set up a
blockade, e.g. 36 . . . �b7 37 'i!Vd3
(analysis diagram) r;.tb6 38 �c4 .l:tg8 39 -..e 3+ r;.tb7
40 'i'xe5 ':c8 41 �c5 :'c7 42
The black rook swings from b5 �d6 and wins, or 36 . . . kId8 37
to d5 and back again. When the 'i!Vxe5+ <it>c8 38 'i!Vc5 <t>b7 39
rook is on d5, the sacrifice "ile7+ <it>b8 40 'i'xa7 :d5 4 1
'ii'x d5+ cxd5 , �xd5 is only a 'ii'xa6+ �c7. With the black
draw. But in the game, Anand pawn on the third rank rather than
tried the second, the blockade fails,
32 l:tg4 e.g . 42 Wa7+ <t>c8 43 'ifxh7
Perhaps he was playing for a :b5+ 44 'itc4 �b8 (44 . . . :'xb2
win? loses the rook after 45 't!6g8+,
33 a6! etc .) 45 b4 .l:!.d5 46 f1f7
33 'iix e3 .l:ta4+ 34 r;.tb3 Iba5 (threateninJ! 47 1Ii'xd5 ) 46 . . . .l:tb5
24 Positional Sacrifices

47 'iWd7 and the blockade crum­ clearly drawn blockade position.

bles after 47 ... l:b6 48 �c5. Note 42 cxd5
that there is no blockade draw 43 'iWxe3 l:c8
with a rook's pawn: 44 'it'e6 l:c6
44 . .l:c5 would allow White to
.

weaken the blockade after 45

'iWe8+ �a7 46 'iWe3 b6 47 'i'e7+
�b8 48 �a2!? with the idea of
b2-b4 undennining the rook.
44 . l:c6 is simpler.
. .

45 'iWxd5 l:b6+
46 <lta3 l:b5
47 'W'd4 �b8
48 b3 �a8
49 �b2 �b8
50 �c3 �a8
51 �c4 �b8
(analysis diagram) 52 b4 <itc7?!
As Anand points out, 52 ... �a8
This is because the rook's is more accurate with a total
pawn only produces one safe draw.
square for the rook, not two. So 53 'iWf4+!
in the diagram position, Black A cunning move, hoping for
loses after 1 'i'd7! and the rook 53 . . 5bc8 34 'fid6 ! when Black
has no safe move: 1.. .l:h6 2 loses since he is in zugzwang,
'iWe8+ �b 7 3 'iVe4+ ':c6+ e.g. 54 . . . b6 55 "fie7 �b8 (or 56
(everything else loses the rook to "fia7 wins) 56 "fid7 <ita8 57 "fic7
a check, e.g. 4 . . 51tc7 5 'iff4+ lld6 and wins.
6 �c5) 4 �b5 a6+ 5 �a5 �c7 6 53 �b6!
'iixc6+! <itxc6 7 �xa6 and wins. But this is good enough to
Hence, Anand was right to play draw .

35 . . . �d5 !' We return to his game. 54 'i'b8 :h5

36 J.c4 e4 55 � b3 I:tb5
37 'iW17 :d8 56 �a4 l:td5
38 "ii'xh7 e3 and a draw was agreed.
39 "ii'h2+ �a8
40 'We5 a6
Pinter-Kasparov
41 'il"e4 lIf8
France 1 993
42 J.xd5
Georgiev realises there is no (see following diagram)
way to progress without this ex­ Kasparov is evidently on the
change. Rut now we reach a brink of defeat. But like all world
 Theoretical and Standard Sacrifices 25

champions, he has an intense will

to survive. Therefore it is not sur­
prising that he found some saving
resources in this desperate posi­
tion. Pinter began with

(analysis diagram)

However, White can play a

positional sacrifice 77 :'a2+ �f1
(keeping the white king away
74 g3 from gl) 78 lLlxf3! .i..xf3 79 h4
The idea is that after and White can begin to advance
74 . hxg3+ 75 �xg3 White can
. . his pawns. Black would have to
play 76 lLlf4 and then advance the play like a world champion to
h-pawn. The black king can be defend such a position.
driven off by :'e l + followed by ii) White captures the f-pawn:
:'a 1 when the rook helps force 76 lLlxf4+. Now 76 . �f2 ! 77
..

through the h-pawn. Kasparov lLlh5 i.e7 (77 ... hxg3+ 78 lLlxg3
found an excellent defence: .i..f4 79 :'f1+ and 80 :'e l + wins).
74 �e2!
7S :at f4! !
Attacking the g6 knight. This is
hy far Black's best chance though
White retains an excellent posi­
tion. We can run through a
checklist which illustrates posi­
tional draws of more or less cer­
tainty:
i) White captures the h-pawn
with his knight: 76 lLlxh4. Now
76 J..xh4? 77 gxh4 f3 78 :'a2+
. . .

and 79 �g3 wins so Black must

play 76 . . . f3 ! . (analysis diag ram)
26 Positional Sacrifices

A bizarre position where it is queening) 100 :f2 �g3 1 0 1 :c2

difficult to see how White can �b7 and White tried for another
escape from the bind of Black's 1 6 moves to win before giving it
pieces. For example, 78 :a7 �d6 up as a draw.
79 :n+ �f3 achieves nothing.
iii) White plays 76 gxh4 ! �h6 Thanks to knowledge of some
77 h5 . White' s best chance to endgame theory, the author man­
win. The knight is defended and aged to escape with a draw from
if he can neutralise the f-pawn the following unpleasant position:
then the h5 pawn will win the
game.
McDonald-Shovel
iv) The game continuation.
London Leag ue 1 994
76 :a2+ �e3
77 lLt3+ <itf2
78 gxf4 .th6
79 1:[a2+ cJit1
80 �a l + �f2
81 Ilgl �xg6
82 :'xg6 .txf4+
83 �hl .tg3!
This was one of the positional
draws for which Kasparov was
angling. White cannot win, since
he can never drive the black king
out of the 'square' of the h-pawn.
Therefore, a l:hg3 or :xh4 sacri­
fice will never win. Black played 47 . . J1h2 and was
The game continued: 84 'itg4 surprised when White answered
�f1 85 �g5 �f2 86 :f5 + �e3 48 :f8 :xa2 491hn l';ia3 . White
87 �g2 �e4 88 :n �d5 89 :'e7 has lost the b3 pawn, but the end­
�d6 90 :e2 �d5 9 1 �f3 �d6 92 game after 50 �d4 llxb3 5 1
�g4 Wd7 93 �f5 'itt d6 94 l:te8 l:H6+ 'it>d7 52 �c4 is a theoretical
�d7 95 :e6 �c7 96 �e4 �d7 97 draw.
<itd5 <itc7 98 lIe7+ �b6 99 %rn
.th2 ! (not 99 . .. �b5 1 00 l:Ib7+ The reader should aim to build
<ita6 1 0 1 :b 1 <ita 7 1 02 �e4 and up a large ' internal stock' of such
White puts his king on g5 then standard draws. Often they will
plays.l:.b4 and �xh4 ! winning the save him a great deal of analyti­
rook pawn when Black's king is cal work, and sometimes they
too far away to stop the h-pawn will save him the game.
3 Sacrifi ces to Create a Passed Pawn

Around the year 1 500, the rules there were 1 6 potential queens on
of chess underwent a dramatic the board! An extra pawn, or the
change. The minister, the most potential to win a pawn in the
pitiful piece on the board - a kind endgame, assumed decisive im­
of circumscribed bishop only able portance.
to limp one square diagonally at a Hence, positional play was
time - became la dame enragee, deepened, since the slightest of
the modem aggressive omnipo­ advantages could eventually yield
tent queen a win. The adage 'a pawn up in
Suddenly, the tempo of chess the endgame is quite enough to
speeded up: it was possible to be win' really means 'to checkmate
mated in two moves, when nor­ you need at least a rook, and this
mally the two opposing armies comes about through being a
did not come into contact for ten pawn up, exchanging off the
moves or so. All the established other pieces and pawns, and then
theory with its emphasis on pa­ queening the extra pawn' .
tient, long drawn out positional The worth of other pieces was
manoeuvring, had to be discarded also changed through empowering
in the face of lightning attacks by the queen and pawns. For exam­
the queen. ple, to lose a knight for three
And yet, paradoxically, the pawns would often be a serious
strategy of chess had in other loss in the old game. Now, on the
ways been enriched by this other hand, although a knight is
change of the queen's power. nominally worth three pawns, in
Previously, pawns could only be many situations, especially in the
promoted into the feeble minister, endgame, three healthy pawns can
a piece which had hardly any in­ outweigh a knight. The capacity of
fluence on the game. This meant the pawns to promote is the vital
that a pawn advantage or more in factor. Three pawns have the
an endgame wasn't of much sig­ 'stored' energy of three queens; a
nificance - and then suddenlv kni�ht is always a knight.
28 Positional Sacrifices

And similarly a pawn on the to a finish. He knows that under

seventh rank is but a pawn; a given circumstances, the kingside
pawn on the eighth rank is a pawns must become overwhelm­
queen! That is why games in ing. '
which one player has sacrificed a In other words, the sacrifice is
piece or even a rook for several not a risky venture, but a well­
passed pawns often lead to mind­ known technical device to drive
bending complications. If the de­ home White's advantage. White
fender succeeds in staving off the would not make the sacrifice if
pawn advance, he wins; if not, he there were any doubt about its
loses. value: after all, in the initial posi­
In this chapter, we begin with a tion he is a pawn up, with many
technical (or 'no risk' ) sacrifice different winning plans to choose
of the exchange to create passed from. 46 lhe6 ! is the quickest
pawns. Then, gradually more and and cleanest way of finishing off
more complex examples are in­ the opponent. The rest of the
troduced, where risk and imagi­ game is self-explanatory: the
nation are as essential as calcula­ pawns sweep all before them.
tion and knowledge. 46 jtxe6
47 jtxe6 l:Ub7
48 jtb3 l:.e8
Alekhine-Fl ohr
If 48 . l:b5, White simply ig­
. .

Nottingham 1936
nores the attack on the c-pawn
and continues as in the game.
49 h6 gxh6
50 g6!
An important moment. The
white pawns are very strong in
united formation, since they can
protect each other as they edge
forwards. 50 gxh6?? on the other
hand would be a terrible posi­
tional mistake, since the pawns
would become isolated and weak.
In fact, in this case, Black could
continue 50 . . :h7 and immedi­
.

In this position Alekhine sac­ ately win one of the pawns. When
rificed the exchange with 46 advancing passed pawns, you
lhe6 ! and comments in the tour­ should always try to keep them in
nament book: 'One of the combi­ a compact formation.
nations that an experienced 50
T'hY("r d0("S n"t 111"("1 t('\ ('nkl11�t("
 Sacrifices to Create a Passed Pawn 29

52 �e2! Davies, the player of White in

Not 52 �f7? which gives this game, is fully aware that if he
Black the chance to counter­ does nothing active, it will be the
sacrifice with 52 . . . l:.xf7! when he same story as in the Alekhine
has equal chances. The bishop game above: 'Under given cir­
and pawns are worth much more cumstances, the passed pawns
than a rook. must become overwhelming, etc. '
52 h5 So instead of trying to directly
53 :d6 l::le 7 stem the tide of the passed pawns,
54 f6 :e1+ he launches a counterattack:
55 'ii? d2 :0 28 :ae1 tOe5
56 n h4 28 . . . :xf5 ! looks a better way
57 l:.d7 1-0 to kill off White' s counterplay,
58 g7 is threatened and if e.g. 29 :e4 c3 30 ':'c4 (30 %1g 1
57 . . . l1g 1 58 :e7 followed by 59 c2) 3 0. . .tZk5 3 1 l:.xb4 tDd3 32
%1e8+ wins. :b7+ 'iitc 8 or 32 :c4 ':'xf4+ 33
Here is a more difficult modem :xf4 tDxf4 34 �xf4 c2 and wins.
example: 29 :e6!
White sacrifices the exchange
himself in order to tum his passed
Davies-Onischuk
pawns into a potent force. Unfor­
Budapest 1 993
tunately, it does not prove enough
to save the game, but it is still the
best fighting chance.
29 tDxe6?
Better is 29 .. Jbf5 ! ? 30 ':'xd6+
'iite 7, when both 3 1 :c6 tDd3 32
l:lxc4 tDe5+ 33 <t>e4 %1xf4+! 34
<t>xf4 tDxc4 and 3 1 :e 1 + �f7 32
%1c6 :xf4+ 33 'it>xf4 tDd3+ win.
30 fxe6+ �e7
31 <t>e4 b3
Continuing as in Alekhine­
Flohr with 3 1 . . .:'xf4+ is not
clear, e.g. 32 'it>xf4 c3 33 'it>e3 c2
B lack played: 34 �d2 b3 35 f4 ! ? �d4 36 \tc 1
26 .:texf5! and the pawns are stymied.
27 gxf5 �xb2 32 �e3 a6
Black now has two menacing 33 �d4 �xd4
connected passed pawns, while 34 �xd4 b2
White's own passed pawns are Now an interesting rook and
doubled and ineffectual. Nigel pawn endgame results. Black has
30 Positional Sacrifices

to play with great accuracy to 54 :xa6 �h 2 +

win. 55 Wf1 �h3+
35 :bl 0-1
If 35 'it;xc4 �xf2 and the dou­ The rook is lost after 56 'it;g l
ble threat of 36 . . l1xh2 and 'ii'g 3+ 57 'it;h l 'ii'el+ 58 'it;h2
.

36 . . . l:.c2+ followed by 37 . . . 1Ic 1 'iYe2+ or 56 'it;e l 'fie3+ followed

forces White to transpose into the by . . . 'i'd3+.
game continuation.
35 ltxf2 In the above games, the ex­
36 'it;xc4 :xh2 change was sacrificed with the
37 :0 h5 simple and direct aim of forcing
White has no answer to this through the pawns to the queen­
passed pawn. ing square. In the following ex­
38 :f7+ �e 8 ample, the threat of queening is
39 .l::(b7 h4 used as a diversionary tactic to
40 : b8 + c:t;e7 pin down the defender's pieces
41 'it;d3 h3 and make him unable to resist an
42 :b7+ �f6 attack on the other wing.
43 :f7+ c:t;g6
44 .:to l1 g2 !
I.Sokolov-Riemersma
45 e7 bl(i') +
Leeuwarden 1 993
46 kIxbl 1;f7
Material is now equal, but the
e7 pawn is easily neutralised by
the black king. Meanwhile, the
black h-pawn is very strong, and
if White tries to approach it with
his king, he will lose both his a­
pawn and his d-pawn.
47 'it;e3
If 47 ]:te l 'it;e8 and Black's
rook starts by gobbling the a-
pawn.
47 1%a2
48 'it;f4 1;xe7
49 .l:ib7+ 1;f6 Here White played:
50 iib6 l:tf2+! 21 :c6 .llxc6
Freeing the g5 square for the 22 dxc6 'fic7
king. White now has two strong and
51 'it;g3 h2 completely unassailable passed
52 Itxd6+ 1;g5 pawns. Black will never be able
53 <;i;>xf2 h t (iii') to arranl!e a . . Jhb5 counter-
 Sacrifices to Create a Passed Pawn 31

sacrifice or dislodge the white The first direct threat: 27 fxg6

bishop with . . . d6-d5 (assuming of hxg6 28 e5 .1xe5? 29 'ii'x g6+.
course that White remains vigi­ 26 .lte5
l ant) . Therefore, White's sacrifice 27 g3 :e8
is excellent in that he can hardly Black flounders about, unable
lose. to find a good plan. 27 .. :iia7 was
But how is White to win? The better, though Black remains
queenside is blocked and Black's passive.
bishop controls the al square, so 28 'iib3
White cannot hope to penetrate Now White plans 29 'iia 2 fol­
along the a-file. The passed lowed by i.a5 to force through
pawns, although strong, can be his queenside pawns.
firmly blocked on the dark 28 :f8
squares in front of them. How­ 29 �g2!
e ver, two factors are in White's There is no reason to hUrry.
favour: Black can do nothing, so why not
i) The black queen, although wait and see if he self-destructs?
doing an excellent job blocking 29 'ifb6
t he passed pawns, is shut away 30 J.h6!
from the centre and the kingside. Forcing the black bishop to a
The rook on b8 also cannot ven­ worse square before embarking
t ure too far from the b-file. on his queenside acti on.
ii) The recapture 22 dxc6 has 30 .ltg7
opened up the a2-g8 diagonal for 31 i.d2
White's bishop. The f7 square Now White is ready to carry
could become vulnerable. out his plan of 32 'iia2 and 33
So White decided he should .1a5 .
hegin an attack on the kingside. 31 gxf5
The potential threat from the 32 exf5 .lteS
queenside pawns ties down a To answer 33 'iia2 with
large part of Black's army and 33 . . . d5 ! when Black breaks free.
i mpedes his mobility. But now disaster strikes on the
23 f4 :a8 other side.
Black is afraid of White's at­ 33 J.h6 .ltg7
t ack and s o forces an exchange of 34 J.xg7 ! rt;xg7
rooks. The problem of course is 35 'iie 3 e6
t hat Black now has fewer pieces Sokolov gives the variation
to block the pawns and defend his 35 . . . 'iic7 3 6 'ii g 5+ �h8 37 'iih 6
kingside. \t>g8 3 8 f6! exf6 39 J.d3 l:te8 40
24 l':b l :al .l¥..xh7+ <ith8 41 b6! ! which dem­
25 lbal J.xal onstrates the two facets of
26 f5 White' s winning scheme: king-
32 Positional Sacrifices

side attack and queenside pawn haps combining with lLId2-c4 and
advances. it.d2-c3, putting intense pressure
36 "gS+ �h8 on e5.
37 .f6+ 'it>g8 S it.cs
38 fxe6 1-0 9 lLIf2 it.xf2+
Black's game collapses. 10 r.txf2
So Black has prevented lLId3,
We shall now look at some ex­ but at the cost of exchanging his
amples in which a piece is sacri­ long-range bishop for the knight.
ficed to create passed pawns. A White's bishop on c l now looks
piece is of course a much greater like the best minor piece on the
material investment than the ex­ board.
change, and so the compensation 10 lLIc6
must be proportionally more 11 it.e3 �e7
tangible and 'tactical ' . The fol­
lowing game is given in its en­
tirety since it demonstrates how a
sacrifice can flow logically out of
the strategical requirements of a
position.

Yusu p ov-C h ristiansen

Las Palmas 1 993

1 d4 d6
2 e4 lLIf6
3 f3 eS 12 lLIa3!
4 dxeS dxeS Another sideways deployment
S "xdS+ '1txdS of a knight (see move 8), and
6 �c4 �e6 another exclamation mark. The
7 �xe6 fxe6 idea is to play c3, depriving the
Black has blunted any aggres­ black knight on c6 of the central
sive intentions White may have d4 square, and then re-route the
had. He hopes that the e-pawns, knight from a3 to c5 via c2, el
although doubled and isolated, and d3. Such a slow manoeuvre is
will prove defensible since they perfectly feasible in view of the
cannot be frontally attacked. fixed and static nature of the po­
Yusupov, however, finds a way sition.
to begin a siege of the e5 pawn: 12 a6? !
S lLIh3 ! Black wants t o centralise his
With the idea of lLIf2-d3 , per- rooks without being bothered by
 Sacrifices to Create a Passed Pawn 33

the possibility of ttJbS . However, pawn before ttJcS. So White

this care less pawn move was ex­ switches his thoughts back to the
actly what White was hop i ng for. e-pawn.
12 .. J;lad8 was better, when if 1 3 20 ii.f2! b6
ttJbS a6 1 4 ttJxc7? :d7 I S �b6 White's plan is to play ii.g3
nc8 wins the knight . After and then b4, a4 , !':tb 1 and bS,
1 2 . . . kiad8 13 c3 ktd7 fo llowed by forc i ng away the c6 knight and
. . Jlhd8 Black would have had a winning the eS pawn. Here we
safe position. see why White retained one rook
13 e3 l:thd8 - a rook is necessary to support
14 �e2 h6 the undermining queenside pawn
Another superfluous pawn advance. So Black tries to set up
m o ve . 1 4 . . . l:Id7 was better. a pawn barricade but it further
15 ttJe2 :d7 weakens his queenside pawns.
16 l:[hdl :ad8 21 ii.g3 lLlf7
Imagine if Black had not 22 a4 as
wasted a move on 1 2 . . . a6, but had So that if 23 b4? axb4 24 cxb4
instead played to double his rooks ttJd4+ and the knight is well cen­
on the d-file immediately. In that tralised.
case, we would reach virtually 23 :tel! �f6
the game position with Black to 24 b4
move. Black could then exchange White offers to concede the d4
off both pairs of rooks with excel­ square to Black's knight since
lent drawing chances. White in after 24 . . .axb4 2S cxb4 ttJd4+ 26
fact wants to exchange one rook �e3 he has a strong potential
to rule out any Black activity passed pawn on the a-file, and the
down the d-file. He does not want eS pawn is still vulnerable.
to exchange off both rooks since Yusupov gives the following
he needs a rook to support his variation in In/ormator 58:
strategy. (This will become ap­ 26 . . .c6 27 as bxaS 28 bxaS l:ta7
parent in what follows.) There­ 29 ttJxeS ! ttJxeS 30 �xeS+ �xeS
fore, if Yusupov had wished to 3 1 l:lcS+ with a winning position.
play for advantage against correct 24 %:td8
play by Black, he would have Black decides to wait patiently.
been forced to leave the black However, White can now prepare
rooks unchallenged on the d-file. a breakthrough on the c-file.
A little lost time can be fatal in 25 b5 CiJe7
chess, even in ' quiet' positions. 26 .iof2!
17 lhd7+ lbd7 A little care is needed. 26 c4
18 tDel! .!Des cS ! 27 bxc6 ttJxc6 (Yusupov)
19 lLld3 lLld6 would leave the black knight
Managing to bolster the b- eyeing d4. But now of course,
34 Positional Sacrifices

26 . . . c5 ? would lose the b-pawn. to the laws of strategy. And fur­

26 l:b8 thermore, the sacrifice is not a
So Black defends the b-pawn. luxury which can be played or
27 c4 c5 declined according to the player's
The only way to stop White style or state of mind (i.e.
breaking through with c5, cxb6 whether he is feeling brave or
and lIc7. not). It must be played. In fact, as
Yusupov points out, 28 �xc5 !
was even stronger. After
28 . . . bxc5 29 l£lxc5 l:a8 30 l£ld7+
followed by 3 1 c5, etc., the
pawns sweep through. The reason
why 28 l£lxc5 is inferior will soon
become clear.
28 bxc5
29 .i.xc5 l£ld8
30 �d6 l:tb7
31 l:dl
The first sign that it is not easy
for White. The 'obvious' 3 1 c5
If now 28 bxc6 then 28 . . . l£lxc6 allows 3 1 . . .l£lec6 ! 32 bxc6 (in
29 :b l l£ld6 ! (attacking c4) is fact, 32 <iti>d3 is better, in reply to
satisfactory for Black. And if the which the knight stands its
position remains blocked, then ground on c6) 32 . . . l:tb2+ 33 <iti>f1
Black can claim that his knight is l£lxc6. White is a pawn up, but
better than White' s impeded the black knight is excellently
bishop. Instead White played: placed on c6, blocking the passed
28 l£lxc5 pawn. And we see the limitations
It would be strange if White's of a bishop well illustrated: it
well thought out and well exe­ only controls squares of one col­
cuted plan suddenly led to a our, and the view from d6 is not
strategical 'dead-end' and he had very good - it is virtually incar­
to think about equalising the cerated by its own pawn on c5
game. Chess is normally (though and Black's e5 pawn and knight.
not always) a logical game. Imagine if the bishop on d6 were
Therefore, a sacrifice on c5, giv­ a knight. It could easily escape
ing up a piece for two pawns, is from its prison, jumping over
neither speculation nor a ' lucky other pieces if necessary (whether
find' . There is nothing fortuitous friend or foe) . The c6 square
about discovering a winning sac­ would not be a 'no-go area' ei­
rifice in a position which has ther, since the knight could con­
been carefully nurtured according suire with the rook to oust its op-
 Sacrifices to Create a Passed Pawn 35

posite number. 41 �c4+

31 lDdc6 42 <it>c2 l:b8
Black still makes the sacrifice, 43 c6 �e8
hut in far less favourable cir­ 45 l:c3 lDb6
cumstances - the white bishop is 46 as l:xa5
not blocked in by his c5 pawn, 46 . . . lLlcs 47 l:c5 is hopeless.
lind the white rook is more ac­ 47 c8('ii') 1-0
t i vely placed. However, the sac­ The strategical course of the
rifice was probably necessary game above was fairly clear. In
sooner or later. the following example, difficult
32 bxc6 lDxc6 tactical variations obscure the
33 .ic5 logical flow of the game as White
Keeping the knight out of d4 battles to prove the correctness of
lind preparing to penetrate with his sacrifice.
his rook.
33 l:b2+
Serper-Ni kolaidis
34 l:d2 l:b3
St Petersburg 1 993
35 l:d6 lDe7?
According to Yusupov, the last
chance for Black was 35 . . . lDd4+ !
36 ..txd4 exd4 37 Ihd4 :h4.
Black wins the a-pawn and is left
' only ' a pawn down in a rook and
pawn endgame. This would have
heen the way for Black to punish
White for his wrong choice of
sacrifice at move 2S .
36 ..tb6 lDg6
36 . . . � cS fails to 37 ..t d8+ and
38 llc6. Black soon loses his a­
pawn and his 'counter-attack'
proves fruitless against the passed Black should play 1 5 . . . �f4
a- and c-pawns. followed by 1 6 . . . 0-0 with a rea­
37 ..txa5 �f4+ sonable game. Instead he tried:
38 �d2 �xg2 15 lDfS
3 S . . . l:a3 39 ..t d S + and 40 a5 with the laudable intention of
wins. manoeuvring his knight to the
39 cS lDe3 excellent d4 square. He had mis­
40 ..td8+ r:3;f7 takenly assumed that because the
41 ltd3 position is semi-closed, White
A voiding all tricks. Black cannot exploit the time he wastes
could already resign. in retreating the knight. Serper,
36 Positional Sacrifices

on the other hand, is alert to his 18 f5!

chances of striking a quick blow A valiant attempt to keep the
before Black is able to castle. He white knight out of e4 (the threat
begins with: was lDe4-d6+). The black rook on
16 a4! b4 h8 may one day be able to swing
This provokes the sacrifice, but to the centre of the board via h7.
1 6 . . . bxa4 1 7 :'xa4 intending 1 8 Also, 1 9 . . .f4, winning the bishop
':fa 1 would be positional suicide. on e3, is introduced as a threat.
17 lDdS! cxdS The obvious defensive idea
Again forced, since otherwise would be to move the knight on
he loses the b-pawn. It is better to f8 and castle. However, all knight
face a huge attack than be a pawn moves fail:
down for nothing - any sacrifice i) 18 . . lL'ld7 19 lL'le4 0-0 20
.

carries an element of risk for the llac 1 and White is ready to push
sacrificer. Black off the board with c5-c6
18 exdS and d5-d6.
ii) 1 8 . . .lDe6 1 9 dxe6 �xe6 20
'ii'xb4 0-0. Black's king is safe
but he is a pawn down.
iii) 1 8 . . .lDf4 19 lDe4 (better
than 1 9 �xf4 exf4 20 lDe4 �e5 !
intending . . . lL'lh7 and . . . 0-0)
19 . . �e2+ 20 'if'xe2 followed by
.

lDd6+ with a strong attack.

19 d6 .c6
Serper points out in Die
Schachwoche that 1 9 . . . ...d7 20
c6! "'xc6 2 1 %lfc 1 "'d7 22 %lc7
wins. The black queen is curi­
White's compensation for the ously short of squares and will be
piece is obvious. Even if Black's lost to :'e7+.
pieces were on good squares and (see/ollowing diagram)
his king safe, it would be difficult Now it seems as if the pawns
to prevent the united passed are blockaded, but Serper has
pawns from rampaging down the another surprise waiting:
board. But as it is, with Black's 20 �bS! !
rook sitting uselessly on the Once again a sacrifice to con­
wings, his minor pieces disorgan­ vert a closed position into an
ised, and his king stuck in the open one. Black is never allowed
centre, only one verdict is possi­ the one tempo he needs to play
ble: White has a winning advan­ . . . lDe6 or . . . lDd7 and . . . 0-0.
t�p"c . 20 axbS
 Sacrifices to Create a Passed Pawn 37

'i'xc8 25 'i'd5 stopping castling

and threatening 1:.a8 or c5-c6; for
example 25 . . . tLlhf6 (25 . . . tLldf6 26
'ii'x e5+) 26 'iVe6+ �f8 27 tLld3
(threatening tLlxe5) 27 . . . e4 28
tLlf4 'iVe8 29 1:.a8 ! 'ii'x a8 30 tLlxg6
mate. A better defensive try is
23 . . . .i.f6 intending to answer 24
l:ta l -a7 with . . . l:.h7 ! activating his
rook. However, 24 'iVxb4 :h7 25
'ua l -a6 ! looks decisive.
24 :'la7 tLld7
24 .. .fxe3 25 'iVd5 ! ! is devastat-
21 axb5 'iVxb5 ing.
Or 2 1 . . .'iib7 22 c6 'iVb8 23 b6 25 ':'xcS+!
(a very aesthetic array of pawns) Under no circumstances must
23 . tLlf6 24 'i'xb4 tLld5 25 'fib5
. . Black be allowed to castle.
tLlxe3 (25 . . Jha l 26 c7+ ! ) 26 25 'iVxc8
�xa8 'ii'x a8 27 b7 'ifb8 28 c7+. A 26 'ii'd5 fxe3
variation which demonstrates the 27 'it'e6+
enormous po�er of the passed Not 27 tLle4 tLlf4 ! .
pawns. 27 �f8
22 l:ba8 'ii'c 6 28 :'xd7 exf2+
23 'ufa1 29 �f1
Bad is 29 �xf2? 'it'xc5+.

23 f4
Black decides he may as well Despite Black's two extra
be hanged for a sheep as for a pieces, he has no good defence.
lamb. The attempt to develop Serper gives the variation
with 23 . . tLld7 fails to 24 llxc8+!
. 29 �!!3+
. . . 30 hx!!3 'ifxd7 31
3S Positional Sacrifices

"xd7 hxg3 (threatening . . J:th l+, 38 "d6+ ciitt7

queening) 32 "e7+ <iii' g S 33 39 .. d5+ ciitf6
'iWeS+ i.f8 34 "xg6+ followed 40 "d6+ ciitt7
by 'iWxg3, and White wins. We A little repetition in time pres­
can add another interesting varia­ sure. Now White plays the win­
tion: 29 . . ...a6+ 30 <iii'x f2 "e2+! ? ning move.
3 1 �xe2 ltJf4+ 3 2 � d l ltJxe6 33 41 'ifxe7+ � xe7
c6 and White wins after 33 ... ciit e S 42 c8('i')
34 1la7 i.f6 35 l1aS+ followed by The c- and d-pawns have had a
36 :xh8 and 37 c7, or 36 c7 im­ glorious career. One cost Black
mediately, as appropriate. Or if his queen, the other became (and
33 . . . cifa>gS 34 l1e7 ciith7 35 l1xe6 stayed) a queen.
and wins - the passed pawns 42 i.h6
march through. The last variation 43 'i'cS+ �e8
demonstrates the enonnous 44 "bS+ �d8
power of the passed pawns. 45 "b6+ �d7
29 "e8 46 "xg6 e2+
30 1:tt7+ ! Hoping White will capture the
Yet another sacrifice to clear wrong pawn and lose his queen to
the way for the pawns. a knight fork.
30 "xt7 47 �xf2! i.e3+
31 "c8+ "e8 48 �e1 1-0
32 d7! The board was not showered
Black now has the huge mate­ with gold coins after this fantastic
rial advantage of a rook, bishop display, but Serper did receive
and knight for a pawn - but what 250,000 roubles for the Best
a pawn! Black must now give up Game prize.
his queen.
32 �t7 Our final example is a compli­
33 dxe8('i')+ l1xe8 cated struggle in which both sides
34 "b7+ :e7 acquire passed pawns.
35 c6! e4
Although Black still has a
Teske-P . Szilagyi
nominal material advantage, his
B udapest 1 99 1
pieces are too disorganised to
stop the c-pawn. (see diagram overleaf)
36 c7 e3
A last desperate fling by Black. White can claim a very small ad­
Perhaps White will play 37 cS(") vantage in this arid-looking posi­
and allow 37 . . . e2+ 3S �xf2 tion: his rook controls the only
e l (iV) mate? open file. However, Black has
�7 �d�-!- �f6 only to play . . . i.b7 and . . . l1cS,
 Sacrifices to Create a Passed Pawn 39

exchanging rooks, and any advan­ obtain some passed pawns of his
tage will completely vanish. Evi­ own.
dently Teske did not like the look 24 Ita7 'ilxa7
of 22 �xc8+ Itxc8 23 "ii'x a6 l::c 2, 25 bxc4 .ta6
when 24 .te l (24 lDd3? ':'xe2) 26 lDxd5!
24 . . . ,Uxa2 or 24 . . . 'ii'c7! ? is un­ This should not save the game
clear, so he played: but it is the best try.
26 exd5
27 .txd5

22 l:lc6
expecting Black to move to de­
fend his knight. Instead, Black Now things do not seem so bad
played the highly ingenious for White. He has connected
22 "ifd7! ! passed pawns in the centre and an
letting White carry out his active bishop. As yet, Black has
threat. Actually White should no passed pawns of his own, and
now retreat his rook, with an the rook on b8 has no open file.
equal position. However, this However, if we look more closely
would be a psychologically diffi­ at the position, we will see that
cult decision to make, so White White' s pieces are in fact on vul­
accepts the challenge. nerable squares. This is almost
23 llxa6? lDc4! entirely due to the pressure of
Now after 24 bxc4 .txa6 fol­ Black's bishop on a6 which ter­
l owed by 25 . . . dxc4 Black's rorises the c4 pawn. If White
passed pawns would steam-roller could evacuate his queen from d3
through the queenside. For ex­ and achieve the advance c5 with­
ample, 24 bxc4 .txa6 25 'ii'b 1 out losing his knight on e2 or his
dxc4 26 .te4 �b5 27 d5 c3 28 bishop on d5, or indeed the pawn
.tc 1 exd5 29 .txd5 a4 followed on c4 itself, then he would have a
by . . b4-b3 and wins. So White
. safe game. But this is not easy to
prepares a counter-sacrifice to achieve, especially if Black plays
40 Positional Sacrifices

27 . . . :c8 ! . This threatens play 3 3 i.g2 and 34 c5 . White

28 . . ...c7 winning the c-pawn. has wasted three moves preparing
The only reply is 28 "c2 but then to bring the knight into an attack­
28 . . . i.xc4 29 i.xc4 "c7 30 ing position on the kingside when
i.xf7 + 'Yt;xf7 3 1 ..e4 (3 1 'ii'b 3+ the correct strategy was the ad­
'if c4) 3 1 . . . �g8 with the idea of vance of his passed pawns.
. . :ifc2 100ks very strong. 33 'ifb4!
Alternatively, Black could play 34 i.xt7+
27 . . ...d7 threatening 2 8 . . ...xd5. 34 "f1 was the last chance.
Now 28 "b3 a4! or 28 "e4 :c8 34 �f8
or finally 28 �f4 g5 ! are all bad 35 'I'd
for White. However, 28 e4 ! holds This is White's idea. Unfortu­
on, though Black is better, e.g. nately he has underestimated the
after 28 . . . 'ifxh3 ! ? strength of Black's passed pawn.
Black played an inferior move 35 i.xe5
in the game: 36 i.a3
27 "c7?!
28 "c2
Now 28 . . . :c8 can be answered
by 29 c5 when White has
achieved his central pawn ad­
vance.
28 "d7
29 �f4?
White does not take advantage
of Black's hesitancy. He should
retreat his bishop to g2 and then
play c4-c5 , perhaps followed by
d4-d5 and c5-c6. This would ac­
tivate his passed pawns. 36 i. d6?
29 a4 Here 36 . . . �xf7 wins (but not
30 �d3 b3 36 . . . b2? 37 'ii'xb2 ! ), e.g. 37 i.xb4
31 axb3 axb3 !'txb4 38 'ii'b2 (38 dxe5 b2 39
32 "c3 "d6! 'ii'b l i.xc4 40 �h2 lIb7 ! fol­
To answer 33 �c5 with lowed by . . . i.d3 wins, or 38 "a3
33 . . ...xc5 ! winning. 32 . . . 'ifxh3 :b8 ! 39 'ii'xa6 b2 wins) 38 . . . i.c7
33 �c5 i.c8 34 �xb3, on the 39 c5 (39 "a3 IIb6 40 c5 b2)
other hand, would be unclear. 39 . . . i.b7 followed by . . . i.d5 and
33 �5 . . J!a5-a2, etc. forcing the pawns
The culmination of White's through. Evidently Black became
mistaken plan which began with confused in time pressure.
':" C) �H Tf " · ;1 S stHl n f't too ' �t� t(' 37 ..t 1Ch4 ltxb4
 Sacrifices to Create a Passed Pawn 41

38 ..t d5 b2 .i c6 mates, or if 44 .. . �g7 45

39 'iib l ..tc7 '1'17+ leads to a mate similar to
Black plans to put his bishop that given at move 40. Since
on c3 and then play . . J�a4 and White is threatening 44 'ii'f6+
. . J:ta l ousting the white queen. �e8 45 .llc6 mate, Black has no
White's passed pawns do not choice but to return a piece, but
seem dangerous at present: if this is hopeless as well. No doubt
White ever plays c4-c5, then he was feeling somewhat ag­
. . . �d3 ! will be the instant riposte. grieved, and with good reason.
So is Black still winning after all?
Unfortunately for Black, another
factor has become important:
king safety. The white bishop on
d5 and the queen would be a
deadly force against Black's ex­
posed king; assuming of course
the queen can free itself of block­
ading duties . . .
40 �g2! ..ta5
Carrying on with his plan. An
attempt to flee with the king fails,
e.g. 40 . . . �g7 4 1 'fie4 ! b l ('I') 42
'i'e7+ �h6 43 'fif8+ �h5 44 43 'fif5
..tf3+ � g5 45 h4 mate. 44 e4 'ii'g5
41 c5! .td3 45 'iVxa5 l:tb2
Black had no choice, since 4 1 46 'i'a8+ �e7
c5 opened not only the a6-fl di­ 47 'i'a7+ �d8
agonal but also the a2-g8 diago­ 48 'iVt7 1-0
nal: 42 ..ta2 ! followed by 'fie4 Even if he avoids being mated,
was threatened. a rook is no match for a bishop
42 'it'xd3 bl('i') and four passed pawns.
43 'ii'a6!
Despite his extra rook, Black is An exciting and instructive
lost. 43 .td8 can be answered by
... game, despite its flaws. It illus­
44 'ilie6 ! when 44 . . :b7 45 'l'g8+
. trates very well that when there
We7 46 'it'f7 is mate, or 44. . . 'i'f5 are passed pawns, the value of
45 'ifg8 + <J;; e7 46 \'Ig7+ �e8 47 pieces can fluctuate enormously.
4 Sac rifi ces to Destroy the Opponent's
Ce ntre

One day, scientists may discover sacrifices the exchange on his

a manageable formula to find the opponents K3 (e6 or e3) for posi­
'best' move in any chess position. tional compensation.
If they do, then part of the equa­
tion will undoubtedly address the
Hulak-Gabrlel
question of central control. Many
Slovenia 1 993
opening and middlegame moves
appear meaningless or bizarre
until one realises that their aim is
to seize control of the centre or
prevent the opponent from doing
so. Control of the centre gives the
pieces extra mobility and opens
up all sorts of possibilities of at­
tack.
It is no wonder, therefore, that
the sacrifice to destroy or weaken
the opponents centre is one of the
most common themes in posi­
tional chess. In fact, such sacri­
fices are so frequent that they At first glance, Black appears
should be in every player's ar­ to have a safe and solid position.
moury. However, White completely dis­
In a single chapter it is not rupted Black's position with the
possible to classify and illustrate sacrifice 28 ':'xe6 ! fxe6 29 �f4! .
every example of this sacrificial There is no answer to the threat
motif, so we shall restrict our­ of 30 �xe6. Black tried 29 . . . b5
selves to looking at some specific but lost quickly after 30 cxb6
instances. First, we will examine 'ii'xc4 3 1 i.xc4 i.h6 32 b7 :ab8
snmt" (,�SI"C:; wht"Tf" Whitf' or Black 33 �xe6 i.xd2 24 �xd8+ <it>h8
 Sacrifices to Destroy the Opponent 's Centre 43

35 liJti+ fj;g7 36 l:d l .tc3 37 were becoming menacing. As is

liJd6 liJf5 38 liJxf5+ gxf5 39 well known, two bishops are of­
�xa6 and Black resigned. ten stronger than a bishop and
A fairly simple example: Black knight, especially when they are
was defenceless after the sacrifice on aligned diagonals and the
since he had no answer to the centre is open (which was the
tactical threats. We now consider case before White' s sacrifice
a more heavyweight positional here).
battle, where chances are more iii) The disappearance of
evenly balanced. Black's bishop pair and the dis­
ruption of his centre has taken a
lot of the dynamism out of
Raecky-Krasen kov
Black's position. It is hard to
Rostov on Don 1 993 imagine Black carrying out a suc­
cessful kingside attack, for in­
stance. Therefore, White' s static
or long-term advantage is now of
much more value, since Black
has little active counterplay.
However, we should not over­
estimate White's position. He
certainly has compensation for
the exchange, but we cannot yet
speak of a clear advantage, let
alone a win. White merely has a
safe game and the initiative. The
player who shows the best quali­
White played: ties in the resulting struggle will
20 l:txe6 fxe6 win.
Assuming that the e6 pawn is 21 h4
doomed, White has sacrificed a White is in no hurry to capture
rook for a bishop and pawn. His the e-pawn. First, he makes an
compensation: escape square for his king, should
i) Black's pawn centre is shat­ he ever be checked on the back
tered, and his kingside weakened. rank (which happens a couple of
Black could have serious prob­ moves later), and introduces the
lems on the a2-g8 diagonal from idea of h4-h5, further weakening
an attack by White's queen and Black's kingside.
knight. (This actually occurs in 21 �h8
the game.) 22 lLle4 l:f8
ii) White has eliminated Black must play cautiously. If
Black's pair of bishops which 22 ... :tbd8 23 'ihe6 'ii' xa4 24
44 Positional Sacrifices

lLlg5 threatens a smothered mate 24 itxd8 'i'xd8

after 25 lLlf7+ �g8 26 lLlh6+ 25 'i'xe6 'iWd1+
�h8 27 'iWg8+ ! and 28 lLlf7 mate. Black has achieved a great
deal. His queen is reactivated and
he has exchanged a pair of rooks.
As a general rule, if the opponent
has sacrificed or lost the ex­
change, the defender should aim
to exchange off the other rook.
26 <it;h2 'ilfxa4
27 'WIxe7
This is good enough for equal­
ity. The attempt to play for ad­
vantage with 27 c6 rebounds after
27 . . . bxc6 28 bxc6 'ilfc2 29 'it'd7
a5 30 c7 a4 (not 30 . . . .i.e5 3 1
23 lId1 'i'xe7 winning) 3 1 .i.h3 a3 (but
This leads to a favourable ex­ not 3 1 . . :iVxe4 32 c8('WI) l:.xc8 33
change of rooks for Black, whose 'iWxcs mates) and the black pawn
queen is then poised to enter into runs through to queen.
active play with a check on 27 'ii'xb5
White's back rank. In his notes in 28 lLld6 'i'b3
Informator 58, Raecky says he Of course, taking the c-pawn
was afraid to play 23 .i.h3 be­ loses the queen. 28 . . . 'iWb2 29 f4
cause of 23 . . . i.d4 with the threat 'i'f6 30 'i'xb7 is unclear, says
of 24 . . . .i.xf2+ 25 lLlxf2 l:.xf2+ 26 Raecky.
�xf2 'ii' d 2+ winning the rook on 29 f4 as
c l . However, 23 . . . .i.d4 could be Now Black also has a danger­
met by 24 �g2! when 24 . . . .i.xf2? ous passed pawn. A race devel-
fails to 25 'iib 2+ followed by 26 ops.
lLlxf2 winning. Black could an­ 30 .i.xb7 a4
swer 24 �g2 with 24 . . . e5 but 31 c6 a3
then after 25 l:.d 1 White keeps a 32 c7 'ifc2+!
slight advantage. He can prepare Not 32 . . . a2 33 c8('i') ':xc8 34
to queen a pawn on the queenside .i.xc8 threatening 35 'ii'e 8+ 'i'g8
(c5-c6) or attack on the kingside 36 lLlf7 mate.
(h4-h5). The difference between 33 'ri'h3 a2
this variation and the game con­ 34 c8(1i') 'ifxc8+!
tinuation is that Black's queen Well played. 34 . . . l:.xc8 35
remains shut out of the game on .i.xc8 h5 (36 'ir'eS+ was threat­
a5 . ened) 36 J.. e 6 leads to a quick
23 l:tbd8 mate, e.g. 36 . . . a l ('if) 37 'i'e8+
 Sacrifices to Destroy the Opponent 's Centre 45

c;t>h7 38 'i'g8+ �h6 39 1bf7 very strong. This threatens

mate. 39 . . . h5+ followed by . . ..xg3+
35 .i.xc8 al('ii) and mate. After 39 �e5+ c;t>hS 40
36 �rT+? 'iVa3 (there is nothing else)
After such an exciting struggle, 40 . . . h5+ 41 'iit g 5 �h7
White, no doubt in time pressure, (threatening 42 . . . i.h6 or 42 . . . .i.f6
carries on playing for a win when mate) 42 �xg6 :f6, White has to
it was time to settle for equality . play 43 f5 lhe6 44 fxe6 'iVd5+
Here he should have played 36 when Black wins the knight, with
.i.b7 when 36 ... 'i'a7 (pinning the excellent wlnnmg chances.
bishop) 37 �f7+ ..t>g8 38 �h6+! Similar is 43 i.f5 ':xf5+ 44 �xf5
�h8 (not 38 . . . i.xh6? 39 .i.d5+ 'iic 2+, again winning the knight.
winning the queen) 39 �f7+ is We may add another variation to
perpetual check. Krasenkov's analysis: 43 'ii' d3
36 � g8 'ii'x g3+! ! 44 .xg3 ':xg6+ and
37 .te6 'i'hl + wins. Interestingly, Raecky did
Driving the king out. Of not see the strength of 3S . . ..g2
course, if White had played 36 even in his home analysis. Ac­
.i.b7 this would have been im­ cording to his notes given in In­
possible. formator, White was always at
38 �g4 least equal throughout the course
of the game.
39 c;t>g5 �h7
40 g4 'i'gl
Raecky mentions 40 . . . 'i'd l , a
nasty move which threatens
4 1 . . Jhf7 42 'i'xf7 "'dS mating,
or if 42 i.xf7 'i'xg4 mate. White
had to play 4 1 �e5 when
4 1 . . :ifdS (or 4 1 . . .'.d6 ! forcing
the same reply) 42 �xg6 'i'xe7
43 �xe7 i.f6+ 44 �xh5 i.xe7
45 f5 is an interesting material
balance - rook and bishop against
38 h5+? bishop and three passed pawns.
Here it is Black's turn to go After 45 . . . .:a8 46 g5 White will
wrong. We should remember that sooner or later play f5-f6, when
chess players, even Russian Black will have to play . . . .i.xf6
grandmasters, are not automatons with a drawn endgame (otherwise
but flesh and blood, prone to the passed pawns would be too
nervousness and error. Krasenkov strong) .
Doints out that 3R 'iV1[2 1 was
. . 41 �t"5
46 Positional Sacrifices

"'fl + 44 <iite4 1IIg2+ 45 �d3

'ii'f1 + 46 �c2 'i'e2+ 47 �b3
'iib 5+ and White's king cannot
escape perpetual check. Perhaps
Black was still looking for the
elusive win.
43 AdS! hxg4
Black's last chance, threaten­
ing 44 . . :l'f5 mate. Hopeless is
43 . . :l'b5 44 f5 ! 'i'b l (44 . . .'i'xd5?
45 ltJf6+) 45 ltJf6+ �h8 46
�xg6, or 43 . . . ':'c8 44 ltJf6+ �h8
45 ltJe8.
41 'ii b l 44 ltJf6+ �h8
When Krasenkov played 45 J.e4 ! 1-0
40 . . . 'ifg l , he may have planned The killer move. Now Raecky
4 1 . . .'ifc5 here. The queen is taboo gives 45 .. :ifa l 46 i.xa8 'ii'x a8 47
because of 42 . . . i.f6 or 42 . . . i.h6 <it>xg6 'iff8 48 'ifxfS i.xfS 49
mate, and 42 ""7 'ifb6 ! 43 'ifd7 ltJxg4 and wins. It is ironic that
'ifxe6 ! 44 'ifxe6 i.h6 mate does the white king, which Black had
not help. But White has the devil­ been hunting down, is trans­
ish move 42 i.g8+! ! and after formed into a strong attacking
42 . . . 'iitx g8 or 42 . . . :xg8, White piece. Despite the mistakes, this
can safely capture Black' s queen was an absorbing struggle.
and there is no mate. If it were In the game above, although the
not for the 'flukey' 42 J.g8+, all initial sacrifice was based mainly
Black's previous moves would be on an assesm s ent of positional
justified, and Krasenkov would factors, there were some difficult
have scored a fine victory. We tactical variations to calculate. In
may talk about positional chess, the following struggle, Tiviakov
but unexpected tactical strokes playing Black makes the mirror
are always lurking there to upset image of White's sacrifice on e6.
the logical flow of the game. In­ This time, however, the battle is
credibly, instead of 4 1 . . ..,, 1 ,. decided by verbal 'stream of con­
Black should force a draw with sciousness' reasoning rather than a
4 1 . . .:xf4 ! 42 �xf4 'iid4+ fol­ mass of variations.
lowed by capturing the knight.
42 ltJd7! lh8?
Razuvaev-Tivlakov
Evidently the struggle has
Rostov on Don 1 993
proved too much for Black. As
Raecky points out, he could still
draw with 42 . . . :1xf4 ! 43 <iit xf4 Evidently Rostov on Don is a
 Sacrifices to Destroy the Opponent 's Centre 47

happy breeding ground for posi­ may try c4-c5 , giving up a pawn
tional exchange sacrifices. Here in order to open the b-file for his
Tiviakov played: rooks (or the d-file if Black an­
swers c4-c5 with . . . dxc5 and
White then plays d5-d6) . So
Black will be vigilant: he will
make sure that c4-c5 is either
prevented or not good for White.
iv) White has no good plan,
since his pawns cannot co­
operate in any aggressive action,
and every good plan requires the
use of pawns at some stage. This
last point may not be obvious but
will become clear as the game
progresses.
15 :xe3! 17 'i'd4 lbd7
16 fxe3 d6 18 IU4
What has Black gained for his White elects to act aggres­
sacrifice? sively on the kingside. Black
i) White's e-pawns are in the must be careful, e.g. 1 8 . . . lbc5
worst fonnation possible - dou­ would give White the chance to
bled and for all intents isolated play 1 9 :aft f6 20 .te4 with the
(since the d5 pawn won't ever idea of :h4 (Tiviakov) . Black
help their defence). Furthennore, would then have to face an attack
they stand on an open file. The e3 or play the positionally disagree­
pawn has the choice of being able 20 . . . lbxe4.
weak on e3 or e4. 18 'i'e7
ii) The black knight gains the 19 :an lbe5
beautiful central square e5 where Not 19 :e8 which gives
. . .

it is completely unassailable. White the chance to make a fa­

(However, it should be pointed vourable counter-sacrifice with
out that the black knight shields 20 ':'xf7 ! 'i'xti 2 1 ':'xf7 �xti 22
the white e-pawns from frontal i.h3 followed by 23 i.e6+. One
attack, so Black will not auto­ of White's main problems is his
matically place it there.) incapacitated bishop on g2, so
iii) Black has an absolutely Black would be very foolish to
solid pawn structure. Rooks allow it onto e6.
thrive on open files, yet White 20 .te4 :f8
only has one semi-open file, the 21 i.d3 .te8
f-file, and Black has time to for­ 22 'ii'e4 g6
tify it. In what follows, White 23 ':'f6 �g7
48 Positional Sacrifices

24 :6f4 (with a slight advantage

to Black), but such moves are
easier to recommend than to play.
24 R.d7
2S h4?
This weakens the kingside and
allows Black's next move. 25
'ifh4 or 25 :6f4 were still best.
2S 'i'xf6!
26 :xf6 �xf6
Now we have a different ma­
terial balance (which is fully ex­
amined in another chapter). How­
So White has done all he can. ever, the nature of Black's advan­
He has posted his rooks on the tage has not greatly changed.
semi-open file; he has manoeu­ White's pawns are weak, he has
vred his bishop to a more active no way to puncture Black's solid
square (true, it is a target there); formation, and his bishop on d3 is
and he has provoked a slight an inactive, vulnerable piece. One
weakness in Black's kingside. of the greatest skills in chess is to
But now what? He needs his retain an advantage in spite of
pawns to advance somehow to simplification.
open lines for his pieces and ex­ 27 'ilVd4 :te8
ploit whatever chinks there are in 28 <ifl>f2 �g7
Black's position. There are two 29 <ifl>el f6
problems with this plan: 30 <ifl>d2 lbg4
i) The lack of good pawn ad­ Provoking 3 1 e4 in order to
vances. further impede the bishop on d3.
ii) The absence of chinks In 31 e4 lbes
Black's position. 32 'i'c3 :e7!
This means that, although at Now Tiviakov expounds his
first glance there is nothing plan in schematic terms:
wrong with White's position, it i) He will play his bishop to e8,
has no potential. He can only wait his knight from d7 to c5, his rook
to see if Black finds a way to to e5, then . . . g6-g5 and . . . .i.g6;
strengthen his position. Razu­ ii) or his knight to f7, rook to
vaev, however, does not want to e5, king from f8 to e7, then
wait, and pretends he is doing knight from d8 to b7 to c5,
something, which only weakens bishop to e8, . . . g6-g5 and . . . .i.g6.
his position further. In both cases, White will lose
24 c4 his e-pawn without gaining ob­
Tiviakov sU!l:gests 24 'ifh4 or taining any counterplav. After the
 Sacrifices to Destroy the Opponent 's Centre 49

demise of the e-pawn, all White's passivity) was not exactly inspir­
other pawns will be more vulner­ ing. Black could have carried on
able. That Black can plan so as outlined at move 32.
deeply and without considering 39 dxcS
White's intervening moves shows 40 'fibS lbd6
the lifeless nature of White's 41 .c6 <iti>d8
position. 42 .tel
White finds his brief flourish
of counterplay has faded. If 42
'iVa8+ .ics and the e4 pawn is
hanging. Now he must pay the
price - Black's knight has been
given the d6 square and Black has
a passed pawn.
42 �c8
43 .a8 gS!
44 'iVc6
After 44 hxg5, Black can
choose between 44 .. .lhg5
(winning the g3 pawn but allow­
33 .a3 as ing the clearance sacrifice 45
34 .b2 lbt7! e5 ! ?) and 44 . . . fxg5 followed by
Better than 34 . . . �eS 35 c5 ! advancing the kingside pawns to
with counterplay after 35 . . . dxc5? create another passed pawn.
36 d6 ! or 3 5 ... bxc5 36 'it'bS. Now 44 gxh4
of course if 35 c5 dxc5 36 d6 45 gxh4 %thS
lbxd6. 46 eS
35 .tc2 %teS The best chance.
36 .a3 cwt>f8 46 fxeS
37 .ta4 .tg4 47 .a4 e4
Black does not want to ex­ Cutting off the defence of the
change his good bishop for h-pawn.
White's feeble cleric. 48 .b3 .tb7!
38 'iVd3 r/;e7 Black must keep control. If
39 cS? 4s Jlxh4 49 .c3 ! and White's
..

White sees Black's plan queen enters on f6 or hS, harass­

gradually unfolding, and loses ing Black's king with checks.
patience. Grandmasters know that 48 . .th7 vacates cS for the king,
..

you need counterplay, and it is and also after the elimination of

not surprising that Razuvaev the d5 pawn, the c6 square be­
lashes out. Tiviakov's recom­ comes a flight square for the
mendation of 39 .tc2 (with total king. If the king can reach b7, it
50 Positional Sacrifices

will be perfectly sheltered behind 59 � d2 e3+ 60 �c 1 lbf4 6 1 .td l

the queenside pawns. c 3 is not clear after 62 'ikn, but
49 'iWc3 .txd5 59 . . . c3+ 60 �c 1 lbb4 wins, e.g.
50 �f6+ �d7 61 .i xe4 c2 threatening
51 a4 <it>c6 62 . . . lb a2+, or 6 1 h5 :g5 1 fol­
Tiviakov points out that lowed by . . .l::t g l + and . . . c3-c2.
5 1 . . .l:tf5 followed by 52 . . . IU7 59 c3
was simpler. The rook is slightly 60 i.xe4 tDb4
misplaced on h5, which gives 61 i.g6 i.xa4
White some hope. However, the 62 h5 c2
ultimate result of the game is not 63 i.xc2 .ixc2
affected. 64 'ikt7 a4
52 'fie7 <itb7 65 h6 a3
53 <itc3 .tc6 0-1
54 �b2 c4 Black' s pawn queens with
55 �c3 tDc8! check.
An excellent move. In order to
keep h4 defended, White's queen We will now leave our discus­
has to temporarily give up its at­ sion of exchange sacrifices and
tack on the h7 pawn. This allows consider the use of pawns to
Black to play . . . l:ic5 with gain of shatter the opponent 's centre.
time. Here is one of the earliest known
56 �f6 l:tc5 examples of such a sacrifice :
Now Black threatens 57 . . . h5 ,
safeguarding his h-pawn, fol­
lowed by gradual preparation of a
queenside pawn advance. White 's
one hope is to win the h-pawn
and create his own passed pawn.
57 �h6 tDe7!
Now we see another reason
why 55 . . . tDc 8 1 was such a good
move. The knight reaches d5 and
White' s blockade of the queen­
side passed pawn crumbles.
58 'ii'xh7
The passed h-pawn is a glim­ Apparently this sacrifice was
mer of hope for White, but it is played in Baghdad in the tenth
too late. century by the great Arab player
58 lbd5+ as Suli, the leading haliyat or
59 �d4 grandmaster of his time. Old rules
Going backwards is no better. of course apply: the queen is the
 Sacrifices to Destroy the Opponent 's Centre 51

weakest piece. White's opening apparently has a safe game. It

system was in fact known as the seems that in a few moves time
' torrent' pawn, since it sweeps the opponents will be shaking
into Black's position and wreaks hands and agreeing a draw.
destruction (or so it was hoped) . White's next move, offering the
White played 1 f5 ! to break up exchange of queens, strengthens
Black's pawn phalanx. There this impression. Unfortunately for
followed: 1 . . .exf5 2 gxf5 gxf5 . Black, there are some hidden
Unfortunately for Black, bishops reefs in this position.
could only move two squares in 23 "al :a8?
those days, so 2 . . . .txf5 is not le- Not suspecting any danger.
gal. However, 2 . . . g5 keeping his Better was 23 . . ...xa l 24 ':xa l
pawn structure intact was better. exd4 25 .i.xd4 :a8 . White can
After 2 . . . gxf5 3 .i.h3 !i:Je7 4 1:.£1 then claim a small advantage
lU8 5 !i:Jg3, White regained his since his bishop pair are very ac­
pawn with a clear positional ad­ tive. However, the a- and b-files
vantage. are stripped of pawns, so Black's
short-range knights are unlikely
The rules of chess may have to be overstretched in their de­
changed over the years, but the fensive task.
spirit of the 'torrent ' pawn lives 24 "xa2 lIxa2
on in modem chess, as the fol­ The rook may look impressive
lowing games demonstrate. on this square, but it would be
better on a8, defending the back
rank.
Karpov-Romanlshin
2S dS .i.b7
Tilburg 1 993
Equally unpleasant is 25 . . . .i.a8
26 l:b l , and already there is a
threat of 27 c5 ! dxc5 28 !i:Jc4
followed by capturing on e5 , de­
molishing Black's centre.
26 :bl (D)
26 .i.a6
Romanishin may have thought
he could play 26 . . . !i:Jc5 in this
position, but then 27 !i:Jb3 ! !i:Jxb3
(27 . !i:Ja4 28 .i.a5) 28 :lxb3 1:a7
. .

29 c5 (threatening 30 c6) or 29
iLa3 ! ? is strong.
However, Black should have
Black has withstood some admitted that his rook on a2 is
early pressure from Karpov, and misplaced and tried 26 . . J:ta7.
52 Positional Sacrifices

Now if 27 ..i.a5 then not 33 ..i.f4+) 33 i.xf6 gxf5 34 exf5

27 . . . llxa5 28 :xb7 l:c5 forces mate.
(defending c7) 29 ..i.f1 ! threaten­ 30 i.xg4 hxg4
ing 30 �b3 winning, but 31 :al!
27 . . . .i.xd5 ! when all Black's Planning to exchange rooks,
problems are over. So White which will free his king to march
would have to answer 26 . . . 1:a7 up the board and capture one or
with a different approach, e.g. 27 more of Black's loose pawns.
f4! ? combined with c4-c5, trying Romanishin has no wish to be
to break through in the centre. slowly tortured by Karpov's fa­
mous technique, and throws him­
self on the sword.
31 lle2
32 llxa6 llxe3
33 L8+ �h7
34 e6
The triumph of White 's strat­
egy. This pawn cannot be
stopped.
34 fxe6
3S dxe6 �f6
36 e7 :lc1+
37 ct>r2 :c2+
27 eS! 38 �e1 1-0
Now 27 . . . �xc5? allows mate
and 27 . . . h6 28 cxd6 cxd6 leaves If you look at the position at
Black's d-pawns vulnerable to a move 23, it is hard to believe that
subsequent i.b4. Nevertheless, Black will be in serious trouble
this is how Black should play; within five moves. Such is the
after the game continuation, his strength of Karpov's style (and the
centre disintegrates. centre-busting 27 c5 !). In the fol­
27 dxcS lowing example, another seem­
28 f4! ingly solid position is fragmented
Completing the demolition. by an unexpected pawn stab:
After 28 . . . exf4, White can con­
tinue 29 gxf4 and his central
Aldama-Vera
pawn mass will quickly become
C u ba 1 993
overwhelming.
28 hS (see following diagram)
29 fxeS �g4
If 29 . . . �xe5 30 .i.xe5 llxd2 3 1 Nonnally, Black's two bishops
l:lb8+ �h7 32 ..i.f5+ g6 (32 . . .<ith6 would outshine White's knights.
 Sacrifices to Destroy the Opponent 's Centre 53

However, the knights have the and Black's position was sud­
excellent d4 square which is im­ denly hopeless. Whichever way
pregnable to pawn attack. This he captures the pawn, his centre
means they are securely central­ falls apart and then his king faces
ised and defend each other in a massive attack. And quite right
chain-like fashion. Black's queen­ too. Black is justly punished for
side pawns have raced forwards in his neglect of the centre.
an attempt to attack White's king. 34 jlxe2
Although they have forced the The alternatives make grim
concession b2-b3 from White, it is reading for Black:
Black's king, not White's, that has a) 34 . . .gxf5 35 g6 ! (the the­
been left more exposed. matic undermining of the centre)
35 . . . .i.xe2 (35 . . . �xd4 36 g7
wins) 36 'i'xe2 ! �xd4 37 g7.
Now since 37 . . . �xe5 38 g8('i')
defends against 38 . . .... g 1 +, Black
has to try 37 . . . axb3 38 cxb3 'ii'a 3
(38 . . . jle3 39 'ii'b 5+ ! ) 39 'ii'b 5+
it.b6, but now 40 l:Ib8+! �xb8 4 1
'ilfxb6+ and 42 g8("') wins.
b) 34 . . .exf5 35 e6 ! jlxe2 36
'ilfxe2 ! it.xd4 37 exfl l:Ixfl
(37 . . . .t c5 38 fS ('ii' ) jlxfS 39
'i'b5+ 'ii'b 6 40 lIb8+ �xb8 4 1
"'xb6+ <itc8 42 "'xg6 wins) 38
But how is White to exploit the 'iib 5+ .tb6 (38 ... 'i'b6 39 "'d5+
Black king 's lack of pawn cover? wins the rook) 39 'i'xd5+ �a6 40
If the white queen tries to go "'c4+ ! �a5 (40 . . . �b7 4 1 "'c8
'round the edge' with 34 � l and mate) 4 1 l:Ie5+ and wins.
35 'ii'h 8 , the knight on d4 is cap­ These variations are given to il­
tured; and advancing pawns on the lustrate how violent the world be­
queenside would only endanger comes for Black's king once the
White's king. Besides, it would centre collapses. In the game, it was
make no sense to attack Black di­ by no means necessary for White to
rectly where all his pieces are carry out such an exhaustive
massed. If the white pawn were on analysis before playing 34 f5. A
g4 rather than g5, we would sug­ little calculation would be sufficient
gest the plan of f4-f5 to break up to discover that the pawn sacrifice
Black's centre. But as it is, the was very promising.
advance f4-f5 is impossible. Or is 35 lLlxe2 axb3
it? White in fact played 36 axb3 exfS
34 fS! ! Losing quickly. but 3 6 . . . �xf5
54 Positional Sacrifices

37 g6 fxg6 38 1he6 :d7 39 ':xg6 pressure against his d6 pawn with

leaves Black's king bereft of all 22 g5! ?
shelter and his pawns weak and If the bishop now retreats to
scattered. c 1 , then 23 . . . d5 ! ? looks good,
37 �xd5+ �b6 exploiting the pin on the e-file.
Perhaps Black was hoping to Instead, Pigusov dismantled
play 38 . . . 'ilfa6 and 39 . . J�a7, but Black's centre with a tactical
this counter-attacking idea is eas­ motif that is common in the
ily defeated by White' s immedi­ King ' s Indian defence:
ate pawn break in the centre. 23 ttJdb5! cxb5
38 e6 fxe6 24 i.xd6 "i¥c8
39 Ihe6+ �a5 25 i.xe7 lhe7
40 ttJd4 1-0 26 ttJxb5
Black resigned since 4 1 ttJc6+
is a winning threat, e.g. 4o . . . 1Wb7
4 1 ttJc6+ \t>b5 42 \i'd3+ followed
by 43 ttJd8+.

As a final example, we see a

dramatic encounter in which
White demonstrates that the
question of central control can be
solved by a revolutionary rather
than an evolutionary approach:

Pigusov-Akop i a n
The dust has settled and White
Novosibirsk 1 993 has a rook and two pawns for a
knight and bishop. In a middle­
game position, with all the other
pieces still present, two pieces
normally outweigh a rook and
pawns, especially when the ad­
vantage of the bishop pair is
added.
However, Black IS very
cramped in the game position.
The knights have no safe central
squares and will be dislodged by
White' s pawns. On being driven
back, these knights will interfere
Black to move tried to ease the with the action of Black's other
 Sacrifices to Destroy the Opponent 's Centre 55

pieces and stop them from func­ 28 :e6

tioning properly. 29 b4 liJb3
26 liJe8? Whereas White 's knight is
Black anticipates 27 e5 and dominant on d5 , this knight be­
voluntarily retreats his knight. comes cut off from the action in
The attempt to ease his cramp by the centre and on the kingside.
exchanges also fails: 26 . . . �xb5 30 f4!
27 cxbS (threatening b3-b4) So that after the inevitable e5,
27 . . . 1:.c7 28 b6 :c6 29 e5 leaves the bishop on g7 is shut out of the
all Black's pieces hanging. But game, and the knight on b3 will
26 . . ....b8 ! was much better. It be further isolated (it can no
rules out 27 e5 , dissuades the longer be played to d4 under the
plan of f4, and leaves the knight gaze of the g7 bishop) .
on f6 to challenge control of d5 30 gxf4
(see White's next move). After Hopeless is 30 . . . hS 3 1 eS h4 32
27 b4 axbS 28 axb4 .txb5 29 f5 ! hxg3+ 33 �h l and wherever
cxb5 liJbd7, Black has solved his the e6 rook goes, a fork on e7
problems. So White would will be fatal.
probably continue with 27 liJc3 31 gxf4 �
when a tense game would be in Ruling out the e7 fork men­
prospect. tioned above, but Black's posi­
27 liJc3! tion becomes more and more
Now the knight is redeployed constricted with every move.
to the excellent dS square. 32 e5 l:tc6
27 a4 33 "'f2!
The white knight proves so The queen finds a way to break
strong on d5 that 27 . . . .i.xc3 was into Black's position. Of course,
worth considering. To give up the 33 . . Jlxc4? 34 lbb6 wins.
powerful dark-squared bishop is 33 .i.f5
anathema to Kings Indian players 34 c5!
(see the section on the Indian The white pawns now hold
bishop) . However, the position Black's pieces in a pincer-like
remains unclear after 27 . . . �xc3 grip.
2 8 'ii'x c3 b6 29 b4 axb4 30 axM 34 h5
lbb7 3 1 f4 .i.e6, etc. In any case, 35 lbe3 !
the game continuation is so bad The knight has done its duty on
for Black that he should have d5. Now it is necessary to clear
tried this line. the d-file so that White 's heavy
28 liJd5! pieces can penetrate into the heart
Not allowing Black a second of Black's position.
chance to play . . . .txc3 ! (e.g. 28 35 .i.h6
b4 .i.xc3 ! ) . The poor bishop exchanges one
56 Positional Sacrifices

blocked diagonal for another. �xc6.

36 'it'h4 36 l:tg6
Threatening 37 l:td8 'it'e6 38 37 I:d8 l:.xg2+
'it'xh5 �xf4+ (what else?) 39 Desperation, but if 37 . . . 'it'e6 38
<t>h l and not only is the bishop on �d5 wins (38 ... 'it'e7 39 lLlxf5).
f5 hanging, but there is also mate 38 lLlxg3 'it'c7
in two with 40 'i'h8+. Note that 39 'i'f6 1-0
White was not side-tracked by the Black resigned. His position is
mere win of an exchange with 36 in ruins.
 5 Sacrifices to Open L i nes

When the British army in the Su­ defensive line. The Hadendowa
dan was attacked at EI Teb on 29 placed in ambush saw their
February 1 884 by a superior force chance and suddenly swarmed
of Hadendowa tribesmen, they into this gap. The invincible de­
formed the customary 'British' fensive formation had been bro­
square, a defensive formation of ken by a little cunning.
incredible strength. It consisted In chess, we also try to cajole
of one row of troops within an­ our opponents into weakening
other row, deployed as the name their line of defence. We saw in
suggests in a square formation. the last chapter how a pawn can be
As one row stood and fIred, the used as a battering ram to break up
other knelt to reload, so that con­ our opponent's formation. But if
tinuous fIre was possible. Mean­ the defensive line is strong, we
while, a double row of bayonets often need a more subtle form of
presented a formidable obstacle sacrifIce: pawns must be deflected
to anyone daring to approach at from their important function of
close range. The British boasted guarding a key central square, or
that the square had never been enticed by the lure of material gain
broken. into relinquishing the blockade of
The Hadendowa tribesmen, a key diagonal. Once the removal
however, had other ideas. They of the pawns has been achieved,
concealed some warriors in a ra­ our pieces - the tribesmen waiting
vine, out of view of the British in ambush - can seize control of
troops, then the main body of the vacated squares and diagonals.
tribesmen attacked one side of the This is the theme of this chapter.
square. As would be expected,
they were beaten off with heavy
I . Sokol oY-Holzl
losses, and had to retreat.
Bru n n 1 99 1
The British, sensing victory,
advanced in pursuit, but in mov­
ing forwards, a gap opened in the 1 d4 liJf6 2 c4 cS 3 dS a6 4 'ii'c2
58 Positional Sacrifices

e5 5 e4 �d6 6 liJc3 �c7 7 .ltd3 liable to attack with f2-f3 and g2-
d6 8 liJge2 liJh5 g4.
Played to avoid an unpleasant 9 'fif6
pin after 9 .ltg5 . A remedy that is worse than
the disease, since now Black's
queen becomes a target. Perhaps
9 . . . h6 was better.
10 �g5 'ii g6
11 l:gl !
Black may have underesti­
mated this move. The threat is 1 2
g4 liJf6 1 3 f4! h6 ( l 3 . . . exf4 14 e5
wins the queen, as does
1 3 . . . .axg4 14 f5 'ii'h5 1 5 liJg3) 1 4
f5 'iWh7 1 5 .ae3 and White will
castle queenside followed by g4-
g5, suffocating Black on the
Black has shown much mis­ kingside.
placed ingenuity in manoeuvring 11 h6
his bishop to c7. If the position 12 �e3 'ii'f6
remains closed, Black can com­ Black should develop with
plete his development and may 1 2 . . . liJd7 and defend doggedly,
one day be able to justify his ec­ e.g. 1 3 g4 liJhf6 14 liJf4?! 'fib7 ! ,
centric play. Though even in the though it must be admitted that
middle game, the bishop would be his position after the alternative
better on e7 (to exchange itself 14 h5 'ii'h7 1 5 0-0-0 'ii'g 8 1 6 liJg3
for White's good bishop with (intending liJf5) does not inspire
. . . .ltg5) or fianchettoed on g7 (to confidence.
bolster a plan of . . . f7-f5 attacking 13 0-0-0 liJf4?
White's centre). White must at­ This is inexcusable reckless­
tempt to punish Black immedi­ ness. Evidently, Black completely
ately for his faulty strategy and missed White's devastating sac­
neglect of development. To do so rifice on move 1 5 . 1 3 . . . liJd7 was
he needs to open lines. Therefore now the only chance.
he began with: 14 liJxf4 exf4
9 h4! 15 e5! !
This threatens 1 0 .ltg5 when An explosion which releases
Black must play either 1 0 . . .'iNd7, all the pent-up energy in White's
when the queen deprives his position. The d3 bishop gains an
knight of the natural d7 square, or open diagonal and the knight on
1 O . . .f6 1 1 .td2, when the knight c3 the lovely central e4 square.
on h5 has its retreat cut off and is Black now discovers why the law
 Sacrifices to Open Lines 59

of rapid development has held loses the bishop on c7 (the only

sway since the days of Morphy. contribution this ill-starred piece
has made to the game) . But if
Black does not play .. .fxe6, he
can never hope to develop his
bishop on cS . This means that his
queenside rook and bishop re­
main spectators as their king is
hunted down.
19 <ifi1f8
20 :dgl .td8
The bishop decides to cut short
its holiday on c7, since the news
from the kingside is most alarm­
ing.
15 "xh4 21 f4
Despair. Dislodging Black's one active
1 5 . . ....xe5 16 i.xc5 0-0 piece.
( l 6 . . . dxc5 1 7 :ge l ) 1 7 i.d4 ! 21 i. f6
'*115 1 8 g4! gives a decisive at- A valiant try, since after
tack, e.g. 1 8 . . ....xh4 ( l S . . . i.xg4 2 1 . . .lDxd3+ 22 "xd3 the fl
1 9 f3 ! wins) 19 g5 h5 ( l 9 . . . hxg5 square collapses.
20 :h I and 2 1 .th7+ wins) 20 22 fxeS!
l:[h l "'g4 2 1 :dg l .d7 (2 1 . . .•f3 Its easy to be brilliant when
22 .te4) 22 :'xh5 and wins. your opponent has a rook and
1 5 . . . dxe5 1 6 i.xc5 is also bishop shut out of the game.
hopeless, since Black can never 22 i.xg7
castle, as 1 6 . . . i.d6? 1 7 lDe4 wins. 23 lbg7!
16 g3! One can sympathise with the
But now the g-file is opened black bishop. After many adven­
and Black's king loses any hope tures it arrives at its correct
of finding refuge on the kingside. square - the one it should have
16 fxg3 gone to at move 7 or 8 - but only
17 l:xg3 lDd7 in time to be part of a winning
Much too late. Black is in a combination by the opponent.
hopeless dilemma: if he castles 23 �xg7
kingside, he is mated; if his king 24 .. g2+ <it>fS
stays in the centre, he also faces a 25 ...O! f6
massive onslaught. There is no way to defend fl . If
18 e6 lDeS 2S . :fle7 26 exd6 'ife8 27 e7+
.

19 Ibg7 �g7 2S "g3+ soon mates.

If Black plays 19 .. .fxe6, he 26 exf6
60 Positional Sacrifices

Our familiar passed pawns. Apparently, Black has a safe

Although Black has the theoreti­ and promising position. His
cal material advantage of two knight is well anchored in the
rooks for two pieces, his queen­ centre, and he is ready to begin
side forces remain entombed. an attack on White's king with
26 :g8 . . . a7-a5-a4, etc. White, on the
27 'ii'f4! other hand, has difficulties in at­
An elegant fmishing stroke. tacking Black's king in view of
After 27 . . . 'ii'xf4 28 .i.xf4, d6 can the blocked nature of the terrain
no longer be defended. Black on the kingside.
cannot even give up a rook to However, a closer look at the
stop the passed pawns, e.g. position reveals that Black has
28 . . . :h8 (28 . . .:g7 29 Ji.xh6 ! ) 29 some serious weaknesses on the
Ji.xd6+ <ite8 30 f7+ <it'd8 3 1 e7+, dark squares, especially along the
etc. a 1 -h8 diagonal. If White could
1-0 give a queen check on this diag­
So Black resigned. His queen's onal, it would be mate. Likewise,
rook and bishop never even if a knight could get to f7. The
moved. black knight on e4 is also not as
securely placed as may appear at
The following example is less first glance, and can be under­
drastic. Here, White purposefully mined.
improves his position until a All these considerations sug­
dark-squared highway opens up gest White's correct plan: he
into Black's position. must stage a breakthrough in the
centre which both dislodges
Black's knight and gains access
Andersson-M i l os
for the queen to the key dark­
Tilbu rg 1993
square diagonal leading to
Black's king. Andersson's play
shows us how this is done.
20 lDd2! lDd6
The knight voluntarily retreats,
anticipating White's next move.
Note that 20 . . . lDc3+ 2 1 bxc3
bxc3+ 22 lDb3 followed by 'ii'xc3
is not dangerous for White.
21 f3 cS
22 dxcS lDxcs
23 e4
Carrying out his plan. Black
tries to block the position and
 Sacrifices to Open Lines 61

prevent the fragmentation of his fxe3 (29 . . . 'illxe3 30 'ii"xb4) 30

centre. 'ii"g3 and White has dangerous
23 dxe4 attacking chances.
24 fxe4 f4 However, Black could have
25 e5 �xd3 tried 27 . . . ike3 . The black queen
If 2S . . . ttJfS 26 e6 ! is strong, swoops into the centre and chal­
e.g. 26 . . . 'ifxe6 27 .i.xfS ii'xf5 28 lenges White's queen. Best play
"fIxe7 or 26 ... ttJe3 27 ttJc4 lbxc4 is then 28 �c4 (28 'tifl �g3)
28 .i.xc4, and again Black has to 28 .. :i'xe l 29 lIdxe l and White
worry about sudden death on the has a clear advantage, e.g. 29 . . . aS
al -h8 diagonal. 30 ttJb6 :b8 3 1 �xc8 ltbxc8 32
26 �xd3 �f5 �eS ttJd6 33 ttJt7+ or 29 . . Jlb8
Black may have been feeling 30 �ceS .1xe6? 3 1 �c6. If Black
very pleased with himself in this attempts to improve his king
position. He has only to play position, then simply �xb4 will
27 . . . �e6 and the troublesome a 1 - give White a good endgame.
h8 diagonal will b e cleared for­ It is notable that when Black
ever. Then he can continue his deals adequately with one mating
own attack on White's king be­ theme - a queen check on the a 1 -
ginning with moves such as h 8 diagonal - he is brought down
. . . �e3 and .. J�gc8, and would by the secondary mating theme of
have good chances of success. ttJeS-t7.
However, White frustrated all 28 ttJe4 l';tb8
Black' s hopes with a thematic Not 28 . . . .i.b7 29 ttJdcS fol­
pawn sacrifice: lowed by 30 'it'xb4. 29 ttJxgS ! ?
27 e6! may also be strong.
29 �xb4 ltb5
This leaves the back rank
weakened, allowing White a
winning combination. However,
there was no other good defence
against the threat of 30 'ii"c 3+.
30 'ii'c 3+ 'it'e5
31 �xg5!
Now Black's position collapses
in a few moves.
31 ':'xg5 32 :td8+ l:tg8 33
.•

l1xg8+ �xg8 34 'i'xc8+ �f7 35

'ii'c4+ e6 36 ttJd3 'i'b8 37 :el
27 'i'xe6 I:[b6 38 b3
If 27 . . . �e3 28 �fl ! .txe6 (the Here Black overstepped the
knight dare not move) 29 4\xe3 time limit. He is of course quite
62 Positional Sacrifices

lost - he is already a pawn down 23 ct>d7

and the f-pawn is also dropping, but not surprisingly his king
to say nothing of his exposed proved fatally exposed:
king. 24 h5 ltJe7
25 'it'xg7 l:.g8
Note how all sorts of tactical 26 'it'f6 'ile8
possibilities appeared for White 27 �g5!
after the sacrifice 27 e6 ! . Here is This threatens 28 :'a l followed
a similar example: by 29 :a7 winning. Hence
Black's desperate reply.
27 h6
Rodrig uez-Sori n
28 .t xh6
Matanzas 1 993 Or 28 jth4 followed by 29
:ta l .
28 l:.g4
29 i-f4 'iWg8
30 g3 'ii'a8
31 'iWe5 'i'a5
32 h6 1-0
The pawn queens.

Let us return to the diagram

position above. Imagine if after
2 1 . . .ltJg6, White had played an
inconsequential move, for exam­
ple 22 r.tth l , instead of 22 e6. Then
Black tried Black could play 22 . . . i-f5 ! elimi­
21 ltJ g6 nating White's powerful light­
and was struck down by squared bishop, followed by . . . 0-0
22 e6! with a safe game. So clearly after
that move again! 2 1 . . .ltJg6, White must act fast,
22 jtxe6 since 22 . . . .tf5 is a strong posi­
23 h4! tional threat. That is why he
There is no defence against the played 22 e6 ! to rule out 22 . . . .tf5
threat of 24 h5 followed by, after - the bishop is pinned along the e­
the knight retreats, 25 'ii'x g7, file after 22 . . . .txe6. Also, White
decimating Black's kingside. gained time after 22 ... .txe6 23 h4
Black cannot even escape the to threaten 24 h5, since with the
worst by sacrificing the ex­ bishop no longer on g4, the white
change, since 23 . . 0-0 24 h5 ltJe7
. queen's route to g7 is unobstructed
25 i-h6 g6 26 'iWe5 ! wins a piece. and there is no option of . . . .txh5.
So in the game, he played 2 1 . . .ltJg6 therefore failed to
 Sacrifices to Open Lines 63

solve Black's problems. Perhaps 25 'ii'xg7 wins) 24 .txe7 �xe7 (if

he should simply have castled? 24 . . . i.g6 25 .tc5 leaves Black
Then 2 1 . . .0-0 22 e6 i.xe6 is unable to castle and facing the
nothing. Or if 22 i.xh7+ �xh7 threat of l:ta l -a7 combined with
23 'fIh4+ <ii? gS 24 'ilixe7 'iie6 and 'it'd6) 25 'iixg7+ �d6 26 :a l
the presence of opposite-coloured .ta4 (else 27 .:r.a7 follows, with
bishops makes White's winning threat of mate on e5 or e7, e.g.
task very difficult, despite his 26 ... :gS 27 'ii'e 5+ �d7 2S lla7+
extra pawn. However, White can 'It>eS 29 'ti'd6 'ti'dS 30 'iixe6+ and
play 22 .tg5 lLJg6 23 'ii'd 3 ! to mate on f7) 27 b3 ! and White
stop Black exchanging the light­ regains his piece with a winning
squared bishops. Now he is ready attack on Black's king, or after
to retreat his bishop from g5 and 27 . . . .txb3 wins as in the brack­
begin a general central/kingside eted note above after 2S 'ti'e5+.
advance beginning with f2-f4 and
f4-f5, etc. In view of White's
strong bishops and superior pawn
structure, Black would face an
arduous defensive task. From
these variations, a general con­
clusion can be drawn about the
position: if Black is to equalise,
he must exchange light-squared
bishops without being immedi­
ately punished.
So why not play 2 1 . . . .tf5 im­
mediately, in order to challenge
White's good bishop? This looks (analysis diagram)
good, since 22 .txf5 �xf5 fol­
lowed by . . . 0-0 is best avoided by However, there is a third op­
White. However, White can still tion: 22 . . . 0-0! and Black escapes
try the pawn sacrifice 2 1 . . . .tf5 22 punishment, e.g. 23 .tg5 .txc2
e6 ! ? (D) . 24 i.xe7 lteS 25 .tf6 (25 exf7+
Now Black is best advised not �xf7 26 'ii'f4+ 'iif5) 25 . . . .tg6 26
to take the bishop: 22 . . . i.xc2 23 exf7+ <ii? xf7 27 IbeS fixeS.
'i¥xg7 :f8 24 .th6 �g6 25 e7 But finally, after 2 1 . . .i.f5,
llhS 26 'iifS+ ! ! and wins. White can play 22 'ti'xg7 :gS 23
Taking the pawn is also fraught .txf5 ! 'ii'xf5 24 'iif6 with an extra
with danger: 22 . . .fxe6 (22 . . . .txe6 pawn, instead of 22 e6 ! ? It is not
23 '1i'xg7 Il.gS 24 '1i'xh7 is totally always best to sacrifice!
bad) 23 i.g5 ! i.xc2 (23 . . . 0-0 24 By now, the reader will be
i.xe7 or 23 . . . 1i'b7 24 .txe7 i.xc2 aware that even the simplest or
64 Positional Sacrifices

the most positional of sacrifices

often requires long and detailed
tactical calculation. But in truth,
this complexity applies to any
manoeuvre on the chessboard, not
just a sacrificial one. Chess is a
very complicated game, and
every position has to be adjudged
on its unique features. We should
rejoice in this, since otherwise
chess would have succumbed to
the 'draw death' predicted in the
1 920s. Anyone who had mastered
the laws of strategy to a high de­ a) If Black plays 1 8 . . . gxfS,
gree would be invincible. Bot­ then 1 9 .thS+ and now 1 9 . . . �ti
vinnik, despite his venerable age, 20 .txti + �xti 2 1 'ii'h S+ leads
would still be world champion to a mating attack after 2 1 . . . �g7
(even if incapable of scoring 22 �h l followed by :g l +; or if
more than a draw against young 2 1 . . .�g6 then 22 �f4 :g8 23
rivals such as Fischer or Kas­ �h l followed by :g l , winning
parov) ! the pinned knight. (Kamsky
Fortunately, chess has re­ probably only calculated as far as
mained too complex for even the 2 1 'ii'h S+ r:3;g7 22 �h l . His
brightest minds to fully master. judgement told him there must be
Here is a good example of a a win with Black's king in such
modem struggle between two an exposed position.) So Black
young players who undoubtedly must play 19 . . . �d7.
understand chess strategy. Their
play is full of invention and fi­
nesse - but chess perfection is not
granted to man !

Kamsky-Bareev
Biel 1 993

(see following diagram)

Here White played

18 fS! (analysis diagra m)
with the following ideas:
 Sacrifices to Open Lines 65

White has no immediate win lhe7 wins) 25 l:hh7 lbxh7 26

here, but Black's pieces are hope­ lbb2 �c8 27 ..txh6 lbf8 28 a4
lessly unco-ordinated. His king on .i.a6 (28 . . . .i.d7 29 'iig 8) 29 'ii'g 8
d7 boxes in the bishop on c8, lbd7 30 lbf4 'ii'c6 3 1 'iie 8 ! .i.h4
which in turn shuts the queens 32 lbg2 and wins.
rook out of the game. The knight Alternatively, White could plan
on d8 has no safe move, and a further sacrifice after 1 9 . . . 'iil d7
merely takes away the d8 square to destroy Black's centre and
from the king. This prevents an mate his king. This would begin
'unwinding' with ... �d8, followed with lbf4 and 'iWf3 , followed by
by . . . .i.d7, . . . .i.e8 and . .,ti:Jf7. lbxd5 . White, however, should
Black therefore has no con­ not be in a hurry to play lbxd5
structive plan after 19 . . �d7. . since, as stated above, Black can
Meanwhile, White can strengthen do little to improve his position.
his position move by move. This So for example White could play
is a key feature of a good posi­ l1ad l (after 'ii'f3 of course) so
tional sacrifice. Although the that the rook would be on a gen­
position at first glance may seem erally superior square once lbxd5
fairly equal, one side is playing was played.
with a plan, while the other is But as a general rule, you
floundering around wondering should not risk a sacrifice when
what to do. He can merely re­ you have a clear, simple winning
spond to direct threats and not method. In the position after
undertake any direct campaign. 1 9 . . . Wd7, Black is defenceless
Thus, after 19 . . . �d7, White against White's plan of penetra­
can plan an invasion along the g­ tion down the g-file. Therefore
file, beginning with 20 �h l fol­ unless everything has been care­
lowed by doubling rooks on the fully worked out, so that the
g-file. If Black tries to contest lbxd5 breakthrough is merely a
control of the g-file with his own technical device and not a true
rook on h8, by say 20 . . . :g8 , then sacrifice, the gradual approach
White can simply exchange it off with �h l and :g 1 is to be pre­
by 2 1 :g l , when he will be able ferred. Now we will return to the
to bring up reinforcements, i.e. position after 1 8 f5 ! .
the rook on a I , while Black's b) Bareev in the game an­
own reserves are entombed on the swered with
queenside. A sample variation: 20 18 exfS
�h 1 b6 (Black must try to de­ He may well have rejected
velop somehow) 2 1 ':'g l a6 22 1 8 . . . gxf5 19 .th5+ 'iil d7 as an
bxa6 �xa6 2 3 �g7 .i.b5 24 'iVg l option purely on intuitive
(threatening 25 �xe7+ ! ) 24 . . . �h7 grounds. As the readers experi­
(if 24 . . . �c8 25 lbxb6+! 'i'xb6 26 ence grows, he or she too will
66 Positional Sacrifices

learn to make instant judgements ertoire of ideas. However, he

such as 'Oh, 1 8 . . . gxfS 1 9 .i.hS+ should not imagine that the next
�d7 looks terrible - it 's not even time he reaches a similar position
worth looking at. ' A well­ in a game that such a sacrifice
developed positional sense is a must be the correct strategy.
must for any strong player, since 20 �f4 .i. gS
it saves valuable time and energy Black defends against the
which would otherwise be spent threat to his d-pawn by tactical
looking at 'rubbish ' . Though of means. 21 .i.xdS ? ..txf4 or 2 1
course there is one drawback: �xdS ? .i.xdS 22 .i.xgS .i.xf3 23
there may occasionally be a dia­ .i.xd8 .i.xd 1 24 .i.xc7 .i.xa4 lose
mond in amongst the rubbish! a piece.
The game continued: 21 �cS
19 .i.f3 But now that Black's bishop
Now imagine that Black's b­ has duties on gS, a gap has
pawn were on b6 in this position. opened up for White's other
Then Black could play 19 . . . .i.b7 knight. We begin to see the com­
20 �f4 'it'd7 followed by . . . �de6 pensation that White has gained
and . . 0-0-0. Black's pieces would
. for his pawn. Both knights, which
be excellently co-ordinated and were passively place on the side
he would be a pawn up. White of the board, have found a new
would have great difficulty in freedom: one in a direct way
drawing the game. Instead, the through the vacating of the f4
pawn is on b7, and Bareev could square, the other indirectly
find nothing better than through the cumulative effect of
19 .i.e6 White's pressure. Of course, the
to defend his e-pawn. But now target is the dS pawn and its de­
he cannot utilise the e6 square for fenders. Now for example, there
one of his knights, and the bishop really is a threat of 22 �xdS
proves vulnerable to attack on e6 since the knight on a4 no longer
after �f4 or �cS. White has hangs after 2 1 . . . .i.g8 22 �xdS
good play for his pawn. In other .i.xdS 23 J.. x gS .i.xf3 24 J.. x d8
words, altering the position ever J.. x d 1 2S .i.xc7. Therefore Black
so slightly, by putting a black is obliged to capture the knight.
pawn on b6 rather than b7, turns 21 .i.xf4
the exclamation mark after 1 8 fS 22 .i. xf4 'iie7
into a question mark. That is why The only answer to the threat
every position has to be carefully of �xe6 and J.. x dS .
examined and handled according 23 'iVe2
to its specific features. 1 8 fS ! is Methodical play, preparing
an excellent concept and well 'iig 2.
worth adding to the reader's rep- 23 �f7
 Sacrifices to Open Lines 67

Clearing d8 so the rook can de­ White has a crushing attack,

fend the d-pawn. e.g. 29 . . .•c7 (getting out of any
24 a4! discovered attack by White's
knight) 30 llxf5 ! ..txf5 3 1 .xf5
ltg8 (3 1 . . .11g7 32 �e6) 32 ltf1
�d8 (32 . . . 11c8 33 .e6+ wins) 33
�d7 ! 'i'xd7 34 'i'f8+ llxf8 35
l:.xf8 mate.
The value of 24 a4 is apparent
in this variation: if the pawn were
still on a2, 29 .ta3? could be an­
swered by 29 . . . b6! winning a
piece; but with 24 a4, the bishop
is defended by the rook. Thus,
29 . . . b6 can be answered by 30
�e4 'i'c7 3 1 �f6+ �d8 32 �xd5
Another useful strengthening of winning. I ' m sure Kamsky did
his position. Now Bareev gives not calculate this variation when
24 . . . g5 25 �c 1 �g6 as '\Ulc1ear'. he played 24 a4; but he probably
However, 26 �h5 ! is clearly to did consider in general terms the
White's advantage, e.g. 26 ... �f4 27 idea of �c 1 -a3 or the advance
�xf4 gxf4 28 lbxe6 followed by a4-a5-a6 followed by �b7 and
1lxf4 and the f5 pawn is doomed; �d6+. Its not surprising that fa­
or 26 . . . :g8 27 �xg6 lhg6 28 vourable variations appear
'ii'c2 ! ..tc8 (in fact it's better to al­ 'naturally' from a well thought
low White to play lbxe6 and .xf5, out positional build up.
recovering the pawn with a big ad­ Bareev however rises to the
vantage) 29 ..ta3 ! occasion and finds the correct
defensive plan, a plan he curi­
ously criticises in his own notes
to the game !
24 � gS !
Black plans to eliminate the
bishop which is the persecutor of
the d-pawn, or at least blockade
its action through entrenching the
knight on e4. Kamsky does not
allow this and removes the
knight, but at the cost of making
the plan of �c 1 and .ta3 impos­
sible.
(analvsis diaRram ) 25 �" x�5 hx�5!
68 Positional Sacrifices

If 2S . . . 'ifxgS+ 26 �h l 'ile7 progress? Not on the queenside,

(the only answer to the twin since Black can always ensure it
threats of lL'lxe6 and lL'lxb7) 27 stays unopened. For example, if
'ifg2 l:.d8 28 as keeps up the White plays a4-aS and as -a6,
pressure. There is a strong threat Black replies . . . b7-b6; or if a4-aS
of 29 a6 b6 30 lL'ld7 ! followed by and bS-b6, then . . . a7-a6 by Black
3 1 lL'ld6+ winning the d-pawn. and again the queenside is
After 2S . . . hxgS , Black has acti­ blocked. White may try to force a
vated his king 's rook and hopes favourable closure of the queen­
for counterplay against h2. side, but the actual breakthrough
26 �xe6 must come elsewhere.
Another advantage of 2S . . . hxg5 The centre is blocked, and will
is seen in the variation 26 • g2 remain so, unless White achieves
:d8 27 as g4 ! forcing the bishop the unlikely advance e5 -e6.
away from the attacking diagonal. Therefore, all White' s hopes of a
Kamsky' s move wins back the winning breakthrough rest on the
pawn. kingside. First, he has to try to
26 'ilxe6 force a weakness in Black's pawn
27 'ilg2 0-0-0 structure fS/g6. This will be very
28 'ifxgS �h7 difficult to achieve, since if
29 'ifg2 White plays �h 1 and :g 1 , Black
can simply play . . . :g8 defending
the g-pawn. And it is impractical
to use the h-pawn as a battering
ram to break up Black's pawns.
White's king would be left too
exposed.
Therefore, White has no real­
istic way to improve his position.
So a draw seems the natural re­
sult. But it is not so easy for
Black, and Bareev plays the very
move that White wants to pro­
voke:
Material equality has been re­ 29 gS?
stored, and White is a little better. The intention is noble: 30 . . . g4
Black's queen and queens rook and 3 1 . . . lL'lgS -e4, putting the
are tied to the defence of the d­ knight on a gigantic central
pawn, and the passed e-pawn is a square. Unfortunately, Kamsky
permanent long-term advantage can prevent this plan, and the
for White. weakness of the pawns remains.
But where is White to make Probably best was 29 . . . �b8.
 Sacrifices to Open Lines 69

30 .tdl! blockade will eventually crumble.

This threatens 31 %:txfS ! , which 32 :f2 %:tdg8
also answers 30 . . . g4. It also pre­ 33 ltan tbf8?
pares .tc2, pressurising the f­ A blunder, which should accel­
pawn. Sooner or later, Black will erate the end. 33 . . . :d8 was best,
be obliged to play .. .fS-f4, and but Black would have no answer
then further ways for White to in the long term to White's strat­
strengthen his position will be­ egy outlined above.
come clear. 34 .tdl !
30 �b8 The bishop now comes to g4
31 .t e2 f4 with even greater force. Black
Note that the f-pawn is far must lose the d-pawn, for exam­
more vulnerable to attack than the ple: 34 . . . tbg6 35 .tg4 and e5-e6,
d5 pawn, since frontal pressure or 34 . . .:h7 35 .tf3 l::td 7 36 .tg4
can be applied by White's rooks. (Bareev).
Now White has a logical plan to 34 tbh7
exploit the weaknesses in Black's 35 .tg4 "iif7
kingside: 36 as?
i) Double rooks on the f-file, Simply 36 e6 wins the d-pawn,
which will blunt any attempt by e.g. 36 . . :iWc7 37 'i'xd5 tbf6 38
Black to start a kingside attack 'iff5 ! ; Kamsky may have been
with . . . g5-g4 and . . . g4-g3 . put off by 38 'iig 2 ? 1:txh2 ! ! when
ii) Play the bishop to g4, and 39 'itxh2 ? f3+ wins the queen.
chase the queen from e6. 36 ktd8
iii) Then advance eS-e6 ! at an A respite for Black, but his
appropriate moment followed by position remains unpleasant.
lIe 1 and l1e5 . 37 b6 a6
iv) And finally win the d-pawn 38 :tel !
or g-pawn through the combined Since White did not take ad­
pressure of rook, bishop and vantage of Black's blunder at
queen. move 33, the game continues in
It is now clear that Black's the pattern outlined at move 3 1 .
kingside pawns had two blockad­ 3 8 %:te l begins part (iii).
ing functions. Not only was it 38 'i'e7
their duty' to form a barrier to
• 39 .tf5 tbf6
White' s rooks, but they were also Black cannot prevent White
required to shut out the white from carrying out his plan. Here,
bishop from control of e6. They as Bareev points out, White
have failed in the second of these should play 40 Itfe2 tbe8 4 1 e6.
tasks. This in tum means that Sooner or later, White can win
Black' s queen will be driven the d-pawn with lIe5 and .tg4-f3 .
from e6. and that the whole Instead, in time. pressure. White
70 Positional Sacrifices

played
40 :en?
and agreed to a draw. Black
can try 40. . . lbhS aiming to get his
knight to e6, e.g. 4 1 :e l
(correcting his mistake) 4 1 . . .lbg7
42 .tg4 lbe6. Black has blocked
the hole in his position. White
could try 40 . . . lbhS 4 1 e6! ? lbg7
42 'iWg4 c!tJxe6 43 :e l :h6 44
:fe2 :d6, but Black can proba­
bly weather the storm.
Although strong players nowa­
days are aware of all the standard 31 h6
tactical devices, a sacrifice can His idea is to continue
still surprise even the most battle­ 32 . . . i.e4, blocking the e-file, and
hardened grandmaster. Yuri Ra­ then 33 . . . �h7 followed by
zuvaev is one of the most solid of 34 . . . I:tag8, bringing all his pieces
all Russian grandmasters. He was to bear against g2. If g2 drops,
selected to play board 8 in the then White's position will col­
USSR-Rest of the World match lapse. Black played 3 1 . . .h6 fIrst
in London in 1 984, ahead of so that after . . . .te4 he is not
many illustrious Soviet grand­ troubled by c!tJgS. However, he
masters, since he could draw with had overlooked White's next
anyone. He duly obliged - four move, which crosses his plan and
draws against Hiibner! Yet for all leaves him in a wretched posi­
his experience, in the following tion. He should have played
game he misses a strong sacrifice 3 1 . . . .te4 immediately when, after
that costs him the game. 32 lbgS dS 33 c!tJxe4 fxe4, an in­
teresting position is reached in
which Black has the superior
Ki n derm a n n-Razuvaev
long-term pawn structure (a pro­
Prague 1 992
tected passed pawn on e4) but his
king is somewhat exposed. Evi­
A tense position. White has con­ dently, Razuvaev was looking for
trol of the e-file and a more com­ a more advantageous outcome to
pact pawn structure around his the struggle, but he overestimated
king. Black on the other hand is his position.
exerting enormous pressure 32 dS! .txdS
against the g2 square both diago­ Of course, if Black had played
nally and frontally. Razuvaev 3 1 . . . .te4 last move, he could
played: simply ignore 32 dS and play
 Sacrifices to Open Lines 71

32 . . . h6. 43 llxf5 b4 44 ltb5 bxa3 45

33 �h4! %lg7 bxa3 :n 46 g3 :'c7 47 ':b2 lld7
34 llg3 48 rj;f2 c3 49 :te2 'iitf6 50 Cite3
Now White's compensation for 1-0
the pawn sacrifice becomes clear.
The square d4 has been vacated, In the next example, Black can
so White is threatening 35 1:hg7+ be forgiven for missing White's
r3;xg7 (35 . . . 'ii'xg7 36 Ibd5) 36 clever idea:
"if'd4+ winning a piece. Mean­
while, 35 �xf5 is also threatened,
She rbakov-Karlsson
when Black's kingside is ripped
Taby 1 99 1
apart.
34 Jte4
The only move. If White's d­ 1 d 4 e6 2 c 4 f5 3 �c3 �f6 4 �f3
pawn were still on d4, White's Jt b4 5 .tg5 0-0 6 llel d6 7 g3
attack would now be at an end. lbbd7 8 Jtg2 'ii'e8 9 Jtxf6 �xf6
But as it is, his queen's rook 1 0 0-0 .txc3 11 llxc3 e5 12 dxe5
makes a powerful entrance, at the dxe5
same time restoring material
equality.
35 llxd6
Yes. 32 d5 ! was a good move.
Ask White's queen's rook!
35 �h7
Or the h-pawn would be taken
as well. White is now dominant
on the dark squares, which allow
him to penetrate into Black's
position.
36 llxg7+ 'ii' x g7
37 'ifb6!
Threatening 38 llg6 'ii'f8 39 13 e4! !
'ii'c 7+ 'iith 8 40 .l:.g3 ! followed by An incredible move, giving up
4 1 �g6+. the e-pawn 'for nothing' .
37 'fia7 13 �xe4
Managing to exchange queens It is hard to refuse such a gift,
and thereby avoid the direct at­ but 1 3 . . . f4 ! was better, when
tack, but at the price of a losing White's bishop remains shut out
endgame. of the game. However, this is
38 lbh6+ 'iitg 7 39 :g6+ �n only apparent in hindsight.
(39 . . .rii) h7 40 l1d6) 40 llf6+ 'iitg7 14 lle3 c6
41 �xf5+ Jt xf5 42 'ii'xa7 + l:.xa7 Ruling out 1 5 'ii'd 5+ winning
72 Positional Sacrifices

the e-pawn. How is White to rificing his pawn, White regains

continue his attack now? it and eventually receives the eS
15 b3! ! pawn as interest.
This i s White's idea. H e will 21 �xf5
play 16 �a l and win back the e­ 22 !ii. xe4 axb3
pawn. Note that if White re­ 23 axb3 !ii. xe4
establishes material equality, he 24 l1xe4 :al
will have a good game: the black 25 lbal li'xal +
knight can easily be ousted from 26 �g2 lU8
e4, and the weakness in Black's 27 'i'e2 'iib l
centre caused by . . . f7-fS will re­ 27 . . J�e8 28 f4 wins the e­
main. Karlsson is not so obliging, pawn, and 27 . . . :f5 28 f4 exf4?
however. 29 l:te8+ leads to mate.
15 �e7 28 lhe5 h6
16 'iWal :te8 Black finds to his consternation
17 life1 �c5 that he is mated after 28 . . . 'ilxb3
After 17 . . . !ii. d7 1 8 lDh4 fol­ 29 :e8 'i'b4 30 'iie6+ 'it>h8 3 1
lowed by 1 9 lDxf5 and 20 �xe4, 'ii'f7 ! . Therefore he cannot regain
White has regained his pawn and his pawn.
e5 is kept terminally weak 29 'ike3 l:td8
( 1 8 . . . g6? 1 9 f3 or 1 8 . . . lDf6 1 9 30 �e7 :dl
!!xeS are both unsatisfactory for A desperate counter-attack that
Black). So Black sets a crafty accelerates the end.
trap. If 1 8 lDh4 then 1 8 . . . lDxf2 ! 31 'iVe6+ r;th7
and 1 9 'it>xf2? f4 or 1 9 fxe5?? 32 l1e8 J::tg l +
lDh3+ both lose for White. After 33 'it;>h3 1-0
1 9 b4 ! 'iYxb4 20 �xf2 e4 fol­ Black is soon mated after
lowed by . . . !ii.e 6, Black has three 33 . . . 'ii'f 1 + 34 'it>h4 g5+ 35 'it;>hS
pawns and dynamic play for his 'iVd l + 36 g4.
piece. Sherbakov finds a simple
way to keep control : Tarrasch once wryly remarked
18 'iWb2! 'Chess is a terrible game. If you
Defending f2. Now the plan of have no centre, your opponent
lDh4 leads to a positional advan­ has a freer position. If you do
tage and no complications. have a centre, then you really
18 as have something to worry about! '
19 lDh4 �d4 The Australian grandmaster
20 'iic2 a4 Rogers finds himself trapped in
After 20 . . . lDd6, the retreat 2 1 this paradox in the next game. He
lDf3 is simple and strong. builds a centre in order to deprive
21 lDxf5 his opponent's pieces of open
A t l �st . Fi !!h t m oves after Sll C'- lines. only to find it decimated by
 Sacrifices to Open Lines 73

some flanking blows from when it is kicked out by . . . g7 -g6,

White's restricted army. a slight weakness appears in
Black's kingside.
18 "ifd7
Rosenta l ls-Rogers
Stopping liJf5 and also with
M a l m o 1 993
designs against the d3 pawn. A
g ood alternative was 1 8 . . . .i.c8 !
1 e4 eS 2 c3 liJf6 3 eS liJdS 4 g3 redeploying the bishop to e6
d6 5 exd6 e6 6 .iog2 .i.xd6 7 liJo where it strengthens Black's in­
liJe6 8 0-0 0-0 9 liJa3 .i.e7 10 d3 fluence on white squares such as
b6 11 liJe4 .i.b7 12 a4 "fIie7 13 e6, f7 and f5 which have been
'Wi'e2 :ad8 14 .iod2 l'.tfe8 15 :Sel neglected by the pawn advances
.i.f8 . . . f7-f6 and . . . e6-e5.
19 "ife2
With vague ideas of 20 .i.e4
and 2 1 d4 ! ? attacking h7 . The
queen also moves from the e-file
in preparation for f4.

Both sides have played the

opening somewhat cautiously and
are now well entrenched behind
their own lines. White's next
move provokes Black into ad­
vancing his centre pawns. 19 gS?
16 .i.gS f6 One pawn advance too many.
17 .i.c1 eS Black could have countered the
Now White can claim that idea of 20 .i.e4 with 1 9 . . . liJce7 ! ,
Black has loosened his pawn then 2 0 .i.e4 g6 2 1 liJg2
front, while Black can point to his (preparing f2-f4) 2 1 . ..f5 ! 22
space advantage in the centre. liJxe5 (22 .i.f3 liJb4 ! ) 22 . . . 'iic7
18 liJh4 23 .iof3 .iog7 24 liJc4 liJb4 ! 25
With the idea of an eventual f4, cxb4 (25 "ifd l lbd3) 25 . . . ..txf3
undennining Black's centre. The gives Black excellent play for the
knight itself may also move to f5 ; pawn. The light -squared bishop is
74 Positional Sacrifices

so menacing to White's king that tion:

he will have great difficulty sur­
viving an attack after . . . lDc6-d4
or . . 'if c6 and . . . lDdS .
.

White of course can (and

should) avoid this variation, say
with 20 lDf3, but he is making no
progress at all. After 19 . . g5, on
.

the other hand, he has a clear tar­

get, the flimsy e5/f6/g5 structure,
which is ripe for demolition with
a timely f2-f4.
20 lDf3 lDc7
21 �fd2! ?
White decides that firm meas­ i) The potential pressure which
ures are called for, and so he sac­ White's bishop on g2 exerts
rifices his d-pawn to speed up his along the diagonal h l -a8 is very
attack on Black's centre. The unpleasant for Black. This pres­
conservative 2 1 lId l is ineffec­ sure becomes important in a
tual, since after 2 1 . . .lDe6, the ad­ number of variations; for exam­
vance d3-d4 is hard to achieve ple, see the note at move 23. Put
and White is therefore left with­ the black bishop on a8, a de­
out a plan. fended square, and Black's de­
21 'ifxd3! fensive task is eased. Black how­
White may have some useful ever is never allowed the lUXUry
positional trumps after this move, of a free tempo to play . . . ..ta8 in
but a pawn is always a pawn. We the game.
have seen some beautiful exam­ ii) White's last move dyna­
ples of sacrificial play in this mites Black's fragile kingside
book, but it should never be for­ pawn structure. Despite the ex­
gotten that (to misquote Tartak­ change of queens, the position
ower) it is normally better to sac­ still has the character of a mid­
rifice the opponent 's pieces. dlegame rather than an endgame:
22 'ifxd3 :xd3 all the other pieces are still pres­
23 ttJe4 rt;g7 ent. This means that king safety is
Not 23 . . . ..tg7 24 Ned6 :b8 25 still a priority. After the black
lDxb7 and wins the knight on c6. bulwark eS/f6/g5 is dissolved,
24 f4! Black will face an attack from all
Finally, White has achieved the White 's pieces. Most of Black's
long desired advance. Black has minor pieces, on the other hand,
an extra pawn, but White has are on the queenside and cannot
more than adequate compensa- easily return to the defence of the
 Sacrifices to Open Lines 75

king. (This is another reason why cessful positional chess if you

the bishop on g2 is doing such an have a sharp eye for tactics.
excellent job: it pins down pieces
which would otherwise rush to
the kingside.)
iii) White has a clear plan of
attack and the means to carry out
this plan. Black, on the other
hand, has an onerous defensive
task and can only respond to one
move threats.
Obviously, Rogers did not rel­
ish defending his position and
fails to find the line which
minimises White's advantage.
24 exf4 It seems that Black has no
25 gxf4 h6? good moves after 28 ltf6. Alter­
Black should play 25 . . . g4 ! re­ natives to 28 . . . ttJd8 are no better:
turning the extra pawn to keep 28 . . . h3 ! ? 29 iLh l ! (better than 29
the kingside blocked. Then after iLfl ':g3+ 30 ttJxg3 <it>xf6 ! )
26 lbf2 'udd8 27 lbxg4 .ta8 ! 2 9 . . . l:Ig3+ 3 0 <it>f2 ltg6 3 1 11xg6+
(Black now has the free tempo �xg6 32 l1g 1 + and wins the ex­
needed for this useful move) 28 change by a knight fork on d6 or
ttJce3 ttJe7 ! (defending f5) Black f6. If instead 29 . . . iLe7 then 30
has almost equalised. Rogers iLxh6+ <it>g8 3 1 <it>h2 and 32
holds on to his extra pawn, but J:Ig 1 + is a winning threat since
seems to forget that it is still 3 1 . . . .txf6 32 ttJxf6+ <it>f7 33
possible to be mated by a direct ttJxe8 lbxe8 34 iLxc6 iLxc6 25
attack even after the queens have ttJe5+ wins.
been exchanged. Alternatively, Black could try
26 fxgS fxgS 28 . . . iLe7 but 29 iLxh6+ <it>g8 30
27 h4! :efl .txf6 3 1 lbxf6+ <it>f7 32
The next stage in the destruc- ttJd5+ �g6 33 ltf6+ <it>h5 34
tion of Black' s kingside. ttJf4+ �g4 35 ttJxd3 wins.
27 gxh4 29 iLxh6+ <it>g8
28 �f6 ! (D) 30 iLxfS �xfS
28 lbd8 31 l:g6+ �h8
Rosentalis points out that 32 ttJeS
28 . . Jhe4, hoping for 29 iLxe4 Now a deadly swarm of white
lIg3+ winning a piece, loses out pieces descend on Black's king.
to 29 J:tg6+ ! ! and White wins. We 32 lIdS
repeat: you can only play suc- 33 ttJ�S llfS
76 Positional Sacrifices

The rook on d5 cannot retreat: semi-open file on the queenside.

33 . . . 11d2 34 .ll xb7 tZ'lxb7 35 If White doubles rooks on the b­
tZ'let7+ wins. So White's pressure file, Black simply plays . . . .ll c6
on the long diagonal finally wins (not however . . . b7-b6? which
the exchange. would expose the pawns to attack
34 .ll xd5 .ll x d5 after a3-a4-a5). The situation is
35 c4! �a8 almost as blocked on the king­
36 �d1 1-0 side. The knight on t7, which
If the knight on d8 moves, 37 shields the f-file, also keeps the
tZ'let7+ wins, and 36 . . . 11e8 is met rook out of h8.
by 37 �xd8 . The knight evidently holds to­
The reader should not imagine gether Black's position. So why
that only pawns can effect a not eliminate it with 33 tZ'lxg5?
breakthrough, although they are The problem is that 33 . . . tZ'lxg5 34
of course the most common sac­ l:ixg5 g6 ! leaves the rook trapped
rificial choice - a plentiful and on g5 . If 35 l:tfl �g7 36 :f6 .ll t7
cheap supply of cannon fodder. (but not 36 . . . �h6? 37 Iih5+) and
Next we see how Kamsky solves the bishop moves between t7 and
the problem of penetrating a well­ e8 until White agrees a draw, or
entrenched and apparently invin­ Black can try . . J�c7-c6-a6 and
cible defensive line. attack the a-pawn.
In such a blocked position, a
knight can be a more useful piece
Kamsky-Yusu pov
than a rook. A rook must stop
Linares 1 993
when it reaches a brick wall; a
knight can leap over it. Therefore,
White must find a favourable
moment to play lhg5 ! , returning
the exchange but remaining with
an agile knight against a slightly
restricted bishop. However, it is
hard to imagine that White can
win against good defensive play.
33 l::t b h1
Kamsky decides his best prac­
tical winning chance is to probe
the position. Perhaps he will dis­
cover a way to win, or maybe
White is the exchange up, and Black will lose patience and
at first glance seems to have an weaken himself somehow.
easy win. But how exactly is he to 33 :c7!
break through? There is only one Here for example Black could
 Sacrifices to Open Lines 77

go wrong with 33 . . . lbh6, when 34 tempo. If his king were still on f4,
.l::r.xh6 ! gxh6 35 l:Ixh6 wins - Black could play 54 . . . �g6 draw­
Black loses his e-pawn or his g­ ing. Can Black gain a tempo and
pawn; if 35 . . . 'iiil g7 36 ltf6. draw?
Yusupov, whose nickname is the
Russian Wall, is not to be per­
turbed. He defends quietly.
34 %ih7 :c6
35 lUhS

(analysis diagram)

Yes ! At move 40 (see diagram

above) he can play 40 . . . 'iiile7 ! not
giving White the opportunity to
35 :c8 play 4 1 ltf6 with check. Now 4 1
In his notes in In/ormator 5 7, :f6 :lxa3 42 g5 l: a 1 4 3 g6 llg 1
Kamsky gives this move an ex­ 44 :f7+ rJi;e8 draws, as does 4 1
clamation-mark and says it is the :g7+ <it>f8 42 :xb7 :'xa3.
only move. But what happens So Black could have forced a
after 35 . . . %:.a6 instead? White draw with 35 . . .lIa6 ! . However, it
must carry out his threat of 36 was not easy to see this during
lbg5 4.Jxg5 37 l:th8+ rJi;f7 38 the game - indeed, Kamsky did
Ihg5 g6 ! 39 ':'xe8 (39 lth7+ �f8 not see it in his post-game analy­
40 :xb7 ..tf7! followed by sis. I suspect Yusupov did not
. . . :xa3. The rook on g5 is im­ think he was in any danger if he
mured, so Black draws easily.) defended passively; and to a cer­
39 . . . 'lfo>xe8 40 Ihg6 (D) tain extent, he was right. How­
..40 . �f7 41 :f6+ �e7 ever, this lack of vigilance even­
(4 1 . . . 'ik;g7 ! ?) 42 g5 :xa3 43 g6 tually proves his undoing.
::ta l 44 nf7+ �e8 45 ':'xb7 :tg l 36 'iiile3 as
46 g7 a5 47 :'a7 a4 48 �e3 a3 49 It was still possible to play
rJi;f4 a2 50 ':'xa2 ktxg7 5 1 l:Ia8+ 36 . . . l1c6. However, White's king
rJi;f7 52 l:.a7+ 'ik;f8 53 ':'xg7 �xg7 is one square up the board if
54 'ik;g5 and White wins - bv one White goes in for the variation
78 Positional Sacrifices

with 37 lDxgS, etc . , as above. better, or indeed anywhere else

This could prove crucial. on the second rank. The reason
Yusupov prefers to strengthen the why becomes apparent as the
blockade on the queenside, and game progresses.
evacuate a pawn from the second 43 <i&i>g3 lh6? !
rank. This proves useful in a The rook could still return to
variation such as 37 lDxgS lDxgS the second rank, and no serious
38 l:h8+ rJ;f7 39 l:xgS g6 40 damage would be done.
llh7-t- �g8 4 1 l:xb7 i.. f7 fol­ 44 l:txg5! lDxg5
lowed by . . . �g7, as given by 45 c!Dxg5 i.. g6
Kamsky. White's gS rook is in If the rook were on the second
the now familiar trap. The draw­ rank, Black could play 4S . . . :e7
back to 36 . . . aS is that Black per­ here and there would be little to
manently loses the option of fear. His rook could defend the e­
. . . l:a6 attacking the a-pawn. pawn and be in contact with the
37 %thl a4 kingside. On a6, on the other
Not 37 . . . i.. a4 38 lDxgS ! i.. xc2 hand, although the rook defends
39 lth7-h2. the e-pawn, it is out of touch with
38 lt7h2 b5 the kingside. This gives White
Now Black has fixed the the chance of a breakthrough.
queenside. White's rooks or king 46 :f2 %lb6
cannot hope to break through 47 <i&i>h4 l:a6
there. However, more and more 48 lDh3
Black pawns are ending up on
white squares. This restricts the
bishop on e8. So it is time for
White to consider playing ltxgS !
to force a favourable material
balance: a rook and good knight
against a rook and bad bishop.
39 %tf2
White probes a little while
longer . . .
39 ltc7
40 l:h5 :e7
41 :f1 :b7
42 <t>f2 l:tb6? ! 48 i.. e4?
and Black carelessly plays Black's first serious mistake.
his rook to an inferior square. He had two drawing methods:
Obviously, his sense of danger a) 48 . . . l:ta7 49 lDf4 l:tf7 SO
has been dulled by White's end­ �g3 ':'xf4 ! ! S I <itxf4 i..e 8 and I
less manoeuvres. 42 . . . %tc7 was can ' t see how White breaks
 Sacrifices to Open Lines 79

through. lowed by 5 1 lIf7 infiltrating.

b) 48 . . J::ta7 49 lLlf4 i.. f7 50 g5 50 lLlg5 kte7
planning, after 50 . . . lte7, either 5 1 If 50 . . . i.. h7 5 1 lLlxe6 wins.
g6, 52 �g5, 53 lih2, 54 :h7, 5 5 51 lLlxe4 d xe4
lLlh5 and then a spectacular 52 W g6
breakthrough with 56 lLlf6+! ?, or White now has a winning end­
5 1 g6, 52 �h5, 53 lLlh3 , 54 lLlg 5 , game. The finish is instructive:
55 lLlh7, hoping for a mate o n f8, 52 e3!
or if Black prevents it with . . . l:e8, The path of maximum resis-
a breakthrough on the seventh tance. Black activates his rook.
rank with l:tf7. 53 :e2 lif7
The common denominator in 54 lixe3 li1f2
these plans is 5 1 g6, so Black 55 :'e1 !
must play the ugly move 50 . . . g6 ! Finding the breakthrough plan.
himself. Now Black's bishop is 55 l1xc2
wretched, but it is still doing 56 �b1 lbc3
good defensive work, overpro­ 57 lbb5
tecting e6 and g6 and ready to Material is now even, but
defend the b-pawn with . . . ..te8 if White's king dominates his
necessary. Black's blockade counterpart.
holds, e.g. 5 1 lLlg2 �g7 52 lLle3 57 WfS
l:te8 53 lLlg4 ':e7 and now 54 58 l'lb4 lIxa3
lLlh6 i.. e 8 55 1:.f6 i.. d7 etc., or 54 59 :'xc4 ':a1
lLlf6 :c7 55 �g4 l:c8 56 ':h2 60 g5 a3
l::th8 ! (or 56 . . . i..g 8) stopping 61 ':a4 a2
White's plan of l:.h7+. 62 ':a7 �e8
White's problem after 50 . g6 ! .. No better is 62 .. J�d l 63 l:Ixa2
is that there are no breakthrough lhd4 64 lta8+ We7 65 Wxg7 and
squares for his king - the position the g-pawn will run through.
is simply too blocked. If you re­ 63 �xg7 �d8
move White's e5 pawn then 64 g6 We8
White, despite being a pawn 65 lta6 rii;e7
down, can win by <;i;f4 and 'it>e5, 66 1:.a8 �d7
penetrating with his king. 67 �g8 �e7
49 �h5! 68 l1a7+ 'it>e8
The black bishop suddenly 69 g7 l:thl
finds it cannot get back to g6 to Or else White plays 70 <;i;h7
cover f7. What' s more, the bishop l:.h l + 7 1 �g6 :g l + 72 <;i;f6 and
is a target on e4. The blockade �xe6.
begins to crumble. 70 l::t xa2
49 .:a7 Now if all the centre pawns -
If 49 . . . l:tb6 then 50 lLlg5 fol- on d4, e5 and e6 - are removed,
80 Positional Sacrifices

White has a standard book win lieS l:hl 76 :a7+ �e8 77 e6

with 7 1 :e2+ <it'd7 72 :e4 ! fol­ llh2 78 :t7! : h l 79 e7! :h2 80
lowed by bringing out the king lU8+ �e7 81 :f3 �e8 82 :e3+
and sheltering it from checks with �d7 83 :e4 1-0
the rook (the so-called Lucena White is ready to play 84 <it'f7
position) . So White has to get rid lH2+ 8 5 <it'g6 :g2+ 86 �f6 1U2+
of the centre pawns. 87 �g5 :g2+ 88 :g4.
70 �e7 71 :a7+ �e8 72 ltaS
..• A gritty display by Kamsky.
�e7 73 dS exdS 74 l:1xdS :h2 75 Top-class chess is not for the
fainthearted !
 6 The I n d i a n B i s h o p

The key feature of many modern 10 �xc6 bxc6

opening systems (for example, 11 eS
the King's Indian, Griinfeld, and White, for his part, wants the
Sicilian Dragon) is the fianchetto black bishop kept under lock and
of Black's king's bishop on g7 . key. He plans to follow up with
From g 7 the bishop exerts pres­ 12 f4 (say after 1 1 . . .�eS) when
sure on White's centre, and Black can only activate his
(assuming Black has castled bishop with a subsequent . . . f7-f6,
kingside) also has an important which will leave him with a com­
defensive function. Hence promised pawn structure after
Black's strategy usually revolves exf6 iLxf6.
around this piece. We shall look 11 �e4!
at some examples where, at slight This fine move decides the
material cost, Black succeeded in battle for the bishop's future in
activating his ' Indian' bishop. Black's favour. If White plays 1 2
f4 then 1 2 . . . lLIxc3 shatters his
queenside pawn structure, so he
Adams-Khallfman
must exchange knights. This,
Las Palmas 1 993
however, means that the e-pawn
becomes very hard to defend,
1 e4 cS 2 ltJf3 d6 3 d4 cxd4 4 since if f2-f4, Black captures en
ltJxd4 ltJf6 5 ltJc3 g6 6 .te2 .tg7 passant.
7 0-0 0-0 8 iLe3 ltJc6 9 �hl dS! 12 �xe4 dxe4
White has played an innocuous 13 'i'xd8 l:lxd8
system against the Dragon and 14 lIfdl .te6
Black has thoughts of taking the 14 . . . :'xd l + 15 lIxd l iLxe5 16
initiative, exploiting the power of l:.dS+ 'i;g7 would be very dan­
his king ' s bishop. For example, if gerous for Black. For instance,
10 exd5 ltJxd5 1 1 lLIxc6 bxc6 1 2 White could play 17 .ia6 i.b7 1 8
lLIxd5 cxd5 and the a l -hS diago­ lhaS iLxaS 19 .txa7 when the a­
nal is swept clear (if 1 3 .td4 e5) . pawn would be difficult to stop.
82 Positional Sacrifices

15 .td4 1 8 . . . :d5 19 l:.a3 i..c 2? 20 i.. c4.

White is reduced to artificial Black's best line is probably
methods to defend the e-pawn. 1 8 . . . c5 19 e6+ ! ? i.. xe6 20 i.. xc5
The aggressive 1 5 lhd8+ llxd8 J:d2 2 1 i.. a3 i.. b3 with pressure.
1 6 .txa7 looks suspicious after 17 lhd4!
1 6 . . . .txe5 (and 1 7 . . . .txb2) or Our positional sacrifice. The
1 6 . . J:td2. Indian bishop will finally come to
15 f5! life.
16 a4? 18 I;[xd4 l1b8!
White thinks he has solved the Adams may have allowed the
problem of Black's g7 bishop and sacrifice because he missed or
so plays to gain space on the underestimated the strength of
queenside. His ideas include ad­ this move. After 1 8 . . . i.. xe5? 1 9
vancing the a-pawn to a6 and llb4 White would have fair
then playing the rook to a5 and c5 chances: his queenside pawn
in order to intensify the pressure structure remains intact and his
on Black's split queenside pawns. rook is well placed. On the other
However, he had overlooked hand, after 1 8 . . . l:.b8, not only is
Black's plan. As Khalifman White's rook denied activity on
points out, he should play 1 6 exf6 the b-file, but even worse, he
.li.xf6 1 7 i.. xf6 exf6 with equal­ cannot prevent Black playing
ity. . . .lhb2, winning a pawn and
16 fragmenting the remammg
queenside pawns. However, as
will be seen, all is not yet lost for
White.

17 as?
1 7 c3 ! was better, bolstering
the bishop on d4. Then 17 . . . i..b 3
18 lie l (not 18 �d2 c5) and if 17 :'xd4 is a typical posi­
. . .

1 8 .'�te6?
. . 19 :a3 ! �db8 tional or long-term sacrifice, yet
( l 9 . .li.d5 20 c4) 20 �xb3 ! , or
. . it depends for its validity on the
 The Indian Bishap 83

'tactical ' move I S . . . ltbS . If Black's chances after 20 l1a4 in

White's rook on a l were on any the second variation. The ex­
other first-rank square - say e 1 - change of Black's remaining rook
then I S . . . .l::t b S could be answered is inevitable after both 20 . . Jhb2
by the simple 1 9 b3. Only in the or 20 . . . :d8 ! ? 2 1 l:td l . It is a well­
game position with the rook on known principle that when you
a l is 1 9 b3 impossible because are the exchange up in an end­
1 9 . . .ixe5 wins the exchange by
. game you should try to force off
skewering the two rooks. the opponent's remaining rook,
So was Khalifman lucky? Is it so that he has no piece left which
a sheer fluke that the rook was on can act at long range across the
a 1 and not f1 or c I ? The fact is board. After 20 . . . ':'xb2 2 1 l1xb2
that the dynamics of the struggle .ixb2 22 �g 1 , White has good
never gave White the tempo he survival chances. But note that
needed to play lin or llcl. There the alternative 22 ltb4 .ic3 23
was always something else that l:lb7 .ixa5 24 l1xa7 is not satis­
he was required to do. Tartak­ factory for White. As has just
ower defined chess as 'the trag­ been remarked, the rook acts at
edy of one tempo' , while Bron­ greater range. Therefore it helps
stein talked of 'the most powerful Black to exchange off the pawns
weapon in chess - the next at the periphery of the board,
move! ' The time element in chess since he is less likely to be over­
- the 'rhythm ' of the game, as it stretched. 22 .ic4 .id5 ! is also
were - is so important that it not good.
should not be surprising that the Black's winning hope is his
most positional of sacrifices impressive mass of pawns in the
should stand or fall on whose tum centre, but its hard to believe that
to move it is. White should lose if he defends
19 f4? well. White can also generate
Now in order to win the e­ counterplay with 22 a6, 23 :1b4
pawn, Black must exchange off and 24 :b7. So the assessment
his own strong e4 pawn, and al­ 'slight advantage to Black' seems
low the white bishop some activ­ more appropriate. Of course, this
ity. If 1 9 .ic4 .ixe5 20 .ixe6+ does not devalue the sacrifice,
�xe6 2 1 !:tc4 'itt d 5 is hopeless, or which was undoubtedly correct. It
1 9 l:[b 1 .ixe5 20 :a4 lbb2 does mean, however, that a ques­
(20 . . . �d8 ! ?) and Black has a clear tion mark must be added to the
advantage. These variations are move 1 9 f4.
given by Khalifman in Infarma­ 19 exf3
tar 58. 20 .ixf3 .ixe5
One wonders if Khalifman is 21 !:td3
being somewhat optimistic about Or 2 1 l1d2 lixb2 22 .ixc6 .ic3
84 Positional Sacrifices

winning the c-pawn, e. g . 23 lIe2 to shepherd home the passed

.ic4 24 :Lf2 .id4 25 %:tf4 e5 . pawn. However, in keeping the
21 :xb2 king out, they have allowed a
The difference between the blow from rook and bishop :
g ame continuation and the varia­ 44 l:b3+
tions examined at move 1 9 is that 45 <iti>e2 l:b2+
Black does not have to allow the 46 l:d2
exchange of his rook in capturing If 46 �d3 :txh2 followed
this pawn. The rook is dominant by . . J:tg2, and Black will shortly
on the seventh rank, so the differ­ have five passed pawns.
ence decides whether Black has a 46 .tb4
small advantage or a decisive 0 1
-

one. The pawn will queen.

22 l:el .td6
23 .ixe6 lbe2
Santo-Roman-M.Gurevlch
24 .idS .txd5
Clichy 1 993
25 %lxd5 :&2!
This wins the a-pawn.
26 g3 .ib4 1 e4 d6 2 d4 liJf6 3 liJe3 g6 4
27 :bl %baS .tg5 e6 5 'iVd2 .ig7 6 liJf3 0-0 7
2S :d7 h3 liJbd7
28 %ba5 .ixa5 is equally Here rather than develop (8
hopeless. The outside a-pawn .te2) or restrain Black on the
would be unstoppable. queenside (8 a4) , White tried to
2S �e6 29 %ldS .td6 30 :hS
••• exploit Black's last move with
h5 31 Af8 l:le5 32 %lb2 as 33 S e5! ?
l:la2 .tb4 34 �g2 �d5 35 l:e2 since normally Black plays
a4 36 l:gS l:e6 37 l:dS+ .td6 3S 7 . b5 and answers 8 e5 with the
. .

l:d2+ �e6 39 :as a3 40 Cittf3 counter-attacking g . . M ! .

.

l:e4 41 l:gS l:b4!

42 l:eS+
There was not time to capture
the g -pawn. Khalifman gives 42
:txg6 l:b3+ 43 Citte2 l:b2 44 :g 8
.tb4 ! .
42 Iit>b5
43 l:d5+ c;t>b6
44 :as
The white rooks have fought
valiantly to prevent Black' s plan
of bringing his king to b3 in order
 The Indian Bishop 85

After 8 ... tDe8 9 0-0-0 White threat of . . . .txb2 . White cannot

has the makings of a big attack play 14 0-0-0 without facing a
with h4-h5 . The white e5 pawn is huge attack after 14 . . . 'ib6 1 5 c3
immune after 9 . . . dxe5 10 dxe5 .te6 and . . . l:.c8 (Gurevich) . Sac­
because of the pin on the d-file, rificial ideas against c3 would
and Black' s queen cannot easily flow naturally from such a posi­
move from d8 without allowing tion. Also, . . . d5-d4 is a strong
JJ.. xe7 . Furthermore, 9 . . .f6 cannot threat. In general, it is very dou­
be played without weakening the ble-edged to castle queenside into
central pawn structure . And fi­ the line of fire of a black bishop
nally 9 . . b5 1 0 h4 b4 1 1 tDe4
. on the a l -h8 diagonal; unless of
(attacking d6) 1 1 . . .d5 1 2 tD g3 course you have strong defences
'ilia5 1 3 �b l leaves the bishop on along this diagonal and have a
g 7 shut out of the game and rapid kingside attack as compen­
White ready to play 1 4 h5 . If sation. In the present example,
1 3 . . . h5 ?, then 1 4 e6 ! fxe6 15 .td3 White has neither strong defences
�h7 16 tDxh5 . Evidently, nor a kingside attack. Therefore,
Gurevich did not like the look of 1 4 0-0-0 would be particularly
these variations and found a dy­ foolhardy, and White has to pre­
namic alternative involving an pare kingside castling .
exchange sacrifice:
8 dxeS!
9 dxeS tDdS!
10 tDxdS cxdS
11 .th6
As Gurevich points out, 1 1
'iVxd5 ? tDxe5 1 2 'iVxd8 liJxf3+ 1 3
gxf3 ':xd8 is bad for White. His
initiative is gone and he is left
with weak kingside pawns. In­
stead, Santo-Roman defends his
e-pawn by indirect means. He
hopes to start a strong attack after
1 1 . . .e6 1 2 h4, etc. But now we 14 c3 'fIe7!
see the point of Black's play : Gurevich's best move of the
11 liJxeS! game. His plan is to eventually
Crossing White ' s plans. play . . . d5-d4 increasing the scope
12 ltJxeS .txeS of his e5 bishop. The best square
13 .txf8 ..tt xfS (D) for the bishop on c8 is therefore
Black has seized the initiative. c6 where it will control an open
His bishop on e5 is tremendously diagonal after the d-pawn ad­
strong and there is an immediate vances . It will also aim at the g2
86 Positional Sacrifices

square which could prove un­ nent of a plan, not usually a very
pleasant for White's king. After rewarding strategy.
14 . . JWc7 ! the d-pawn is immune 17 f4?! �f6
for tactical reasons: IS 'tWxdS? 18 }lad 1
�xc3+! , etc. After 1 8 fS gS ! the f-file re­
15 �e2 i.d7 mains closed and White is left
16 0-0 i.c6 very weak on the b8-h2 diagonal.
Now that White has completed He would have to watch out for a
his development, he has to form a sudden . . :i'g3 and . . . i.eS with a
plan. The reader will be familiar winning attack. So all 1 7 f4 has
with the dictum that rooks need done is weaken White's kingside.
open lines. Therefore, White 18 1;g7
must try to puncture a hole 19 �f3 e6
somewhere in Black's solid wall 20 g3
of pawns. But how? If he pre­ Besides further weakening the
pares and carries out the advance kingside, this allows a tactical
c4, then Black can reply . . . dS-d4 blow which leads to a fragmented
with a powerful passed pawn, and queenside. Gurevich recommends
no open lines anyway. Or if he 20 a3 i.a4 2 1 l:1de 1 .f.1d8 22 c;£th 1 .
doubles rook on the e-file, Black Black has the initiative but he
simply plays . . . e7-e6 and there is must still find a way to break
no !>reakthrough. So Santo­ through without allowing coun­
Roman tries using his f-pawn, but terplay.
this proves bad as well. 20 'i'a5!
Since all active plans fail, 21 a3 i.a4!
White should probably wait and 22 l:tc1 d4
see what Black comes up with.
For example, if Black decides on
an eventual . . . e7-eS and . . . dS-d4
to create a passed pawn in the
centre and activate the c6 bishop,
then after White's cxd4 and the
recapture . . . exd4, both the e- and
c-files will be open, and this will
give White counterplay. Black's
position at present is defensively
very strong, but if he wants to
win he has to allow some chinks
to appear in his armour. But it is
psychologically difficult to do The point of Black's play.
nothing and, as we saw in the White's pawns are broken up
chapter on depriving the oppo- since 23 i.xb7? dxc3 24 bxc3
 The Indian Bishop 87

'iib6+ wins a piece, or 23 b4 dxc3 30 c5?

24 ':xc3 'iVb6+ 25 :c5 %ld8 fol­ A time pressure blunder but 30
lowed by . . . .id4+. a4 'ifxa4 was bad anyway. The
23 'ot>h2 dxc3 black a7 pawn would prove very
24 bxc3 :d8 strong in combination with the
25 'ilb2 b6 dominant bishop on f6.
26 'ilb4 30 bxcS
This allows the black rook to Gurevich's only slip in a finely
penetrate. But 26 lln :d3 intend­ played game. 30 .. Jhf3 ! wins
ing . . . i.xc3 was very unpleasant. immediately.
Note the enonnous power of the 31 :'xc5
bishop on f6. 3 1 'ifxc5 avoids an immediate
26 :d2+ finish, but 3 1 . . .:xa3 leaves Black
27 'ot>h1 'ilia6! in total control: Black's kingside
Black does not want to ex­ is rock solid and he can prepare
change queens since the whole the advance of the a-pawn. White
king is vulnerable. would be hindered in his attempt
28 c4 to stop the passed pawn by the
White has no remaining con­ need to protect his precariously
structive ideas. Perhaps he should placed king. Now however it is
have played the desperate 28 g4 all over.
hoping to dislodge the black 31 Ihf3!
bishop with 29 g5. 0-1
28 %ld3! 32 1hf3 'ii'e 2+ wins the rook.
A nasty move to meet in time
pressure. White probably ex­
Novi k-Gallagher
pected 28 . . J�b2 when 29 'ifd6,
Oberwart 1 993
with the idea of creating a passed
pawn with c4-c5, gives White
counterplay. 28 . . J�d3 ! denies the 1 lDf3 lDf6 2 c4 g6 3 d4 .ig7 4
queen the d6 square and unex­ lDc3 0-0 5 e4 d6 6 .ie2 e5 7 d5
pected
y wins the a-pawn. as 8 h3 lLJa6 9 .ig5 'ife8 10 liJd2
29 'it>h2 lDd7 1 1 g4
Ruling out any ideas of. . .lhf3 A typical King's Indian battle
and . . . �c6 pinning the rook. is in full swing. Black's basic
29 �e8! plan is to achieve the . . . f7-f5 ad­
And suddenly the white a­ vance in order to dissolve
pawn is lost. A remarkable retreat White's e-pawn and liberate his
by the bishop. Gurevich is accu­ g7 bishop with . . . e5-e4. So he has
rate to the end, since 29 . .id7, a
. . played . .'i'e8, breaking the pin,
.

more plausible retreat, allows 30 and then . . . lDd7, clearing the way
�cd 1 . for the f-pawn. White of course is
88 Positional Sacrifices

doing his best to obstruct this Black, since White's strategy

plan, or to make sure that it turns would have failed, i.e. he would
out disadvantageously for Black. have lost the initiative. Novik is
So he began with .i.gS, pinning not prepared to give Black easy
the black knight; then he retreated equality after only 1 2 moves and
the knight to d2 in order to bol­ continues to play for advantage: a
ster the e4 square with f2-f3 after psychologically understandable,
. . . f7-fS ; lastly he played 1 1 g4 if incorrect, decision.
preventing . . . f7-fS for tactical 13 �n? !
reasons: 1 1 . . .fS? 12 gxfS gxf5 1 3
�hS and wins.

13 c6
As Novik and Nesis point out,
11 �dc5 there was no reason for Black to
This clears the d7 square for avoid the natural 1 3 .. .fS, e.g. 1 4
the queen, so that in the . . . f7-fS gxf5 gxf5 I S exfS �xfS . Black
variation above, �h5 can be an­ has an active position and is
swered by . . .... d7. ready to play . . . e5-e4, opening up
12 :gl the diagonal for his g7 bishop and
Another tactical variation to at the same time introducing the
dissuade .. f7-fS : 1 2 . . .f5 1 3 gxf5
. idea of . . . lDd3+. The annotators
gxfS 1 4 �h6 :f7 I S .i.h5 and also point out the variation 1 5
wins. So Black moves his king �hS (instead o f 1 5 exfS)
out of the potential pin. 1 5 . . . 1Wd7 1 6 lDg3 f4 1 7 �fS (D)
12 �h8! 17 . . . �e4 ! with a clear advan­
White has now come to the end tage to Black. Novik may have
of his tricks to ·prevent' . . . f7-fS. missed 17 . . . lDxe4 when he played
Novik and Nesis in Informator 58 1 3 lDf1 . Otherwise, the plan of
now give the laconic · 1 3 �e3 f5 lDf1-g3-fS would have been a
1 4 f3= ' , but this would undoubt­ very good one: the knight would
edl y represent a minor victorY for he excellentlv placed. the �7
 The Indian Bishop 89

bishops diagonal blocked and the would need an exceptional feel

g-file an avenue of attack against for chess to know intuitively
Black's king. whether or not a move can be
tactically refuted four moves
down the line.
14 lDg3
Now White assumes the initia­
tive again. If 14 . . .fS, White could
play I S gxfS gxfS 1 6 lDhS ! ( 1 6
.ll hS ! 1 ) eliminating the g 7 bishop
with a good game.
14 cxd5
15 cxd5 .td7
16 h4
Planning to ram the h-pawn
down the throat of the g7 bishop.
(analysis diagram) Black barricades his kingside and
looks for counterplay on the
And likewise, Gallagher may queenside:
have avoided 1 3 . . .fS because he 16 bS
did not see the tactic 1 7 . . . lDxe4 17 .d2 b4
which saves him from unpleasant 18 lDdl f6
pressure (and gives him a big ad­ 19 .te3 .e7
vantage). If this interpretation of 20 h5 gxh5
events is correct, then it demon­ 21 f3!
strates how important it is to be 2 1 lDxh5 allows 2 1 . . .lDxe4. Or
alert to the subtleties of the game. if 2 1 gxhS fS ! and Black has
Both players made a positional achieved his thematic advance.
error because they missed a tacti­ White therefore prefers to sacri­
cal point. Of course, if you have fice a pawn in order to keep a
very fine positional judgement, or grip on the fS square. If he can
have a strong sense of justice, prevent Black playing . . . f7-fS,
you could play 1 3 .. .f5 ! without then the bishop on g7 - the key
thinking and stumble 'by acci­ black piece - will either remain
dent' on 1 7 . . . lDxe4. The problem shut out of the game, or vulner­
is that chess is not always a just able to exchange by lDbSxg7,
game - moves which are 'natural ' when the dark squares in Black's
or 'logical ' or 'interesting ' often kingside will be left weak.
founder 'undeservedly ' on an Grandmaster Gallagher is fully
obscure tactical point. There is aware of the danger facing him
not always a move like 1 7 . . . lDxe4 and, realising that passive play is
to save the innocent. And you useless, returns the pawn to free
90 Positional Sacrifices

his bishop. bishop completes its journey to

activity. After 2S 0-0-0 :g8 26
liJgh l ! the knight which terror­
ised Black with ideas of liJfS ends
up on a ridiculous square. The
expression Black has dynamic
counterplay seems appropriate
here.
25 �xa6?!
With this exchange White
takes off the pressure of e4 before
playing lLlh5 eliminating the
black f6 bishop. The immediate
2S lLlhs allows a combination
21 hxg4 2S . . . �h4 26 i.xa6 lLlxe4 27 "ii'e2
22 fxg4 fS! liJxf2 28 �xf2 i.xf2+ 29 "ii'x f2
Playing the impossible move. lhfS (Nesis and Novik) . The
Now 23 liJxfS �xfS 24 exfS e4 piece sacrifice has broken open
and the bishop sees daylight; if the centre and we are suddenly
2S i.d4 liJd3+! 26 �xd3 �xd4. reminded that White also has a
23 exfS e4 ! is similar. Therefore, king. However, the assessment
White has no choice: clear advantage to Black in In­
23 gxfS i.f6 formator seems incorrect, since
24 liJa :ac8? there is no clear continuation for
Black after 30 "ii'h2 . The rook on
c8 is attacked and the threat of
llg7 may be dangerous. Hence
the sacrifice is dubious. There­
fore, White should play 25 lLlhS
immediately when after 2S . . . :g8
26 ':'xg8+ ':'xg8 27 0-0-0 we
reach a position similar to the
game, but with the bishop still on
e2 and the knight on a6. The dif­
ference favours White since the
offside knight is a much worse
piece than the bishop.
Curiously, Nesis and Novik do 25 liJxa6
not criticise this move, although 26 liJhS 11g8
it amounts to a loss of tempo, but 26 . . . i.h4 27 i.h6 is bad for
they do mention 24 . . . i.h4 ! which Black, but now he loses by ex­
is undoubtedly stronger: the change his prize dark-squared
 The Indian Bishop 91

bishop.
27 lbg8+ :xg8
28 0-0-0

34 lbe6
In time pressure, White missed
the strength of 34 d6 ! llxc5 35 d7
28 �e5 .i.xd7 (35 ... 'ii'd8 36 "d6 :c7 37
Black cannot keep the bishop: 'i'xc7 ! wins, or 36 ... 'i'xd7 37
28 . . . .i.h4 29 'ii'e2 ! �c5 30 .f3 .fS mate, or 36 . . . :c6 37 'ii'xe5+
.i.e8 3 1 l:h l .i.xf2 32 �xf2 �g8 38 :g 1 + wins) 36 "xd7 and
.g5+ 33 �e3 .i.xh5 34 �xg5 Black's king is defenceless, e.g.
.i.xf3 35 .i.f6+ :g7 36 llg I wins, 36 . . . 'i'fS 37 :hI 'ii'g 8 38 'ii'e7
or at move 3 1 , 3 1 . . . .i.xh5 32 1Ic8 39 1Ixe5+ "g7 40 f6 and
'ii'xhS .i.xf2 33 �xc5 ! �xc5 34 wins (Nesis and Novik) . In the
f6 .c7 35 f7 11g 1 + 36 rli>d2 absence of the g7 bishop, Black's
.i.e3+ 37 �d3 and mate on h7. king is often vulnerable.
These pretty variations are given 34 c3!
by Nesis and Novik. 35 11c2 a4
29 �xf6 'i'xf6
30 ..i.xc5 dxc5
31 �d3
White has eliminated the In­
dian bishop and has a much su­
perior pawn structure. He now
plans to answer 3 1 . . . c4 with 32
�c5 followed by d6, when the
passed pawn is very powerful.
31 �b5
32 'it>bl
The c-pawn is of course taboo.
32 c4
33 �c5 lIe8 36 llhl?
92 Positional Sacrifices

Panic in time pressure. White's 43 tiJc7

problem is that he has wrested a A desperate attempt to block
defmite positional advantage the c-file. White's king faces the
from the tactical melee, but now all-out attack of Black's pieces,
he can't find a way to kill off the and there is hardly any pawn
dynamism in the position and cover. It is no wonder his position
quietly enjoy this advantage. The soon collapses.
threat of 36 . . . b3 37 axb3 axb3 38 43 :i'e3+ 44 �b2 lIg8! 45
•.

'ilixb3 c2+ should be met by 36 lbb5 'it'e2+ 46 �a3 .i.c2 47 l:[el

bxc3 ! lhc3 37 'ilib2 .i.d3+ 38 .:a8+ 48 �b2 'ii'x el 49 �xc2
�a l when 38 . . . 'ii'e7 (defending 'ile2+ 50 �bl .l:lg8! 0-1
b4) 39 'it'h2 'it'f6 40 lIh l wins The change of front is decisive.
(Nesis and Novik) . But it is diffi­
cult to be rational in time trouble.
Arenclbia-Akopian
36 b3!
Biel 1993
37 axb3
37 'it'h2 .i.d3+ 38 �a l cxb2+
39 'iixb2 lIc2 and 40 . . . b2+ wins.
37 axb3
38 'it'xb3 .i.d3+
39 �c1 'ii'e 7!
But not 39 . . ..i.xe4 40 1i'b7 !
threatening mate by both 4 1
'Wxh7 and 4 1 'it'xc8+. Now, with
his flag about to fall, White has to
find a move that doesn't lose.
According to Nesis and Novik,
there is only one move: 40 'iib 6 !
to prevent the threatened
40 . . .'�i'a7 and introduce the idea Black began the process of un­
of 4 1 d6. Of course, it is no sur­ dermining White's centre with
prise that White does not find this 10 b4!
move. Now White should try 1 1 c4
40 bxc3? .i.xe4 when Black undoubtedly has the
41 %:tel better of it since White is weak
The only way to prevent on the diagonal a I -h8 and the d4
4 1 . . .'ii' a 7 which would threaten square in particular is a 'hole' .
both 42 . . . 'ii'e 3+ and 42 . . .'iVa l +. However, White's centre struc­
But now the queen infiltrates ture would remain intact and the
through a different route: closed nature of the position
41 'ii' h 4! would minimise Black's advan­
42 lIdl 'ii' g3 tage. In the game, White did not
 The Indian Bishop 93

want to admit his strategy had White has attacking chances

gone awry and so he , played a against Black's king (Akopian) .
perfunctory developing move: 19 Jl gS 1i'c7
11 Jld2? bxc3 20 lDd2 JlfS
12 .i.xc3 Black's pieces combine well
If 12 bxc3 c4! splits White's with his massive pawn centre.
centre pawns. Perhaps Arencibia There is no rush to advance the
thought he had neutralised the pawns since White has no coun­
Indian bishop, but he is in for a terplay or means to fortify his
rude awakening: position.
12 dS! 21 ltJel e4
13 eS 22 i.f4 JleS
White's centre now loses all Now Black allows the ex­
cohesion, but as Akopian points change of his dark-squared
out, 1 3 exd5 lDxd5 14 .i.xg7 bishop since White's knights
rJi;xg7 is clearly to Black's advan­ have been driven far away from
tage. Both b2 and f4 are attacked, the black king's defensive pe­
and if 1 5 1i'c 1 then 1 5 . . . ..,, 6 in­ rimeter.
troduces ideas of . 1i'xb2 or . . . c5-
. . 23 .i.xeS 1i'xeS
c4+ and . . . lDg4. An Indian queen to replace the
13 lDhS! Indian bishop.
Already Akopian is planning a 24 :bl d3
sacrifice to demolish White's 25 Jl g4 1i'd4+
overstretched centre. 0-1
14 1i'ct f6! 26 �h l e3 wins the bishop on
15 d4? g4 to start with.
Still refusing to accept how
badly he stands. He should try 1 5 Black however does not always
exf6 .i.xf6 1 6 i.xf6 ':xf6 1 7 g3, have it his own way. If White
though 1 7 ... .i.h3 or 1 7 ... 1i'b6 is succeeds in keeping the Indian
still clearly good for Black. The bishop immured then he will have
game continuation is hopeless. an excellent game.
15 ltJxf4!
Now White's whole centre dis-
Eplshl n-Gheorg h l u
appears with remarkable rapidity.
Geneva 1 993
16 1i'xf4 fxeS
17 ft4 exd4
18 .i.d2 eS! (see following diagram)
Black does not allow the ex­
change of his treasured dark­ 13 as!
squared bishop: 1 8 . . . lUb2 1 9 White has a space advantage .
.i.h6 followed b y 2 0 lDg5 and His natural plan is to prepare the
94 Positional Sacrifices

pawn advance b2-b4 which be­ ' good' bishop for White's 'bad'
gins the process of 'peeling ' bishop. This would mean that
Black's centre. However, if Black would be left with his mis­
White carelessly plays 1 3 lib 1 ? erable bishop on g7, which is
then Black can answer 1 3 . . . aS ! blocked in by its own pawns,
which completely blocks the while White's bishop on h4 could
queenside and frustrates White's be re-routed to active play after
plan. Hence the move 13 as, an eventual lDd2, f3 and .i.f2.
which also has another useful Play could continue 1 4 . . . lDh7 1 5
point that becomes clear on the .i.a4 .i.d7 16 .i.xd7 'iixd7 1 7
next move. 'ii'a4 ! 'iixa4 1 8 lDxa4 (threatening
1 9 .i.e7) 1 8 . . . �f6 1 9 b4 !

13 'if'e8
Black's plan is consistent with (analysis diagram)
the theme of this chapter: he pre­
pares the advance . . . f7-fS in order Now 1 9 . . . cxb4 20 c5 .i.xh4 2 1
to attack White's centre and per­ liJxh4 is clearly to White's advan­
haps eventually liberate the en­ tage: he will soon win the pawn
tombed bishop on g7. The game back and remain with a much
will be decided by how success­ better pawn structure after
fully Black is able to implement 2 1 . . .dxcS 22 liJxcs followed by
his strategy. 1iab 1 , etc. The protected passed
14 .i.c2! pawn on d5 would be very strong.
White has seen through No better for Black is 19 . . . .i.xh4
Black's intentions. If Black now 20 bxc5 ! .i.f6 (20 . . . .i.e7 2 1 cxd6
plays 1 4 . . . lDh7 (or 1 4 . . . lDhS), i.xd6 22 cS and 23 d6 regains the
aiming for the ... f7-fS advance, piece with a winning position) 2 1
then I S .i.a4 ! forces him to agree cxd6 lDa6 22 c5 and the passed
(after I S . . . .i.d7) to the position­ pawns are worth more than the
ally unfavourable exchange of his piece. e.g. 22 . . . :fc8 23 l:ifc 1 fol-
 The Indian Bishop 95

lowed by 24 nab 1 or 24 c6. preparation of the b4 advance,

In the game, Black avoided the since Black's queenside is well
bishop exchange but there were fortified by three minor pieces
other problems to face: and a rook.
14 liJ d7
15 g4!
If the knight were on h7, Black
could now play 1 5 . . . f5, achieving
the freeing . . . f7-f5 advance and
breaking up White 's kingside.
But because the knight is on d7,
obstructing the action of the
bishop on c8, White can simply
answer 1 5 .. .f5 with 1 6 exf5 gxf5
1 7 �xf5 winning a pawn. There­
fore, Black's natural plan is
thwarted, at least for the moment.
Epishin tries to achieve a vice­
like grip on the f5 square, and so 20 b5?
permanently prevent . . . f7-f5 . Gheorghiu tires of his passive
15 .t f6 defence. He should play
16 � g3 20 . . . .te7 ! . Then 2 1 tt:Je3 liJf6 at­
White does not allow Black to tacks g4, and now 22 liJh2 liJd7
ease the congestion in his ranks 23 liJf3 (else the h-pawn drops)
by exchanging bishops. 23 . . . liJf6 is a draw by repetition.
16 �g7 Alternatively, White could sacri­
17 'ii' d2 'ike7 fice a piece: 22 liJxe5 dxe5 23
18 �g2 %:th8 it.xe5 followed by the advancing
An admission that his plan of of his centre pawns en masse. But
. . . f7-f5 has been defeated. Black White's preparations for this sac­
waits patiently to see how White rifice are not as complete as they
can strengthen his position. are in the game at move 28 when
19 h4? ! he plays a similar sacrifice. Black
More accurate was in fact 1 9 can continue to fish in troubled
liJd l ! immediately. waters with 23 . . . �g8, e.g. 24
19 'ikd8 'iVc3 (better IS 24 l:lad l )
20 liJdl 24 . . . liJce8, etc.
White brings his knight to the If White had played 1 9 liJd 1 !
excellent e3 square as the prelude instead of the premature 1 9 h4? ! ,
to a direct attack on Black's cen­ Black would not have had this
tre and kingside. This strategy is defensive option: the g-pawn
now more promising than the would remain defended. Gheor-
96 Positional Sacrifices

ghiu however misses his chance 'itxgS 3 1 lLlxd6 wins. But now
and chases some will-of-the wisp comes a thematic sacrifice:
on the queenside. 28 lLlxeS! dxeS
21 axb6 'Dxb6 29 .i.xeS+ f6
22 'De3 'Da6 30 .i.e3!
23 l:[hl 'Db4 White has destroyed Black's
A rather useless manoeuvre centre, acquired two passed
with the knight which chases the pawns and gained strong attack­
white bishop to a better square ing chances against Black's king.
and does nothing to stop White's Epishin's marvellously restrained
methodical kingside build-up. It 30th move threatens 3 1 d6 when
would have been better to keep the bishop on e7 dare not move,
both knights on or near the king­ or 3 1 gxf6+ .i.xf6 32 eS powering
side, where they would help de­ through in the centre. The bishop
fend against White's inevitable on c3 is enormously strong.
breakthrough. 30 lDa4
24 .i.dl .td7 Black's only hope is to elimi­
2S .i.e2 as nate the white bishop.
Black's 'attack' on the queen­ 31 l:ha4!
side continues. Unfortunately for But this hope is dashed. White
him, there is nothing to attack. does not flinch from a further
Now White feels ready to act. exchange sacrifice, since he
26 gS! hxgS knows that his central juggernaut
27 hxgS will sweep all before it.
31 i:[xhl
32 Wxhl .i.xa4
33 'Dg4
Now f6 drops and Black's po­
sition collapses before the ram­
paging pawns.
33 'iih8+
34 Wgl Wf8
3S gxf6 .i.d6
36 eS .i.e7
37 'ile3! 'Da6
38 e6 .i.d6
39 e7+! 1-0
27 iLe7 39 . . . 'ite8 40 f7+ wins the
As Epishin points out, queen, or 39 . . Wf7 40 ..... e6+ and
.

27 . . . .i.xgS? 28 'DxgS 'i'xgS 29 mate next move. A fine game by

'Df5+ 'itf6 (forced) 30 'i'xgS+ Epishin.
7 The ' Ka rpov i a n ' Exc hange Sacrifice

Every strong player knows that but his activity apparently out­
you need a plan. It may be a weighs his positional deficien­
grand strategy that lasts many cies. For example, if 20 i.d3 :g8
moves, or a simple tactical ma­ intending . . . f7-f5 (Karpov), and
noeuvre, but it is important to be Black will have enormous pres­
doing something to enhance your sure along the a8-h l diagonal. In
position. Otherwise, the opponent the game, Karpov found a way to
will have endless time to kill off the dynamism in Black's
strengthen his game, and if his position. He began with:
position gets better, then perforce
your position will get worse.
Karpov 's speciality is depriv­
ing his opponent of a plan.
Petursson, the Icelandic grand­
master, has played both Kasparov
and Karpov. He sums up their
different styles as follows: 'When
you play Kasparov, you know he
wants to annihilate you; when
you play Karpov, nothing special
happens, but you lose. '
Here are some examples of
nothing much happening fol­ 20 l:tdS! !
lowed by top-class opposition when Gelfand initially turned
biting the dust: down the exchange offer. It is
worth making a few observations
about the position after the hypo­
Karpov-Gelfand
thetical 20 . . i.xd5 2 1 exd5 l:te7
.

Linares 1 993
(D) .
Rooks thrive on open lines.
Black' s pawn structure is ragged, The disappearance of Black's
9S Positional Sacrifices

bishop on b7 means that Black squares such as b5, a6, b7 and c6

can no longer hope to strike a would prove vulnerable.
blow against g2 (after . . . �gS). So White's absolute control of
the white squares and the lack of
open lines negates Black's mate­
rial advantage. And when we add
White's strong passed pawn and
Black's exposed king to the
'evidence' , it is easy to conclude
that White has all the chances.
Gelfand knows that a position
without a plan contains the seeds
of defeat. Hence, he refuses the
sacrifice.
20 �he8
20 .. .f5 2 1 lLle5 looks danger-
(analysis diagram) ous.
21 .td3 �d8
Therefore, the only means of Still hoping for some counter­
activity for Black's rooks is along play after 22 l:r.xdS 'ii'x dS 23 'i'c2
the e-file. However, after 22 :d l f5 ! .
IDleS 23 .td3 , for example, all 22 lId1
the breakthrough squares in Karpov keeps an iron grip on
White's position are guarded. the position. Black isn't allowed
Black's rooks are dressed up with a ghost of activity.
nowhere to go. 22 l%ed6
So much for the rooks. It' s also 23 .lte2 .ltxdS
difficult to imagine how Black's Gelfand finally gives in. How­
bishop on h6 is ever going to take ever, 23 . . . 'ii'c 6 24 l:hd6 �xd6 25
part in a co-ordinated action of lhd6 'i'xd6 26 'ii'xf7 was bad, as
the pieces. The opposite-coloured 26 . . . .ltxe4? fails to 27 'ifeS+; and
bishops greatly favour White. 23 . . Jhd5 24 exd5 'i'd6 25 'ii'c2 !
There are beautiful open diago­ is also unpleasant as 25 . . . .ltxd5?
nals for the white bishop. From 26 .ltc4 followed by 'ii'd 3 leaves
d3, it can threaten .ltxh7, acquir­ Black in a fatal pin, and mean­
ing a passed pawn which will while White is threatening 26
prove useful in an endgame, or 'i'xh7 or 26 .t c4 solidifying d5 .
after .ltf5 it could help to push Now Black at least has the ex­
the d-pawn to d7. Finally, if change as consolation for his
White chose to launch a direct misery, but we are effectively in
attack on Black's king, the bish­ the situation referred to in the
ops control or potential control of note to move 20.
 The 'Karpovian ' Exchange Sacrifice 99

24 exd5 1i'b7 would be left without a con­

structive plan. It would be nec­
essary to wait and see if White
could strengthen his position.
It's no wonder that this approach
did not appeal to Gelfand. Nev­
ertheless, patience is still a vir­
tue and White would have to
work hard to generate real win­
ning chances. After 26 . . Jhd5 on
the other hand, Black is left with
weaknesses everywhere, and
does not even have any material
compensation.
25 'ilh4! 27 .ixd5 l:lxd5
Karpov's play consists of 28 :el !
grand strategical strokes and an Of course he keeps the rooks
alertness to tactical nuances. If he on; this not only ensures the
were to lose the d5 pawn, then his safety of White's king from back­
whole positional build up would rank checks, but also preserves
be proved wrong. But Karpov has threats against Black's exposed
calculated that he has just enough king.
time to get in the vital .i.c4 move. 28 :d8
He gives the following note in 29 .xf6!
Informator 57: Karpov is determined not to
a) 25 . . . :xd5? 26 :xd5 1i'xd5 give his opponent any counter­
27 .xh6 1i'd l + 28 .i.f1 ! , or play whatsoever. The obvious 29
b) 25 . . . .i.g5 26 1i'g3 .c7 27 .xh7 allows 29 . . . 1i'd5 with the
.i.c4. idea of . . . c5-c4 ! ? followed by
Hence the d5 pawn is inviola­ . . . .i.c5, when Black may get
ble. In his earlier calculations, some attacking chances against
Gelfand may have overlooked 28 f2 as well as White's queenside
.i.f1 ! in the first variation, or per­ pawns. After 29 'ii'xf6 .d5
haps in the second variation he Black does not threaten 30 . c4. .

missed the quiet 26 .g3 , gaining because of the simple 3 1 bxc4

time for the vital .ic4 by pinning since Black's rook on d8 needs
the rook. defending.
25 .i.f8 29 'fIc7
26 .tc4 l:lxd5 30 g3! .i.d6
Black hastens to counter­ 31 tLl g5
sacrifice the exchange. After the If Black had prevented this
alternative 26 . . . h6 27 .e4 he move with 30 .. h6 then 3 1 tLle5
.
1 00 Positional Sacrifices

would have been equally strong.

We now see the point of 30 g3 ! .
Lutz-Karpov
If the white pawn were on g2,
Dortmund 1 993
Black would have the trick
30 . . . i.e7 ! avoiding the worst.
Now however he must allow
White to penetrate into his back
rank with fatal consequences.
31 :'d7
32 :e8+
Again Karpov prefers to keep
control rather than allow Black
counterplay after 32 lDxh7 c4! .
32 <ii' b 7
33 liJe4 i.e7
34 'iff5
White sets his sight on the h­
pawn. White can play 34 "ilxfl White has just played 2 1 'ii'hS
safely, but there are many ways hoping for 2 1 . . . .i.xf4 22 gxf4
to win. l:xf4 23 e5 ! winning - Black
34 "ilc6 must now defend or exchange his
35 �g2 ltc7 rook when 24 'iixh7+ is decisive,
36 :'h8 'iVg6 and if 23 . . . g6 24 .i.xg6! . Note
It is too late for 36 . . . c4 37 how strong the white bishop be­
bxc4 "ilxc4 3S :xh7 "ilxa2(?) 39 comes in this variation. Karpov
'iWbS+ 'it>cs 40 :hS+ .i.dS 4 1 found a way to frustrate White's
lDd6 mate. plan:
37 'i'dS+ 'i'c6 21 :'xf4!
38 'ifxc6+ �xc6 22 gxf4 .i.xf4+
39 Ibh7 �dS It's vital that this is check.
39 . . . c4 40 bxc4 �d7 4 1 c5 is Otherwise, White would have
hopeless for Black. time to play 23 e5.
40 lDd2 .i.f6 23 �bl "ileS !
41 lDc4 'itd4 Now, however, Black can en­
42 ':'h6 sure that White's bishop remains
The bad bishop plagues Black blocked in.
right until the end. 24 'i'xeS
42 ':'c6 After 24 'iff3 IUS 25 :f2 g6 !
43 g4 :e6 Black can play the manoeuvre
44 h4 �dS . . . i.h6-g7 aiming at b2. The
45 gS 1-0 black queen is much more power­
The passed pawn wins the day. fully placed than its white coun-
 The 'Karpovian ' Exchange Sacrifice 101

terpart, so Lutz decides to ex­

change queens.
24 .ixe5
25 h3
Its now time to sum up the
consequences of Black's sacri­
fice. He has an absolutely safe
game, without any weaknesses.
The bishop on eS is impregnable,
and exerts influence across the
whole board. Black's 2-1 major­
ity of pawns on the kingside
could later be converted into a
passed pawn. There is no har­ Now 26 . . . a4? could be an­
mony among White's pieces and swered by 27 b4, keeping the
pawns. The pawns on c4 and e4, black rook out of as. The pawn
besides being vulnerable to at­ on a4 would also be a target - if
tack, shut the bishop on d3 out of the rook moves away, then .i.c2
the game. White's rooks have followed by b4-bS would threaten
few active opportunities. Their to win the pawn. However, in
only hope of penetrating into playing 26 b3, White has com­
Black's position is along the f­ promised his pawn structure
file. Naturally, Karpov makes somewhat on the queenside. This
sure that this hope is frustrated. slight weakness proves crucial
Assuming that White can do later on.
nothing aggressive, how should 26 d6! !
Black play to increase his advan­ A wonderful example of Kar­
tage? Karpov answers this ques­ pov's positional play. He realises
tion in the course of his subse­ that the bishop on c6 should go to
quent play. g6 in order to thwart any white
25 as! activity on the f-file. And by
First, Black threatens 26 . . . a4 moving from c6, the way is
which would undermine the c­ cleared for the rook to enter into
pawn. It would be a fixed target the game via c8, cS and eventu­
and White would no longer be ally hS. This is planning on a
able to defend it safely with b2- grand scale!
b3 . After 26 . . . a4, Black would 27 lld2 .i.e8
also have the option of . . . llaS 28 �c2 .ig6
followed by . . . .i.f6 and . . . llhS , 29 lldf2
bringing his rook into the at­ The rooks may look impres­
tack. sive, but there is no breakthrough
26 bJ square. Karpov's subtle play has
1 02 Positional Sacrifices

reduced his opponent to passivity. If he waits with 40 ltg3 then

He now continues his plan. 40 . . . h6 4 1 J:tg4 i.. e3 followed by
29 :te8 . . . .tf4 wins the e-pawn. However,
30 <it;dl .ltd4! Karpov would probably have
The rook must move away . chosen a slower, more methodical
from the f-file since 3 1 :f3? .th5 plan of execution.
or 3 1 :f4 i.. e 3 32 :h4 .ltg5 win 40 gxh4
the exchange. 41 Itxh4 l:[ g5
31 :a2 :te5 As Karpov points out 4 1 . ..d5 !
Now that the white rook has is decisive after 42 lUf4 (42 %:te l
been forced away from f2, this .tf2) 42 . . . .te3 43 l:Ug4 .tf2 44
can be played without allowing lIh2 .td4+ followed by . . . dxe4.
32 :tfS mate. After winning the e-pawn and
32 ltg2 creating his own passed pawn on
All White can do is mark time e4, Black could gradually plan
and defend against any direct the advance of the h7 pawn.
threats. White would not have the re­
32 :h5 sources to fight off both passed
33 :0 .lte8! pawns. However, Karpov's
Now that his rook has been ac­ 4 1 . . .:g5 is also ultimately win­
tivated, Black begins the next ning.
phase of his plan. This involves 42 l:.h2 :tg3
utilising his kingside pawn ma­ 43 .te2 .ltd4+
j ority. 44 �c1 a4!
34 <;te2 g6 Breaking up White's queenside
35 .te2 :e5 pawns. Compare the note after
36 .td3 �g7 White's 26th move.
37 %Lg4 g5 45 bxa4 l:txa3 46 �d2 :g3 47
Karpov is finally happy with i.. d3 :tg5 48 11thl .te5 49 Ith3
his preparations and the pawns i.. d4 50 :Jh2 :tg3 51 l:tn .tf6
begin to advance. 52 :tbl .ltg5+ 53 <it;c2 .te3 54
38 lIn .ltc5 :al .ltgl 55 %1d2 <;tf6
39 �b2 After interminable manoeu­
Here Karpov recommends 39 vres, designed no doubt to de­
a4. However, White' s pawns then moralise his opponent, Karpov
lose their flexibility on the brings his king into the centre in
queenside, and he has no hope of order to pressurise White's de­
ever activating his game with b3- fences even further. There is also
b4. the looming threat from the h­
39 .ltg6 pawn, which White will find very
40 h4 hard to stop.
Lutz decides to force the issue. 56 a5 bxa5 57 l:.xa5 .tc5 58
 The Karpovian ' Exchange Sacrifice
I 103

:a1 �e5 59 lUI .tg1 60 :dd1

This loses control of the sec­
Sion Castro-Karpov
ond rank, but White had to do
Lyon 1 993
something against Black's plan of
advancing his h-pawn.
60 ... .te3 61 :f8
An aggressive move, but what
is there to attack?
61 ... :g2+ 62 �b3 :h2!
Black wants to play . . . .th5,
activating his bishop and driving
the white rook away from the
good defensive square at d l ,
without allowing :h l in reply.
63 .tb1 .th5 64 %:tel .tf2 65
:.n .tc5 66 :tel .te2 67 .ta2
White loses more and more
ground. White has just played 18 .tc3 .
67 ... l:h3+ 68 <iii>b2 .ta3+ 69 He hopes that the threat of 1 9 d5,
�a1 .td3 70 .tb1 discovering an attack on the rook
This allows a winning combi­ on h8, will force Black either to
nation. However, the game could abandon the defence of the h­
not be saved. Even if White sur­ pawn or weaken himself in some
vived direct threats to his king other way. Karpov, however,
(for example, 70 l:.g8 .tb4 7 1 sticks to the principle that the
:tel :th l ! ! 72 :txh l .tc3 mate) best answer to a threat is to ig­
the h-pawn would prove unstop­ nore it. He played:
pable. It's curious that Karpov 18 �b8!
never needs to use this pawn in White had the temerity to call
the game. Karpov ' s bluff with
70... .tb4 71 l:.ct .td2 72 :d1 19 d5 cxd5
72 .t xd 3 .txc 1 is hopeless de­ 20 .txh8 l:.xh8
spite the opposite coloured bish­ Already, we can discern some
ops. The h-pawn runs through. similar features to the Lutz­
72 ... .tc3+ 73 'ii7a2 .txc4+ 74 Karpov game above. Black has a
�a3 .te2 0-1 strong dark-squared bishop which
White was probably relieved to has no rival. This gives him as­
be able to resign. Wherever his cendancy over some key central
rook goes, it will be lost to a dis­ squares such as e5, f4 and c5 .
covered attack, e.g. 75 l:te l (However, the domination is not
.td2+. A fine display of sus­ as great as in the Lutz game.) The
tained pressure from Karpov. white rooks have no open lines.
1 04 Positional Sacrifices

Their plight is even worse than in 24 . . . :xh6 25 'i'xh6 ttJxg4 26

the Lutz game, where they at 'ii'h 8+ r3;a7 and White will lose
least had the f-file. If White tries one or more of his kingside
to open lines with c2-c4 then the pawns, e.g. 27 f3 ttJf2 (best) 28
bishop on b7 will become very !:te l ttJxd3 29 cxd3 iLxh2. In
strong after . . . dxc4. Besides, the trying to prevent the dislocation
square c 1 is controlled by Black's of his kingside, Sion falls under a
bishop so it is not clear that direct mating attack.
White would be able to profit by
the opening of the c-file. The al­
ternative method of opening a file
is h2-h4, but there are great prac­
tical difficulties in achieving this
advance. (The game continuation
will make this clear.) Assuming
that c2-c4 is unwise and h2-h4
impractical, White has no plan.
All he can do is wait and see if
Black can improve his position.
21 ttJd4 a6!
Not 2 1 . . .iLxh2 22 ttJb5 'i'f4 23
![h l winning the bishop. But now 24 iLf5 ttJc4
22 . . . iLxh2 is threatened. 25 lixh6
22 ttJxe6 If 25 :de l 'iic 5 ! with ideas of
' Patience' - 22 1Wfl ttJc5 23 h3 . . . 'it'b4 or .. :iid4 or . . . 'iix f2.
l::tc 8 - leaves White without any 25 Ihh6
constructive ideas. Black can 26 'it'xh6 'iie 5!
prepare to advance his centre A powerful entrance, threaten­
pawns and drive back White's ing mate. The position of Black's
pieces. Having watched the slow queen and bishop may remind the
and excruciating demise of Lutz, reader of the Lutz game.
one cannot be too critical of 27 'ii'f8 + <iia7
Sion's decision. 28 'iib4 ttJd2+!
22 fxe6 Once again Karpov demon­
23 1:txe6 ttJe5 strates that he is a masterly tacti-
Although White has nominal ciano
material compensation, the two 29 �al 'ife2
minor pieces far outweigh the Now the rook has no good
rook and pawns. White's rook on squares. 30 l:lc 1 ttJb3+ or 30 ':h l
d l is still passive and his kingside d4 fail to solve his problems.
pawns are weak. If, for example, 30 llgl 'ii'xfl
White plays 24 :xh6 then 31 f!dl "it'e2
 The 'Karpovian ' Exchange Sacrifice 105

Back again, with one white force of Black's minor pieces.

pawn less. (The same plan would apply if
32 l1g1 'iix h2 White had played 34 %:tg l .) In a
33 l:td1 'iie2 middlegame-type position involv­
Ditto. ing a direct attack on a king, two
34 l:th1 minor pieces are generally of
Now at last the rook sees day­ much more value than a rook. If
light, but it's too late. we remove the queens in this
34 a5 position, White's chances are
35 'Yic3 greatly improved.
An understandable blunder af­ 35 d4
ter all that White has been 0-1
through. Karpov gives 35 'ii'f8
lLlc4 followed by 36 . :�e5 or
. Yes, Karpov is a mean player
36 . . . .ll e 5. White's king wouldn't who loves to torture his oppo­
last long against the combined nent.
8 Queen for Rook a n d Bishop Sacrifices

To the beginner, the loss of the perpetual check) 26 'i'h4 b2 27

queen is almost as devastating as 'iVd8+ �g7 28 'iVxb6 :xh2 and
the loss of the king. Perhaps this White resigned since . . . :11 1 +
is why players often miss the followed by queening i s an un­
chance to make an effective stoppable threat.
queen sacrifice: they never quite
overcome their early impression
of the queen's omnipotence.
Here are some examples to
persuade the reader that it can be
a good idea to part with the
queen. The examples are grouped
in descending order, from the
brilliant to the bad.

8.1 Completely winning

Vulevi c-Regez
No fine judgement was needed
Wettingen 1 993 in this case. Indeed, the finish
(see following diagram) could be calculated from
22 . . . nxb2 to 28 . Jbh2, in view
.

Here Black played 22 . . . :xb2 23 of the forced nature of the play.

lbxd5 .i.xd5 . One cannot call this
a ' real' sacrifice - it is a simplify­ 8.2 Winning
ing combination, after which the
queen cannot prevent the rook
Reeh-Lau
and bishop shepherding home the
Germany 1 993
b-pawn. The game finished 24
'ilfxh5 lbg2+ 25 �f1 e6 (the rook
rln r? thw�rts White' s last hope. (see following diagram)
 Queen for Rook and Bishop Sacrifices 1 07

a piece to get an attack on the

white king in this position; but
there is no possibility at all here.
Meanwhile, White has a clear and
simple plan to increase his advan­
tage. He can put his bishop on e5
and double rooks on the seventh
rank.
21 i.. eS
22 fLe7 'ii'f7
23 i.b5
The bishops flex their power.
23 i.. d7
Here White played: 24 i.d6 AdS
20 'ii'x e7! :'xe7 The valiant 24 . . . fLxb5 comes
21 Axe7 out a piece down.
We note the following features 25 i.e5
of the sacrifice: Finally the bishop reaches its
i) The rook infiltrates to the ideal square. White is in no hurry
seventh rank, where it is excel­ to play ':xa7.
lently placed. Black cannot con­ 25 "'eS
test control of the open c-file - 26 i.. a6 :as
the only open line on the whole 27 i.b7 l:.bS
board. 2S i.. a6 :as
ii) The power of the bishop on A bit of cat and mouse. Now
a3 is greatly increased by the dis­ White begins the final stage of
appearance of Black's dark­ his plan - to play his knight to c5,
squared bishop. The dark squares threatening to capture the bishop
in Black's kingside are indefen­ followed by Abb7, crashing
sible. Meanwhile, the black through on the seventh rank.
bishop on b7 is a miserable piece. 29 lbb3! lbe4
iii) Black has absolutely no 30 i.. xe4 dxe4
counterplay. His queen is blocked 31 lbc5
in by the pawn structure on the White has carefully calculated
kingside. White' s position con­ that his attack gets in before
tains no chinks. The king on g I is Black can set up any threats
perfectly safe in its fortress. The along the newly opened a8-h i
question of · counterplay and king diagonal. It is frustrating for
safety will come up again and Black's bishop that it sees day­
again in our discussion of queen light too late. After 3 1 . . .i.. c6 32
sacrifices. Black would willingly 1:.g7+ �f8 33 lth7 ! Black loses
give up one or two pawns or even his queen to the threat of 34
1 08 Positional Sacrifices

:11 8 +. The black queen is so White's favour. So rather than

constricted that it has no safe recapture the piece Black played
square to flee to. this zwischenzug, trustingly at­
31 l:d8 tacking White's queen. White's
32 .tf6 l:tb8 reply was unexpected:
33 l::t xb8 "xbS 21 .txf6! .txd3
34 lIb7! "'18 22 .te5
35 lilxd7 "'a3 White only has two pieces for
Or 35 . ....n 36 llb8+ mating.
. the queen, but Black' s queen has
36 liJe5 '1'18 no safe square. For example, if
37 l:g7+ 1-0 22 . . .'ii'rl7 then 23 l:ad l .te4 24
ltlxe6 ! tears Black apart; or
It was possible for White to 22 . . .'ikc5 23 ltlxe6 l:xe6 24
make the queen sacrifice above .txe6+ �h8 25 l:ac 1 'ike7 26
without any concrete calculation ltc7 ! 'ikxe6 27 .txg7+ and wins.
at all. The following one required This last variation demonstrates
more calculation, though again it the enormous power of the two
is positional in nature. bishops bearing down on Black's
king position.
22 "'e7!
G . Kuzm i n -Turov
23 :act!
Russia 1 993
Black hoped for 23 ltlxe6 when
23 . . .... xe6 ! 24 .txe6+ llxe6 gives
a drawn position. White correctly
j udges that his attack is worth
more than the queen. Now,
23 . . . l:Iac8? would lose a rook af­
ter 24 l:xc8 and 25 .txe6+.
Meanwhile 24 llc7 is threatened.
So Black decided to give up a
rook to at least be free of one of
his tormentors, the knight on d4.
23 llad8
24 tDc6 'ikxh4
25 lilxd8 :lxd8
20 .te4 25 . . . 'ikxd8 26 lled l is a fatal
Here Black saw that 20 . . . .txd4 pin.
2 1 "'xf5 ! or 20 . . ...xd4 (which is 26 .txe6+ �h8
best) 2 1 llxe6 ! l:he6 (2 1 . . .•xd3 27 :c7
22 l:xe8 mate is the simplest The idea of a dominant rook on
queen sacrifice in this book) 22 c7 and raking bishops should re­
.txe6+ �h8 23 .xf5 is in call the previous game to the
 Queen/or Rook and Bishop Sacrifices 109

reader. pieces. The white pawn on a3 is

27 'ii' h6 defenceless, and after it is cap­
28 ':xa7! tured, Black will have a tremen­
Don't hurry ! Before winning dously powerful a-pawn. This
back the queen, White snaffles a pawn can be forced to its queen­
pawn. ing square by the knight and
28 �e4 bishop, while the rooks tie White
29 .Jtxg7+ 'ii'xg7 down in the centre, assuming of
30 lIxg7 �xg7 course that nothing happens in
and White has a winning end­ the meantime.
game. The game finished: White's one hope is the slight
31 f3 'ittf6 32 .Jta2 �a8 33 weakness of Black's king posi­
.l:le6+ 'itt g5 34 �xb6 lId1 + 35 tion, especially the g6 square.
<it>h2 :!d2 36 'itt h3 lId1 37 ':b4 Note that in the previous exam­
l::t d2 38 .Jte6 <it>f6 39 ':b6 <it>g5 40 ple, there were no weaknesses
.l:lb5 �f4 41 lixf5+ <itte 3 42 <itt g3 whatsoever in the queen sacrifi­
':xb2 43 l:teS+ 1-0 cers position. This meant he
could manoeuvre as he pleased
8.3 Promising without any risk at all. Note that
if Black's king were on h8 and
the Black h-pawn on h7, White
Lautier-M . G u revich
would be lost. He would have no
M u n ich 1 993
counterplay.
Here, however, he must tread
carefully. For example, if Black
now played 40 . . . 4Jxa3, White
could respond 4 1 f5 ! g5 42 f6 !
(activity at all costs ! ) 42 . . . 4Jxf6
43 4Jg3 and White has attacking
chances. Since the a-pawn is
doomed anyway, Black tried to
reduce his opponent 's counter­
play before capturing it by return­
ing the knight to the centre (with
the threat of . . . 4Jg4) :
40 4Jf6
Black played: 41 �f1 4Jxa3?
38 'iixd3! In In/ormator 5 7, Gurevich
39 'ii'x d3 lbe3 criticises his decision and instead
40 'iWd2 recommends 4 1 . . Jha3 ! 42 4Jg3
In return for the queen, Black �a 1 43 <it>g2 a3 44 .td3 (44 i.c4
has a good co-ordination of the is met by . . . 4Jc7 and . . . 4Jcd5)
1 10 Positional Sacrifices

44 . . a2 with a clear advantage to

. 55 r.itg2 lLld5
Black. Now White has a chance: 56 h4
42 lLlg3 :tel This makes the win easier
43 <.tg2 lLlg4 since the pawn becomes a target
44 R.e2! lLlxtl on h4. Black now proceeded to
45 Wxel lLlg4 win first the c-pawn then the h­
46 Wd2 lLle3+ pawn:
47 <.ttl lLld5 56 h5 57 r.ith3 lLlf4+ 58 <it>h2
•..

Now White succeeds in elimi­ lLlfe6 59 We7 r.ith6 60 cat>hl R.fS

nating the dangerous passed pawn 61 'ilf6 i.. g7 62 'ife7 lLlfS 63 'ife3
since 48 . . . lLl c3 49 Wd7 followed <it>h7 64 'ife8 cat>g8 65 r.itg2 R.f6
by R. d3 or f4-f5 gives White an 66 cat>h3 lLlde6! 67 "xc6 i.. d4 68
attack. 'ild5 i.. xcs 69 "b7 R.d4 70 "d5
48 lLlxf4 �h7 71 'i'D i.. g7 72 .17 lLld4
49 Wxa4 lLlbl 73 <it>g2 lLlfe6 74 r.ith3 lLlc5 75
50 Wdl! We7 lLld3 76 We3 lLle6! 77 r.itg3
Lautier agrees to enter an end­ lLle5 78 Wb6 lLlfS 79 "d6 lLlf7
game with another material im­ 80 'fIe7 R.eS+ 81 r.ith3 cat>g7 82
balance rather than face the dan­ <ifi1g2 lLlh7 83 r.ith3 lLlf6 84 We6
gerous all-out attack of Black's lLlg8 0-1
pieces after 50 'ifc2 lLlc3 fol­ Black has fmally reached his
lowed by . . . R. d4+. ideal set-up. He will play . . . lLlg8-
50 lLlc3 h6-f5 and . . . i..f6 winning the h­
51 "d7 lbe2+ pawn. Then the advance of his
52 lLlxe2 lLlfxe2 passed pawns will be decisive.
53 b5?
As Gurevich points out, 53
Klovans-Didysko
W xb7 lLld4 is only slightly better
Katowice 1 993
for Black. Black's plan would be
to win the two white queenside
pawns for his own c-pawn and
then advance his own kingside
pawns, sheltering his king with
his minor pieces. Whether he
could win or not is uncertain, but
it will certainly be a long and ar­
duous defence for White.
53 lLld4!
54 bxc6
54 b6 loses to 54 . . . lLle4+ and
55 . . . lLlxc5 .
54 bxc6
 Queen for Rook and Bishop Sacrifices 111

White has sacrificed a pawn

and now tried to force home his
attack with:
26 �f6 ! ?
This threatens 2 7 �xg7 ! fol­
lowed by 28 f6, mating. So Black
defended with:
26 �h8
White then pressed on with:
27 l: d 7
This looks decisive. If
27 . . . 'i'c4 then 28 b3 'ii'c6 29 :xf7
gxf6 30 l:dd7 and mate beckons
on h7, unless Black sacrifices his 33 :n
queen. Black in fact did so Didysko criticises this move in
straight away: Informator 57 without saying
27 gxf6! what White should play. Black is
28 :xc7 l'bc7 threatening 33 . . . :c4, so what
White's attack has vanished. should White do?
All Black's pieces are excellently a) 33 b3 (preventing . . . :c4)
co-ordinated and his bishops are l:xc3 ! 34 �xc3 i.xg2+ 35 �g l
potent. In trying to gain some �c5+ 36 :d4 �xd4 mate.
central squares for his knights, by b) 33 :d6 :c4 34 lDxf6
capturing the e-pawn, White un­ i.xg2+ 35 �g l �3+ wins
fortunately brings Black's bish­ White's queen and remains a
ops to life. rook up.
29 'ii' h4 i.e7 c) 33 lDxf6 l%g7 34 'iVh6
30 lDxe4 l:g8! (hoping to play 35 'ii'x g7+ and 36
Better than 30 . . . l:c4 3 1 �2g3 �e8+ wlnnmg both rooks)
l:g8 32 :d7 ! with an attack. Now 34 . . . i.xg2+ 35 Citg l �c5+ and
Black has strong pressure against wins.
g2. d) 33 :e l :c4 threatening
31 c3 34 . . . �d3 35 :e2 :Xg3 ! 36 hxg3
3 1 lDd6 :xg2 ! 32 lDxf7+ <t>g8 i.xe4 planning a fork on f2.
33 lDh6+ r:3;; g 7 wins, or alter­ To sum up: White is probably
natively 3 1 �xf6 :g7 ! threat­ losing even before 33 :f1 .
ening both 32 . . . i.xg2+ and 33 :c4
32 . . . :c6 winning the knight. 34 :xf4
31 lDd5 Equally useless is 34 :e 1 �d3
Threatening 32 . . . lDe3 and as in line (d) above.
33 . . . l:lg4. 34 exf4
32 lD2g3 lDf4 35 'ii'xf4 l:xg3!
1 12 Positional Sacrifices

36 hxg3 .i.xe4 rial equivalent - he also had many

37 �h2 positional features in his favour:
Of course 37 b3 .i.xg2+ loses. the open g-file for his rook bear­
What follows is White's death ing down on g2 where its power
throes, as his queen is gradually interacts with the enormously
overwhelmed by the numerous strong bishop on a8; a strong
black pieces: solid centre; a well placed knight
37... .i.dS 38 'i'b8+ �g7 39 on b4 which was easily manoeu­
'i'xb6 lba4 40 'i'e7 l:e4 41 'i'aS vred to the even better c4 square;
.i.e4 42 'i'e7 .i.f1 43 �gl .i.d3 and, not least, the c4 square for
44 'i'd7 �e5+ 45 �h2 .i.e2 46 his rook. His king was perfectly
'i'e6 l:e5 safe, and the black bishop on e7
47 g4 hindered any attempt to break
Black threatened 47 . . . .i.g4, through on f6. Meanwhile,
48 . . . l:xf5 and 49 . . . l:h5 mate, so White's knights proved the adage
White makes space for the king. that the worst bishop is better
47 .i.f2 than the best knight !
Now there are ideas of
48 . . . l:e3 and 49 . . . .i.g3+ with a 8.4 Unclear
quick mate on the back rank.
48 g5 fxg5
Genov-S. lvanov
49 g4 .i. x g4
Berlin 1 993
A good harvest.
50 'it'xa6 l:e2
51 f6+ �h6
52 'i'a8
Its too late to do anything with
the queenside pawns .
52 .i.e5+
53 �g3 .i.h5
54 'i'd5 .i.e3
-0 1
It 's mate after 55 . . . i.f4+,
56 . . . g4+ and 57 . . . l:h2.

A drastic defeat. Looking at the

position after 30 . . . l:g8, the pun­ White has sacrificed a pawn
ishment seemed out of all pro­ for attacking chances. He must
portion to White's 'crimes' . Yet now have expected Black to
Black not only gained a healthy move his queen (since 1 8 i.h7+
rook, bishop and pawn for the followed by 19 lhd5 is threat­
queen - the full theoretical mate- ened) when he has dangerous
 Queenfor Rook and Bishop Sacrifices 1 13

possibilities. In Informator 58, g5 :fd8 22 gxh6 4Jxe5 ! 23 4Jxe5

Ivanov gives the variation (23 'i'g3 4Jxf3+ 24 'iixf3 lIg5+
17 .. :iia 5 1 8 'ii'e4 ! g6 1 9 'iWf4 also loses for White) 23 . . . :d l +
�h7 20 oltxg6+ ! c;.t>xg6 (20 . . . fxg6 and mates. Therefore White must
2 1 .l:id7+ followed by 22 'Yi'xh6 be more subtle. He should probe
wins) 2 1 'iif6+ 'it>h7 22 :d3 with with his queen and try to coax
the threat of 23 4Jg5+ ! hxg5 24 Black into weakening pawn ad­
![h3+ and mates. vances.
Instead, Black killed the White Black, for his part, is in some­
attack stone dead by sacrificing thing of a dilemma. He has a very
his queen: secure position and can draw by
17 %lad8! sitting tight. White is unable to
18 olt h7+ �xh7 puncture holes in his position. On
19 l:txd5 :Xd5 the other hand, Black thinks he
has winning chances by advanc­
ing his queenside pawns and
creating a passed pawn. Such a
policy is not without danger, as
the game continuation demon­
strates.
20 h4
This is not the prelude to an
unsound kingside attack. White is
making an escape square for his
king.
20 �g8
Preparatory to 2 1 . . .:fd8; Black
Black now has active pieces, a defends the f-pawn. There is one
solid pawn structure and a safe chink in Black's armour and that
king. Furthermore, he has the is the insecure position of the
theoretical material equivalent of knight on c6, which isn't de­
a queen - a rook, bishop and fended by a pawn. Therefore
pawn - so it is not clear that the 20 . . . :d7 ! preparing to manoeuvre
term sacrifice is justified. the knight to the unassailable d5
What should White do? At pre­ square looks best. Then if 2 1 'ifa4
sent, he cannot hope to directly llfd8 22 :c l liJe7 and 23 'iix a7?
attack Black's king by advancing is impossible because of
his kingside pawns, since the 23 . . .oltxf3 in reply. Or if 2 1 :c l
bishop on b7 is waiting to gun immediately then 2 1 . . .4Je7 and
him down along the a8-h l diago­ Black has nothing to fear. In the
nal if he loosens his position too game, Black prefers to keep his
much; for example, 20 g4 �g8 2 1 rook on d5 and the knight on c6
1 14 Positional Sacrifices

in order to pressurise the eS vances; in particular, b6 beckons.

pawn. But this gives White some To achieve this, White has to give
chances to pressurise the queen­ up his a-pawn for Black's, but
side. this isn't necessarily a disadvan­
21 ltd llfd8 tageous exercise, as will be seen.
22 "'a4 llaS
23 "'b3 :adS
Black has managed to fend off
White's first attempt to promote a
weakness in his queenside. Note
that 23 . . . ltJxeS 24 ltJxeS :XeS 2S
:c7 i.dS 26 'ii'c 3 is bad for
Black. The exchange of Black's
a-pawn for White's e-pawn
would loosen the cohesion of
Black's queenside and rid White
of his weak e-pawn.
24 rt>h2 ll8d7
25 l%c3 bS! ? 28
2S . . . ltJxeS 26 ltJxeS :xeS 27 29 l:ta3
'ii'a4 attacks the d7 rook and so 30 llxa5
wins the a-pawn; after the plau­ Both sides have made progress.
sible 27 . . . :edS 28 'ii'xa7 White Black has eliminated White's last
has achieved a queenside break­ queenside pawn. This is an im­
through, leaving the b6 pawn portant achievement since we
very vulnerable. 2S . . . bS ! ? hopes know that a queen plus a passed
to tum the queenside pawns into pawn is a dangerous combination.
a positive asset, not just an object Furthermore, Black's own passed
of attack. But any advance is pawn could become an important
double-edged. feature if he succeeds in advanc­
26 "'c2! ing it; but this would be difficult.
The insecure position of the White has succeeded in activating
knight reveals itself again. Now his rook on an open file and his
26 . . . ltJxeS 27 ltJxeS :XeS 28 queen is also ready to infiltrate
llc8+ ! won't do. into the black position. If White
26 ltJb4 managed to penetrate to the back
27 'ii'e2 as (D) rank, then Black's king would
28 'ii'e3! find itself in danger. However,
The queen finds a way to slip Black has adequate play.
into the black position via the 30 ltJc6
squares left unprotected by 31 :a2 ltJd4
Bl ack ' s <111�ensi de pawn ad- If 3 1 . . .b4 then 32 �6 threat-
 Queen for Rook and Bishop Sacrifices 1 15

ens 33 ltc2.
32 l:a7
Here 32 liJxd4 llxd4 leaves the
h-pawn hard to defend since 33
g3 ltd l , threatening mate, is dan­
gerous.
32 liJxf3+? !
Evidently a winning attempt.
Black should prefer 32 . . . liJc6 !
when White has nothing better
than 33 ':'a2 liJd4 with a draw by
repetition.
33 'ii'xf3 b4
34 'ii'e2! Black must give up his bishop.
An excellent move. Now 39 :exe5
34 . . . ':'d4 35 'ii'b 5 or 34 . . . b3 35 Ivanov also mentions 39 ... ltb5
'ii'c4 b2 36 ':xb7 ! (Ivanov) are 40 l:hb7 l:hb7 41 'i'xe4 l:bS.
unsatisfactory. So Black must However, this sacrifice seems to
tread carefully. fai1. White plays 42 'ifd4 b3 43
34 lieS 'ii'b2 followed by placing his king
35 ':'a4 l:d4 on b 1 ; then his queen is freed to
36 ':a7 l:xh4+ ? ! chase away the black rook from
A lunch under dangerous the b-file when the b-pawn can be
conditions, as Nimzowitsch captured. Or the queen can be used
would have said. Ivanov points in combination with kingside
out that 36 . . . �e4 37 'ti'h5 �g6 is pawn advances to capture one or
better. White could also answer more of Black's kingside pawns.
36 . . . .i.e4 with 37 'i'e3 when the
battle continues.
37 <it>g3 ':'e4
38 'ii'd l ! l:d5
Here 3S . . . .i.d5 39 'ii'h 5 g6 40
�xh6 ':'xe5 was unclear accord­
ing to Ivanov. In fact, White has
chances with 4 1 l:hf7 ! �xf7 42
�7+ <it>f8 (42 .. .r.tf6 43 'ii'h S+
�f5 ? 44 'fIf8+ leads to mate in
two) 43 'fIhS+ <it>f7 44 'ii'xe5 ;
however after 44 . . . b3 the passed
pawn should save Black.
39 'fIbl (analysis diagram)
1 16 Positional Sacrifices

A sample variation: 43 . . . �h7 between the d5 and f5 squares, so

44 �f3 �g8 45 �e3 �h7 46 5 3 g4, taking away f5, looks
Ciki>d2 �g8 47 �c l g6 (this va­ right.
cates g7 and thus makes it possi­ 53 hxg4
ble to defend f7 with the king if 54 fxg4 ltd5
White's queen forces the rook 55 'iff4 liaS!
away from the second rank) 48 It turns out that the rook has
�b l �g7 49 'iVd4 b2 50 'ild6 safe, albeit unprotected squares.
IIb7 5 1 f4 �g8 52 g4 'lilg7 5 3 56 'ife4+ 'lilh6
'ilc6 and the rook has t o leave the 57 'iVf3 f6
second rank, as if the b-pawn Black could also play 57 . . . Ciki>g6
goes, Black is lost. So: 58 'ii'd 3+ �h6 59 'ifd2+ g5+ !
a) 53 . . J th4 54 f5 gxf5 with a draw.
(54 . . Jbg4 55 f6+ and 56 'iVe8 58 'iVe3 ltg5
wins) 55 gxf5 exf5 56 e6 ! 59 'iVxe6 �h7
(threatening 'ii'c 3+) 56 . . . lIe4 57 liz-liz
'ifxe4 ! followed by 58 e7 wins. Now the black rook can oscil­
b) 5 3 . . . lIb3 54 f5 gxf5 55 gxf5 late forever between e5 and g5
exf5 56 'iff6+ �g8 57 e6 ! IIb7 58 and White cannot break through.
'ii'd7+ and 59 e7 wins . A very instructive game.
c) 5 3 . . . ltb8 54 f5 gxf5 55 gxf5
exf5 56 e6! followed by 57 e7 8.5 Unpromising
wins as 56 . . .fxe6? loses the rook.
40 :'xb7 :'b5
Korchnoi-Stean
41 %le7 %leeS!
London 1 980
Now that the e5 pawn has been
eliminated, Black can sacrifice
his b-pawn and still achieve a
draw. The game continuation will
demonstrate that Black can set up
an impenetrable fortress on the
kingside.
42 :'xe5 ltxe5 43 'ii'xb4 ltf5
44 'ii'b 8+ Ciki>h7 45 'iff8 h5 46 13
�g6 47 'ii' h8 lidS 48 �h4 %l.f5
49 'ii'b 8 ltd5 50 .bl + �h6 51
'i'c1+ 'it?g6 52 'i'e3 IIf5
53 g4
After his fruitless manoeuvres,
White realises that only pawn Black has played the opening
advances can hope to break poorly and now finds himself un­
through. The black rook pivots der intense pressure on the
 Queen for Rook and Bishop Sacrifices 1 17

queenside. If his queen retreats to only weaken White's centre po­

d7 then 19 lLlc5 will win the b­ sition. This then is the set-up
pawn for nothing. Other queen Black is almmg for after
moves allow 20 �xc7. Therefore 1 9 . . . ihc6. He cannot really hope
Stean decides to sacrifice his for supernatural intervention, but
queen for rook and bishop, and if White potters about for the next
trust in the solidity of his posi­ couple of moves, then he will
tion. This is undoubtedly the best have time to carry out his plan of
course of action, since all alter­ securing the queenside from at­
natives are completely hopeless. tack.
18 Jtxe4
19 �xc6 �xc6
At first glance, things do not
look too bad for Black. All his
pieces are well developed, his
king is safe, and there are no ob­
vious pawn weaknesses in his
position. The knight on b4 seems
excellently placed to obstruct any
White attack on the queenside
pawns. However, the game con­
tinuation proves that Black's
queenside pawn structure is fa­
tally weak. Imagine if a genie (analysis diagram)
appeared at the chessboard and
gave Black three extra 20 'ii'c 1!
'positional ' moves. He would use Unfortunately Korchnoi is not
them wisely to play . . . Jt xg2 , a player to potter around. 20 'ifc l
. . . b7-b6 and . . . c7-c5 . Then assum­ threatens 2 1 .i.xh6 so Black has
ing we have allowed White to not got time to carry out his con­
play �xg2, we would reach the solidating plan above.
following position (D) : 20 hS
Black no longer has any wor­ 21 .i.xc6 lLlxc6
ries. His queenside is rock solid. The knight, temporarily at
There is no way that the b6 pawn least, has to relinquish the b4
can be attacked. If the white square, since 2 1 . . . bxc6 ruins the
queen landed on b5, Black could pawn structure.
simply play . . J:td6. Nor has 22 'iVc4!
White any hope of a break­ Now we see the point of
through on the kingside. There Korchnoi 's plan beginning with
are no weak points to attack 20 'ii'c l . He homes in on the
there, and a move like f4 would weak b7 pawn before Black has
1 18 Positional Sacrifices

the chance to play . . . b7-b6 and back a rook with an easily win­
. . . c7-c5 . ning position .
22 kte6
Bolstering c6. 8.6 Conclusion
23 ktel b6
24 'iib S :dd6 What conclusions can we draw
2S 'iia 6! from our examination of queen
Threatening to win the c-pawn for rook and bishop sacrifices?
with 26 'iic 8+. Black has no good To be effective, the player with
defence. For example, 25 . . . I:.e7 the rook and bishop must have:
26 :xc6 ! lhc6 27 'iVa8+ wins; or i) A safe king (absolutely es­
25 . . . c;i;>h7 26 'iib 7 and the pawn sential) .
falls. ii) A solid pawn structure with
2S �d4 no chinks.
26 'iic 8+ c;i;> h7 iii) A well co-ordinated and
27 .i.xd4 exd4 self-defending force.
28 ':'xc7 Also, the queen works very
Now Black's game crumbles. well with passed pawns. In fact, a
There is no hope of a blockade on queen and passed pawns often
the queenside. If now 28 . . . l:Ixe2 overcome a much larger army.
29 .uxf7 threatens 30 'iif8. These are of course only gen­
28 ktf6 eral rules. As always, the impor­
29 'ti'b7 :de6 tant thing is to examine the indi­
30 �f1! 1-0 vidual position in front of you
Simplest. Black is denied any and decide whether in the specific
counterplay at all. Now 30 . . . �g8 circumstances the sacrifice is
3 1 ktxf7 ! lhf7 32 'iic 8+ wins sound.
9 It' s You r Tu rn

The reader is invited to examine (Become aware of the tactical

20 positions taken from recent features of the position, e.g. a
master games. Imagine you are queen and rook separated by the
the player to move. Ask yourself: distance of a knight fork - can
i) What are the important fea­ this be exploited?)
tures of the position (for example, iv) If possible, is the positional
a safe king, a strong centre or a sacrifice tactically sound? Is it
weak pawn structure)? Are my equal, good or winning?
pieces well co-ordinated? What There are clear answers to
about my opponent's pieces? most of the puzzles. However,
ii) Do I stand better or worse? one cannot expect complete clar­
What is my correct plan? ity; otherwise we would be in the
iii) Can I make a positional realm of combinations rather than
sacrifice to implement this plan? positional sacrifices.

1 . White to play 2. White to play

1 20 Positional Sacrifices

3. Black to play 6. White to play

4. White to play 7. White to play

5. White to play 8. White to play

 It 's Your Turn 121

9. Black to play 12. White to play

10. White to play 1 3 . Black to play

1 1 . White to play 14. White to play

1 22 Positional Sacrifices

1 5 . White to play 1 8. White to play

1 6. White to play 1 9. Black to play

1 7 . White to play 20. White to play

 It 's Your Turn 123

�g4 or 26 'i'xe5 ..td6) 26 ... e4 27

b4 l:lxd5 28 'i'b3 (trapping the
Answe rs
rook on a2, but Black is so active
that he gains a decisive kingside
1 . 24 e5! is a thematic break­ attack) 28 . JId3 29 'i'xa2 ..td6
.

through. Now after 24 . . . dxe5 25 30 h3 (if 30 g3 ..txg3 or 30 .. J�d2

d6 'ile8 (25 . . . 'iWd7 26 dxc7 ! 'ii'xd2 gives a winning attack)
27 cxd8('iW) ! wins) 26 'ifg2 .i.b8 30 JIxh3+ 31 'itgl
•• 'ilh4!
27 �b6 l'la6 (27 . . . :'a7 28 �b5) (cutting off the king' s escape)
28 �xc8 lhc8 29 'iWxb7+ and White resigned since 32 gxh3
wins. So Black tried 24 ...fxe5 25 'ii'g 3+ mates. (Miiller-Hort, Ger­
tDe4 (a beautiful square for the many 1 994)
knight, vacated by 24 e5 ! ) 25 ... h6
(preventing 26 �xg5) 26 �xe5! 4. Should White capture the
and Black resigned. After rook or the queen? Actually,
26 . . . dxe5 (or 27 :f7+ wins) 27 there isn't much choice, since 20
d6 'ii'e 8 (27 . . . it'd7 is the same) ii.xc8 :'cxc8 leaves White unable
White has a choice of wins, e.g. to save his d-pawn after
28 .i.xg8 �xg8 29 �f6+ or 28 2 1 . . .lHd8. So White must make a
:f7+ "ilxf7 29 .i.xf7 �xf7 30 virtue out of necessity and play
'ii'f2 + and 3 1 dxc7. (Glek-Dreyer, 20 dxc6! . After 20 'ii'c 7 21
.••

Hamburg 1 993) ':xc5. White has a rook, knight

and strong passed pawn for the
2. White in fact has a forced queen. Moreover, his position is
win: 45 1lxa5! bxa5 46 :b7 'iWc8 solid, his king safe and his pieces
47 ':xt7+ �h8 48 'i'c1 ! and mate well co-ordinated: all the criteria
next move cannot be averted. 1 -0 we adjudged necessary for a suc­
(Maiwald-Masserey, Switzerland cessful queen sacrifice in Chapter
1 993) 8. Evidently there is more virtue
than necessity about White's
3. There is no forced win here, position. This does not of course
but 21 ... b5! undermines White's mean that he must win, although
disorganised queenside. White in the game he did so: 21 .. ,'ii'd 6
did not want to be left with iso­ 22 b4 g6 23 ltdl d3 24 ii.h3
lated pawns after 22 . . . bxc4, so he ii.d8 25 �5 d2 26 It:)c4 (one is
accepted the offer: 22 cxb5 axb5 reminded of Capablanca's adage
23 axb5 'ua2 24 'iVc3. This seems that a passed pawn is either very
to hold things together. However, strong or very weak; Black's
a second pawn sacrifice released passed pawn looked threatening
all the stored up energy in but it lacked effective defence)
Black's position: 24 d5! 25 .•• 26 Ji'e7 27 :'xd2 'it'el+ 28 .i.f1
.•

exd5 It:)f6 26 �ba3 (26 lhe5 ii.b6 29 �xb6 'iWxd2 30 c7 and

1 24 Positional Sacrifices

Black resigned. (Van Wely­ 6. Black has just played

Jonsson, Iceland 1 994) 32 . . . J.b7-c6 exposing an attack
on White's queen. Should White
S. The immediate 20 J.xc6 play 33 'ifxbS or 33 l:txc6? 33
':a7 is nothing for White (and 'fixb8 llxb8 34 ':xbS is two rooks
don't imagine you can sacrifice for the queen, but after 34 . . . �bS
your queen after 20 �xc6 lla7 2 1 (Shirov) there are no immediate
'ilxa7 'ilxa7 22 cxbS axbS 23 targets for White's rooks and
�xbS : we have the familiar rook, Black has the makings of a king­
bishop and passed pawn for the side attack after the bishops are
queen, but the passed pawn is a exchanged and he plays . . . ltJfS.
weakling on a2, and Black can So in the game Bareev-Shirov,
easily gain strong counterplay Tilburg 1 993, White correctly
before it gets very far). We have preferred to sacrifice his queen:
already seen some effective ex­ 33 l1xc6! ':xb6 34 l:tbxb6. An
change sacrifices on e6, and this interesting decision. 34 axb6
is another example: 20 IIxe6! . gives White a strong passed
Now if 2 0. . . fxe6 2 1 'ilxe6 + 'it>f8 pawn, but then 34 . . ....e7 3S ':c7
(2 1 . . . c;t>g7 22 �xc6 attacks e7) 22 'ifgS followed, after the e-pawn is
�dS and wins - 22 . .. �eS 23 defended, by . . . ltJfS, leads to un­
�xc6+. So Black tried 20 lh7 ••• clear play. Bareev prefers to
21 Axg6+! fxg6 (2 1 . . .�f8 22 bring both his rooks into play and
'ilh3 fxg6 23 'ilhS+ �f7 24 J.dS keep Black tied up. 34 .:td8 35
••

mate, or 2 1 . . .�h7 22 'ilh3+ ! ltJg2 "'e7 36 �xa6 ltJf5 37 ':c5

<itxg6 23 �e4+) 2 2 'ile6+ �g7 (preventing 37 ......a3) 37 .:th8
••

23 �xc6. For the exchange, (Black strives for counterplay

White has two pawns with the before the a-pawn begins to ad­
promise of more. But more im­ vance) 38 ':b7 "'b6 39 :b6 "'g5
portantly, he has destroyed 40 lhd5? White's last couple of
Black's pawn centre and opened moves have shown the uncer­
up his king for attack. The game, tainty of time pressure, and now
Karpov-Topalov from Linares he commits a blunder. Shirov
1 993, finished: 23 lId8 24 cxb5
•.. gives 40 ltJf4 (with the threat of
�f6 25 �4 �d4 26 bxa6 'ifb6 41 :xg6+ ! ) 40 . . . :h6 4 1 IIxdS
27 ':dl 'ilxa6 28 l:txd4! IIxd4 29 'fIe7 42 IIb3 ltJxe3 ! 43 llex3 gS
'ilf6+ <it>g8 30 'ifxg6+ 'it>f8 31 as unclear. 40.. ...h5! . Not only
.

'ife8+ <it>g7 32 'ife5+ 'it>g8 33 attacking h2 but also preparing to

ltJf6+ 'it>f7 34 J.e8+ <it>f8 35 penetrate on 4 1 or f3. That is why
"'xc5+ 'ifd6 36 "'xa7 'ilxf6 37 White had to prevent this move
�h5 ':d2 38 b3 Ab2 39 <it>g2 1-0. with 40 ltJf4. 41 h4 "'f3 42 Ibf5.
A rather aggressive variant of the Desperation. Otherwise his king's
Karpovian exchange sacrifice! position crumbles. 42 gxf5 43
•.•
 It 's Your Turn 1 2S

:'b5 "xg3 44 :'xf5 :.xh4 45 sive) 28 ...... cS! 29 exdS?! liJd4

.tc8 :'g4 46 :'f2 ltg5 and White (now White's queen cannot re­
resigned. 47 . . ...xe3 or 47 . . . :'xaS turn to the defence of the king) 30
follows. As we remarked in the dxe6+ fxe6 31 "eS ..txd3+ 32
introduction, often the decisive �al "'c3+ 33 ..tb2 liJc2+ 34
factor is not the inherent strength �bl 1De3+ 35 �al 'i'xeS 36
of the sacrifice, but rather how Ilxd3+ "dS 37 :txdS+ lDxd5 38
well the player handles the result­ ..teS .td6 39 .tb2 lDxf4 40 .tf6
ing position. lDdS 0-1 (Arachamia-Kotronias,
Crete 1 993). Another reminder
7. White played 17 .txb5+ that you can't play positionally
axb5 18 liJdxbS. The idea is no­ unless you check all the tactical
ble: after the queen moves, say details.
18 . . :i'c6, 1 9 liJxd6+ .txd6 20
"xd6 "xd6 2 1 lhd6, White has S. This position arose in Kas­
three passed pawns for a piece. parov-Short, Novgorod 1 994. The
These outweigh the piece and the world champion played 18 fS!
tactics also favour White: opening lines and breaking up
2 1 . . .liJb7 (hoping for 22 l:lb6 Black's centre. 18 ...... h6+
liJxf3) 22 .td4 ! ? liJxd6? 23 .txeS ( l 8 . . exfS 1 9 'i'xdS+ ..n 20
.

wins or 22 . . . liJxf3 23 liJxh8 liJxd6 "'xaS "xa2 2 1 :a3 wins, or

24 ltg3 traps the knight. How­ 1 S . . . 'i'xfS 1 9 :'f3 "g6 20 :'xfS+
ever, White has overlooked a fufS 2 1 lDb6 ! wins a piece) 19
tactic right at the beginning of her �bl :txfS 20 ]:tf3! :'xf3 21 gxf3.
combination: 18 liJcd3+! 19
••• Black is tied up. He cannot co­
�bl ( 1 9 cxd3 :'xbS and ... liJxf3 ordinate his pieces, mainly due to
is threatened) 19 ......d7 20 f4 the looming threat of lDb6.
:'xbS 21 liJxbS "xbS 22 cxd3 21 "f6 22 .th3 cj.Jf7 23 c4! .
•.•

liJc6 (but not 22 . . . liJf3 23 'ii'c 3 I ) . Another blow from the winds that
Black i s at least equal. He has undermines Black's centre. Kas­
avoided the exchange of queens, parov continuously strives to
so that White 's passed pawns open the position. 23 ... dxc4
cannot be utilised - they must (23 . . . d4 24 f4 ! exf4 2S :'xd4 eS
keep back to guard White's king. 26 'ildS+ and 27 'i'xa8 wins a
Meanwhile, Black's centre is piece; an aesthetic continuation:
'bomb proof' . In the game the 1 8 fS ! , 23 c4 and 24 f4 ! ) 24 1Dc3!
minor pieces proved more effec­ 'ike7 25 "'c6 :tb8 26 1De4 lDb6
tive than the rooks: 23 lthl l::tx hl 27 liJg5+ <ii;>g8 28 "e4 g6 29
24 :txhl .ta6 25 l:dl �d7 26 .xeS 1Z.b7 30 :td6 c3 31 ..txe6+
"c3 .te7 27 .tel dS! 28 'fIg7 (a ..txe6 32 :txe6 1-0
dangerous adventure but 28 eS
liJb4 29 d4 would be very pas- 9. White has a wretched pawn
1 26 Positional Sacrifices

structure - two sets of doubled pawn. But I wish to point out an­
pawns and another two isolated other draw: 1 .ixf6 "'xf6 2
pawns - to worry about. How­ ltJxhS! �xh5 3 lIe3 with a
ever, he is attacking two black blockade.
pawns on e6 and g7, and the de­
fensive 24 . . . "'c5+ 25 'it>h2 (25 1 1 . In the game Anand­
1IVd4 ! ?) 1IVe7 26 "'d4 'it>b8 27 a4 I.Polgar, Linares 1 994, the Indian
is unclear: White's queen is ex­ grandmaster played 19 ltJxe6!
cellently placed and he has at­ fxe6 20 1IVxe6. The g6 pawn is
tacking chances on the queenside. now doomed so White acquires
So Black played the crafty three strong pawns, which
24 .. :i'c7! when White took the e­ threaten to race up the board, for
pawn: 25 "'xe6+? Instead, he the piece. Black needs to use her
should try 25 :te l or 25 'i'd4, extra piece to generate some
with only slightly worse chances. compensating attacking chances
2S ......d7! . Now the exchange of against White's king. Unfortu­
queens is forced, and despite nately, the insecure position of
White's extra pawn he has a lost her own king hampers her coun­
position. His flimsy pawn struc­ terattack. 20 'it>b8 (20. . . i.g7 2 1
•••

ture cannot resist the attack of .td4 .txd4 22 :'xd4 leaves al­
Black's rook and queen. The most all Black's pawns hanging)
game concluded: 26 'ilxd7+ (26 21 'i'xg6 :'h3 22 �bl ! (22 .td4
'ii'h 3 'ilxh3 27 gxh3 :'f3 is just as ltxc3 ! 23 .ixc3 1IVxa2 and
bad) 26 'it>xd7 27 l:tbl b6 28
••• 24 . . . ltJc5 gives Black good coun­
llel 'it>e6 29 a4 lIfS 30 lite3 lIxeS terplay; White's game move
31 'ittf2 (the pawn endgame is avoids this counter-sacrifice since
hopeless, but so is 3 1 lIf3 lie 1 + if 22 . . . :txc3 23 i.e 1 ! regains the
32 'it>f2 :ta l ) 3 1 ...l::txe3 32 'it>xe3 material without creating a weak­
hS 33 �f4 as! (drawing back the ness) 22 ...:'f3 23 .id4 :'xf4 24
white king with the threat of "ilfh7 (Black has broken up
' " b7 -b5) 34 'it>e3 'it>eS 35 'itt d2 White's pawn phalanx but the g­
'it>e4 36 �e2 d4 37 cxd4 �xd4 pawn is now unstoppable except
38 �d2 c3+ 39 �e2 'it>c4 40 'it>e3 by sacrificing the f8 bishop,
'it>b4 41 �d3 'it>xa4 42 �xc3 which will leave White a pawn up
�bS 43 �b3 'it>cs 0-1 (A. Nunez­ with a more compact position)
McDonald, Andorra 1 99 1 ) 24 dS (a desperate bid for activ­
..•

ity) 25 g6 .ics 26 .ixcS ltJxcS 27

10. White can draw with the g7 (now the passed pawn will
simple 1 .!be4 i.xc3 (forced) 2 cost a rook) 27 ... �a7 28 g8('i')
ltJxd6 i.xe 1 3 f3 ! intending 4 l:txg8 29 1IVxg8 ltJxe4 30 ltJxe4
ltJe4 and 5 ltJxg5 ! leaving Black Ibe4 31 1IVg7 "'cS 32 llgel l:tf4
with the bishop and wrong rook's 33 1J.e7 "'b6 34 1IVgS c3 35 b3
 It 's Your Turn 1 27

'ii' b4 36 l:tc7 lla4 37 'Wxd5 Wb8 reaches h I he draws, since Black

38 'i'd8+ �a7 39 a3 'ii'xa3 and has the wrong rook's pawn) 60 h4
1-0 before 40 �xc3. �xh4 6 1 �f3 'itth3 62 'itt f4 �h4
(White intended 63 g6 ! hxg6 64
12. The position occurred in 'itt g 5) 63 �f3 'it,(h3 64 'it,(f4 .i.d4
Karpov-Yusupov, Tilburg 1 993. 65 �f3 and I couldn't see a win
White has a clear advantage and (I still can't). So I played 56 . . . a6.
can even win the exchange im­ Then 57 h4 'itt f5 58 'ittd5 .i.xh4
mediately with 47 ttJxa7. How­ 59 �c6 .i.xg5 (59 . . . 'itt xg5 60
ever, the danger is that the posi­ 'it>b7) 60 b7 .i.f4 6 1 Wb6 �e6
tion will become so blocked that (6 1 . . .h5 62 'ittxa6 h4 63 �a7 h4
White's rooks will be unable to 64 b8(,i') and the a-pawn reaches
penetrate into Black's position. a7 after Black queens, with a
So Karpov eschewed the win of draw) 62 �xa6 �d7 63 �a7 �c6
the exchange and instead played 64 b8('i') .i.xb8+ 65 'ittxb8 �b5.
47 f6! . This ensures that the f-file It seems as if White is lost, since
is opened, since if 47 . . . :b7 48 his king is outside the 'square' of
�f1 followed by 49 h7, 50 :h5 the h-pawn. But by making a
and 5 1 ':xg5 will win. 47 ... .i.xf6 feint to support the a-pawn,
48 l:.f1 .i.h8 49 ttJxa7 winning, White snatches a draw: 66 �b7 !
but 49 'u'hf2 is simpler, e.g. �xa5 67 'itt c6 and a draw was
49 . . . f6 50 h7 'itt fl 5 1 ttJxa7 .uxa7 agreed. After the game, Colin
52 �xf6+ .i.xf6 53 h8('iV) . Crouch asked me why I did not
49 .. Jha7 50 l:.h5 (still 50 Ithf2 play 56 i.xb6! . It turns out to be
•••

wins) 50 �e7 51 :txg5 :a8 52

..• more important to keep a queen­
h7 f6 53 l:g8 ':fS 54 c5! (this side pawn than the bishop. Black
breakthrough method will by now wins by a tempo: 57 axb6 axb6
be familiar to the reader) 58 <it>b5 �f4 59 �xb6 �xg5 60
54 . dxc5 55 'it>c4 �f7 56 d6!
.. <it>c5 �h4 61 �d4 �xh3 62 <it>e3
cxd6 57 :xfS+ 'itt xfS and Black 'itt g3 ! (62 . . . 'itt g2 63 �f4) 63 �e2
resigned without waiting for the 'itt g2 64 �e3 h5 65 'itt f4 h4 and
inevitable 58 'itt d 5, 59 �c6 and wms.
60 <it>xb6.
14. Apparently Black has a
1 3 . The author reached this good game: a solid position, well
position as Black against developed pieces and a threat
R.Wynarczyk at Whitby in 1 99 1 . (22 . . ....xc2+). 22 ttJe7+ does not
I looked at 5 6 . . . axb6 and saw it save White: 22 . . . ':'xe7 23 'Vikxe7
was only a draw after 57 axb6 'ii'x c2+ 24 'itt a l ttJc6 ! and White
.i.xb6 (57 . . . 'itt f4 58 b7 i.a7 59 loses his queen because
�d3 is no better) 58 �d3 <iti>f4 59 25 . . . .i.xb2 mate is threatened. So
'itt e2 'itt g 3 (if White's king White should cut his losses and
128 Positional Sacrifices

play 22 'iVxf8+ :'xf8 23 liJe7 king is a portent of doom, so

mate (Van Mil-Reindennan, White surrenders the pawns to
Wijk aan Zee 1 993). We repeat: break up Black's kingside)
tactics come before strategy. All 31 ...:tb4 32 ttJd4 �xd4 33 �xd4
generalised, verbal arguments cut llxbS 34 <it?e4 r3;c7 35 �c3 r3;xc6
no ice if it's mate in two. (White and Black gradually exploited his
had just played the cunning 2 1 material advantage to win on
'ii'd2-b4 and Black fell for it with move 59.
2 1 . . . 'ii'c6?? )
16. 3S g4! gxh4 (or 38 ... hx.g4
1 5 . This position arose in the 39 h5 lit>e6 40 fxg4 f5 - or else
PCA Candidates match, Tiviakov­ White plays his king to e4 and f5
Adams, New York 1994. Rather - 4 1 h6 'iitf6 42 gxf5 and wins) 39
than play 1 8 �xb4 with equal gxh5 'ite6 40 �g2 eM5 41 f4! !
chances, White tried the enterpris­ (the point) and Black resigned
ing IS 'ifxb7?! sacrificing the ex­ sin�e 4 1 .. .'iif e 6 42 �h3 wins.
change for a couple of pawns.
IS ... ttJc2 19 llxaS ( 1 9 .txa5 'iVd3 17. Black has just played
20 'Wixe7 'iVxb3 is good for Black) 22 .. J:ld8-d7? which gave White
19 ...ttJxe1 20 :'xaS 'iVxaS 21 the chance of a breakthrough on
'ii'xaS (of course 21 'ifxe7?? 'iVxg2 the kingside: 23 llxcS+ ttJxcS 24
is mate) 21 ...:'xaS 22 �xel. f5! exfS 25 'iWh6. Note that this
White's passed pawns look dan­ penetration is only possible be­
gerous, but his pieces are loosely cause Black's knight has been de­
placed and until they can be rede­ flected from its role of defending
ployed, they cannot give the the f5 square. 25 J1c7 26 ':'f2! .
••

passed pawns much support. One The threat of ttJg5 hovers over
of White's problems is that his Black and interferes with the co­
other rook has been exchanged. ordination of his pieces. For ex­
Therefore Black's remaining rook ample, 26 . . . ttJe7 27 ttJg5 would be
has no rival. Adams exploited this decisive, since the rook on c7 no
in incisive style: 22 ':a4! 23 cS
•• longer defends f7, so mate in two
�c4! 24 c6 (if the passed pawns with 28 'ii'h7+ is threatened
were well supported, they would (27 ... ttJc6 28 e6 ! ) . Black's one
become more dangerous with chance is if he can gain counter­
every advance. Instead, they are play against White's king with
targets for Black's pieces) . . . 'i'xd4 in reply to ttJg5. So White
24 ... ttJdS 25 ttJn 'iUS 26 ttJe3 does not play ttJg5 but rather safe­
ttJxe3 27 fxe3 'itteS 2S �c3 .tcS guards his king and prepares an­
29 'ittf2 �b6 30 � �dS 31 other breakthrough: e6. A good
�xg7 (White's passed pawns are example of a threat being stronger
stymied and the arrival of Black's than its execution. 26 .a6. Black
••
 It 's Your Turn 129

cannot see any way to improve his gxh3 Ilxf3+ 32 'ii'xf3 lhf3+ 33
position. 27 lie2! 'iWbS 28 � �e2 :'xh3 0-1 (Costa-Gavrikov,
'WIid3 (once again hoping for 29 Switzerland 1994)
�gS 'ii'xd4+) 29 e6. Now Black's
kingside crumbles. 29 ...�d6 30 20. Black needs one more move
exf7+ lhf7. Of course, 30 . . . �xf7 for comfort: . . . .tg7, adding an­
3 1 ttJeS+ wins. 31 'ii'xg6+ ltg7 32 other defender to the knight on f6.
'WIixd6 h4 33 'ii'dS+ �f7 34 lbeS+ But it is White's move and he
1-0 (McDonald-Hastings, London struck immediately with 22 cS! . If
1 993) the bishop were on g7, Black
could now play 22 . ttJxcS safely.
..

I S . White played 27 dS! exdS But as things stand, both

(if 27 . . . :'xdS 2S 'iYeS+ wins) . The 22 . tLlxcs and 22 . . . dxcS could be
..

sacrifice has a triple purpose, as answered strongly by 23 fxeS . So

soon becomes clear: Black can only accept the offer
i) Open the e-file. one way, and this leaves the
ii) Give the knight safe access queenside severely weakened:
to fS. 22 ...bxcS 23 fxeS dxeS 24 tLlc4 (a
iii) Close the diagonal as-h i . beautiful square for the knight,
You could hardly ask more aiming both left, to as, and tight,
from one move ! Play concluded to eS) 24. �g7 (the threat of 2S
..

28 l:te8+ l:txe8 29 'iWxe8+ (point ttJ4xeS obliges Black to overpro­

i) 29 ... �f8 30 ttJh6+ �g7 31 tect the knight on f6, but in doing
ttJfS+ (point ii) 31...<it>gS 32 so the square d6 is underprotected;
:a3 ! . Point (iii) ! The rook swings White's next move exploits this
over and mates by 33 l:.g3+. If fact to eventually win the ex­
the as-h I diagonal were open, change) 25 �c3! (much better
Black could counter with mate in than the natural, obvious and infe­
three beginning 32 . . l:.c l+. (Los­
. rior 2S tLl4xaS) 2S .. J:te6
Laffler, Amsterdam 1 994) (defending against 26 �xaS �7
27 ttJd6, but now disaster strikes
1 9. Black has a lot of firepower on a different square) 26 �xa5
directed against White 's f3 pawn. 'iib 7 27 ttJdS l::txd8 28 i.xd8
One feels that a slight increase in 'iWxbS 29 'iWd3 hS 30 ktbl 'iVa6 31
the pressure will break down the l:tfe1 h4 (Black battles on, but the
defences. This is achieved with material deficit is fatal) 32 !ib3
the sacrificial entrance of Black's 'ikc6 33 �xf6 ttJxf6 34 llb6 'iic8
bishop : 29 ...eS! 30 dxeS iLh3! 35 :'xe6 'iixe6 36 a4 .th6 37 as
undermining the f-pawn 's de­ ttJhS 38 'iVdS 'ii'g4 39 'iixf7! 'it>h8
fender. White ' s position now 40 ttJxeS ttJf4 41 'ti'xf4 1-0
collapses, since 3 1 l:.hg 1 .t xg2 (Ftacnik-Gedevanishvili, Sydney
and 32 �xf3+ is hopeless. 31
. . . 1 99 1 ) .