Simulation vs abstraction in game design

This is an excerpt from Game Architecture & Design, an industry textbook I co-authored with Andrew Rollings. (I wrote the game design bits, Andrew dealt with code, tech and development practices.) The book was originally published in 1999 and a revised edition came out in 2004. In the intervening two decades, a lot has changed, but it's also interesting to see what hasn't...

If I throw a ball and take many high-speed photographs of its flight, I'll see that the trajectory the ball took is a parabola. But the ball didn't follow that path because gravity told it: "Move in a parabola." A parabola is just a symbolic concept in the analytical domain of mathematics, and the universe doesn't know anything about mathematics or analysis or symbols; these are human concepts. In reality, there are just a bunch of physical processes, each of which deals only with the processes and circumstances just before and just after it. So, the ball is at one position, and gravity tells the ball's velocity to change, and the ball's velocity tells its position to change. The balance between kinetic and potential energy over the time the ball is in the air gives you what we call a parabola.

This is the opposite approach to that taken in most software applications. There, processing power is at a premium, so the sooner you can go to symbolic modelling rather than step-by-step simulation, the better. The tradeoff is that software can crash when your symbolic "shortcut" misses something that the one-step-at-a-time approach would have taken in its stride.

Researchers in Artificial Life have identified an analogous problem:

"The classical AI approach has been criticized because the symbols and symbol structures on which planning and decision making are based are not grounded in the real world. The problem is that unequivocally decoding sensory data into a symbol and turning a command without error into its intended action may be unsolvable."

- Luc Steels, "The Artificial Life Roots of Artificial Intelligence" in Artificial Life (MIT Press, 1997)

One big advantage of the way that reality does things is that the universe, being non-symbolic, cannot crash. As an example of the principle at work in a game, suppose I am putting a monster into my new Frankenstein adventure and the idea is that it will jump out of its vat when the player enters the laboratory. Instead of putting in a lot of complicated AI to do with detecting humans and having the goal of wanting to kill them, I just choose the short cut of placing a trigger tile inside the laboratory door. When the player steps on the trigger, the monster will appear and attack.

Okay so far, but what if the player manages to get onto the tower roof, jumps down, and, by some fluke, manages to land safely on the balcony of the laboratory? Now they can explore the lab, get all the power-ups, and read the journal about the monster (an entry that is supposed to be poignant if they've just fought and killed it, but that is meaningless otherwise). Only when the player goes to leave via the door does the monster climb out of its vat and growl, "You shall not steal my master's secrets!"

In the past, the nonsymbolic, step-by-step approach was not practical. The the processing capability wasn't available to deal with that and graphics too. But now much of the graphics work is done by the video card, and computers are doubling in power every eighteen months or so. At last, it is starting to be possible to create "uncrashable" games by avoiding the need to design using symbolic shortcuts.

Comparing Nonsymbolic And Symbolic Design

In the original Warcraft, peasants collected gold by entering a gold mine and bringing sacks back to your town hall. At the start of the game it was always worth spawning peasants because, the more peasants you had, the greater your revenue stream. However, there came a point when the peasants started to get in each other's way. Adding more peasants would then lead to “traffic jams” as the peasants encountered each other on the streets of the town and would have to back up to let others get past. The situation was alleviated by leaving wide streets. Additionally, it was not a good idea to place your town hall too close to the gold mine – giving a little more space also helped avoid traffic congestion.

Now, an economist could derive an equation to describe the flow of gold to the town hall. The factors would be the number of peasants, the placement density of the town buildings, and the distance from the town hall to the mine. We can imagine that it would be a pretty complex equation. The point is that the designers of Warcraft never needed any such equation.* They simply programmed in the basic rules and behaviours and the economic simulation emerged directly from those.

Contrast this with a game like Caesar II, which used underlying equations to create a simulation of an ancient Roman city. This approach is less satisfying because the player is not directly viewing the reasons for success and failure. Instead, when playing a game like Caesar II (or any simulation of its type) you are trying to build an abstract match to the game’s underlying equations in your head. The simulated economy and the gameplay are less visible, lessening the sense of immersion.

And you know what? The same goes for stories. If you construct them from symbolic forms (arcs, paradigms, act breaks) you'll end up with less robust and varied stories than if you allow each micro-event to trigger the next and see where it goes. Which is why in roleplaying terms I'm a simulationist rather than a narrativist. Hey, if it's good enough for reality then it's good enough for me.

* This gives me an excuse to digress onto the topic of AI. Foundation models (or indeed any deep neural net) are sometimes referred to as algorithms. I find that term misleading. In principle you could express all the weights of a billion-node net in the form of "an algorithm" but that's not really an accurate way of talking about what the AI is doing in, say, ChatGPT, which is akin to (though much more complex than) the peasants collecting gold in Warcraft. That too is governed by multiple algorithms (for route-finding, collision detection, etc) but it would be more accurate to talk of it as a model. An algorithm could be derived to express the rate of gold production in terms of all those variables, but the Warcraft system doesn't have that algorithm built in, and nor do AI systems. There is an example here, where the article refers to "a separate algorithm" where they really mean " a separate model".

Principle of Least Action image by Maschen CC0, 

A toolkit for open-world gamebook design

The earliest true gamebook (with rules, that is, and disregarding mere multiple-choice texts) was probably Steve Jackson's Death Test in 1978. That allowed players to backtrack and visit the same location multiple times, but throughout the 1980s the main evolution of gamebooks detoured down a different branch with series like Fighting Fantasy, where the arrow of narrative time pointed in one direction. It wasn't until 1995 that the first open-world gamebooks appeared -- followed by silence until, decades later, series like Steam Highwayman and Legendary Kingdoms took up the torch. Finally it seems that open-world gamebooks might be becoming the dominant form.

I have been sent a succinct and comprehensive toolkit for open-world design that would-be gamebook writers will find very useful. Not just would-be writers, either; I've been doing this since the last millennium and even I found it had some handy tips. You can find that toolkit here. The author prefers to be known only as Ruth, so I will respect her anonymity, though it can only be a matter of time before her own open-world gamebook series appears.

In the Vulcanverse books, when I used what Ruth calls Local Changes, typically I'd include another entry:

So you’d arrive at the location and be presented with the situation that could lead to a change. If you complete that correctly you’re told to turn to the original location and tick the box. Then you get a one-off scene that explains the change and assigns any relevant keywords. If you return to the location subsequently you go to X*A which is the location in its changed state.

Regarding keywords, it’s worth differentiating (and incidentally I say this with 20/20 hindsight) between keywords used as logic flags to denote that an event occurred (“a tidal wave hit the city”) and keywords used to indicate a current state (“the city is in ruins”). In Vulcanverse I could have used keywords and titles respectively for those two functions. I didn’t, which meant that sometimes I’d remove a keyword that was no longer applicable, only to realize later that I’d still like to know if the player had originally had that keyword. Eg if I complete a quest for the king and so he makes me the court champion, even if later on I lose the court champion keyword I still need to have the keyword that says I did the quest. (I'll discuss flags and states in more detail in a future post.)

Regarding rules mechanics generally, I keep promising that next time I’ll create an object-oriented rule set so that I could keep track of, say:

• having one’s face stolen by a mujina
• getting cursed by a fairy to have a donkey’s head
• acquiring blue skin
• being body-swapped into another form

Variously those conditions could mean:

• You’re not recognized by somebody who should know you
• You’re mistaken for someone else
• You are treated differently from other people (potentially in several different ways)

In Fabled Lands I ended up filtering for all those statuses every time I asked the player to check them, whereas if they had been properly organized I could simply have asked, “If you have a status that is listed as transformative and uncanny, go to X”.

Similarly with equipment, where it is useful to have a generic way of identifying whether something can be used as a weapon, or has a blade, or reflects like a mirror, or is made of iron, and so on – rather than having to list the items for each applicable function (such as “if you have the polished shield, the silver mirror, the jade looking-glass…”). Ruth's Multi-Function Item tool covers that one, and could be extended to keywords describing the character’s current status to solve problems like the altered-face one above.

Vulcanverse also used the Current Location feature, a box on your adventure sheet where you record a section number that you are routed to after the current subquest you're on is complete. This allows for encounters to happen at different points around the world, for instance when you're acting as go-between carrying love tokens from Eros to Psyche in The Hammer of the Sun. The Current Location can be used for a lot of things, a common example being when you have a wandering NPC whom the player has to meet up with from time to time, because the different stages of the player's dealings with the NPC don't have to be tied to the geography of the world. You just tell the player to record a Current Location number when the subquest is triggered, then when the next part of their interaction with the NPC (or whatever) is done, you send them back to the Current Location and they resume their travels.

I was going to compile a list of gamebook design resources, but they are many and scattered far and wide. Here are a few to get started. If you know of any others, link to them in the comments.

• "Writing gamebooks" on Lloyd of Gamebooks
• "Wamt to write a gamebook?" on Lloyd of Gamebooks
• "Fetch quests" in the Steam Highwayman books
• "Ambush sequences and the noted passage mechanic" in Steam Highwayman -- which, now that I've seen the article, I realize features the same tool that I call current location (Martin's nomenclature describes it better!).
• "Learning from better gamebooks" on Rand Roll

And while we're on the subject of gamebooks, the Blood Sword 5e hardcover book has just been released in the UK and US. You can buy it at Orc's Nest or from GMS Tabletop Games. The book is a complete 5e roleplaying campaign based on the gamebooks and features some absolutely top-notch design, writing, artwork and production. If you're starting to think about Christmas presents, it's the perfect gift for the avid gamer in your life. Just be aware: they're going to need a bigger stocking.

A hopeful monster

Want to see something genuinely original? There's a small catch: it's not a spectator sport, you have to join in to make it work.

Still interested? Then take a look at Gamete on the Flat Earths gaming blog. And if you want to take part, just leave your variant rules in the comments. Between you and me, it could go anywhere.

A Thunder of Dragons

Tension and excitement fill the room as A Thunder of Dragons begins! Players take on the role of these mighty flying reptiles, soaring above a sprawling 15 by 15 playing board filled with raging rivers, perilous mountain ranges, treacherous swamps and dark forests. Castles, villages, towns, and abbeys are dotted about the game board. Such settlements offer rich pickings for the dragons. But beware! These havens of prosperity are guarded by garrisons of bowmen, knights, foot soldiers and wizards alongside powerful heroes. Bigger and richer settlements are even more fiercely defended. Players swoop in to pillage these strongholds for their treasure and relics: coins, jewels and magical artifacts of great power. They will need crafty tactics to bypass or obliterate the defensive units that stand in their way or else they risk being driven off into the wilderness to lick their wounds. As dragons claim victory they return to their lair triumphant and laden with booty, growing ever richer and stronger. But other players won’t just sit back and watch; they can unleash potent spells from afar in an effort to thwart dragon attacks and aid NPC defenders.

A Thunder of Dragons is a board game I've been designing with Nick Henfrey, co-creator of Conquerors and Spacefarers. (To be honest, all the heavy lifting has been done by Nick while I chip in with suggestions about game balance.) The prototype is a lot of fun to play, and I'm not saying that just because I won our first full game.

You start by shuffling and laying out terrain and settlement cards. This ensures the game board is different every time. Players establish their lairs and can either walk (slow but easy) or fly (fast but uses up power), picking on settlements which they can plunder for treasure, captives (princesses and princes too; no gender bias from us), and spells. You can hold cards to add to your hoard or hand them in to increase power. 

It's really rare for an early prototype of a game to play as smoothly as this. Normally what happens is you start fitting pieces into the rules jigsaw and it's all going well till you hit some part of the design that just refuses to fit with the rest. I've been struggling with something like that in my Jewelspider RPG design (nearly cracked it, though) and I thought Nick and I would have similar problems as we had with finessing the Zombomba boardgame. But no -- we laid out the map tiles and got playing and it all came together like Smaug swallowing a hobbit. One gulp.

My victory in the first game was a bit of a fluke. I began by attacking an abbey. Little did I realize that abbeys are really well-defended and when you're starting out there's a high risk of being driven off and/or being badly injured -- and if you use up all your power in the attack you'll have to try and get back to your lair on foot while pursued by the settlement's defenders and reinforcements. Luckily I survived and carried off a major relic, putting me way out in front. But even that didn't secure a sure victory, because the other players can see who is ahead and will team up to harass them with spells.

As you can see in the picture above, the dragon playing pieces are 3D printed models, making the game as visually appealing and tactile as it is fun to play. But the frustrating thing is we just don't know what to do with it. Patreon and Kickstarter would never raise enough for us to be able to sell physical sets of the game, and nobody is willing to shell out for PDFs of a boardgame. These days, the successful crowdfunded games are all by established games publishers. But if anyone out there can suggest a company we can team up with to turn A Thunder of Dragons from fantasy into reality, please shout it out in the comments.

You can follow A Thunder of Dragons on Facebook and Twitter.

Dalek City

So here's a blast from the distant past, one of the game design concepts I was spitballing with and/or pitching back in the mid-'90s at Eidos Interactive: Dalek City. It would of course have been a spinoff from the BBC's series Doctor Who. That show was pretty much dead and buried back then, so getting the rights wasn't a complete impossibility. And I was pretty sure Terry Nation, who created the Daleks and had joint control of the rights with the BBC, would be keen to hear any ideas for exploiting them.

"It's set on Skaro, the Dalek planet," I began. "The Daleks have been mutated by nuclear war and can only move about in mechanized travel machines. At the start of the game they pick up power by induction, so they can't leave the city."

"Yep, got it." said the Eidos executive I was explaining this to, who may or may not have been a famous gamebook pioneer. "So you have to get upgrades to let them travel outside the city."

"Well, it's certainly possible for the Daleks to get those upgrades. But you don't play the Daleks in this game."

"What, then?" 

"There are all kinds of threats to the Daleks. Various mutants live in the jungle around the city. Natural disasters like meteor showers can occur. There's another race, the Thals, who are their ancient enemies."

"You play the Thals, then?"

"Not really. You don't play anybody. You can step in and help the Thals if you want. Or you could spawn lots of mutant monsters to overrun the Dalek city."

This at least sounded like familiar territory. "That must cost you resources."

"No, your resources are unrestricted, up to whatever the game engine can handle. You could just send in so many monsters, raiders, and natural disasters that the Daleks would be wiped out right at the start. The point is, say you do that a few times. Then you try something different: sending in just one monster to begin with. The Daleks kill it and take it to their labs. They start researching it. Pretty soon they don't have any trouble dealing with that kind of monster—and what they've learned will help them in other ways, too."

"I see; it's one of these Artificial Life things," he said. (It was more of a growl, really.)

"Well... kind of. The Daleks are prime candidates for A-Life because their psychology is so simple. They're paranoid, inquisitive, power-hungry and they hate everything. And their society is like a type of insect hive. The aim of the game, you see, is whatever you want it to be. You can just observe the Daleks going about their duties, like your own little formicarium. Or you can trash their city and watch the little buggers get stomped. Or you can test them with various threats and see how they learn and develop. It's the cruel-to-be-kind method. Eventually you might find you've nurtured them to the point where they can take anything you throw at them. Played that way, the ultimate aim of the game is to make the Daleks into an opponent you can't beat."

For a long while he said nothing. I almost thought he might be considering it. Then he shook his head. "Players don't like games without a clearly defined objective."

New kinds of interactivity promise a world of possibilities that we have hardly begun to explore. To fully realize those possibilities, though, we have to be prepared to let go some of the control that we have come to expect, both as designers and as players. A quarter of a century on from that meeting, I'd like to think game publishers would be more willing to entertain left-field ideas, though I'm not sure how often that's true!

How to create a new world

You’re creating a virtual world. What’s your job, and what isn’t?

You’re god here. You’re laying down the rules of the universe. And you need to be a benevolent god because malevolent gods don’t have many customers. So you won’t populate the world with lots of dangerous critters – or anyhow you’ll provide safe havens from such critters. You’ll start people off with the rudiments of an economy. Maybe some tools to get them started.

But only to get them started. It isn’t god’s job to design buildings and cities. You might want to make a city – gods, like everybody else, gotta follow their bliss. But the laws of the world and the basic tools will mean that pioneers will take one look at your city and go, “Nah, I’m going over here to build a place of my own.”

That’s a good thing. Gods have to let go. For starters, if the people in your world are designing stuff then that’s less work for you. Also it gives them ownership. It means they get the virtual world to be what they want it to be.

Think of it like making the Garden of Eden. You come back a few days later to find Adam and Eve growing trees and herding sheep. A bit later they’re tapping for rubber, carving wood, using sheep intestines to make string. Next time you see them they’ve invented tennis. And all you had to do was the physics and the biology.

So virtual worlds can and should grow to reflect the wishes of the users. Except…

Seen The Deuce? You probably wouldn’t want to live near Times Square in the 1970s. What if your early adopters turn the world into Damnation Alley but you’d really like to attract the vuppies (sic) who’ll bring in the money and numbers? They get turned off by all that dark sleaze and might not come back for a second look. If your world is big enough there’s room for both groups of players, but that’s kind of a cop-out. It’s really two different worlds.

Around the time Second Life was making a big noise, a finance guy asked me to design a virtual world. Actually, he didn’t so much want a designer as somebody to take notes, as he already had a lot of ideas about what the world should be like.

“It’ll look like a modern Western city,” he said. “People will use money and they’ll be able to have sex.”

“So there’ll be prostitution.”

“No, I don’t want prostitution. Also, you can mug other players’ avatars and steal from them but you can’t kill them.”

“So there’ll be rape.”

“No! No rape! What are you saying?”

“What’s to stop somebody from stealing another user’s money and then only giving it back if they agree to sex? Or threatening them for sex? What’s to stop somebody demanding money for sex? If you put those elements in, prostitution and rape are emergent.”

“I don’t want those things in my game.”

He couldn’t see how his design decisions (mugging, theft, sex) had consequences. Tutoring a money guy on how game design works isn’t on my list of enticing jobs. I turned him down. But it did get me thinking about how you police the behaviour of people in a virtual world.

The ideal answer is that god doesn’t need to be a lawyer or a policeman any more than he or she needs to be an architect or a city planner. If people want a civilized society, they can organize it for themselves. Your job is to give them the tools to do that. You might, for example, place a constitution monolith in the centre of the world. Users can vote to add, remove or modify rules. Those could become laws of nature. Swear words, say, might be auto-censored. Or they could be behavioural guidelines that users are expected to abide by, and policing that social contract becomes a community matter.

There’s no hard and fast rule, but my working principle would be to give the users the tools and let them decide how to use them, whether it’s to make towns or to make laws. If there’s a demand for something – cars or buildings or whatever – then somebody will supply it. There will be businesses springing up with car designers and architects. The world developers are free to focus on the important stuff, like gravity and sparrows.

I’m talking about the big virtual worlds here. The ones with their own economies, where famous rock stars come to give concerts and studios hold movie premieres. But there are also what we might call boutique worlds, effectively cosplay theme parks. They will appeal to smaller groups of players with more specialized preferences, who come to play and typically won’t fuel much of an internal economy. The difference is that they will pay you more than the casual everyday user of a big world, and in return they expect you to do the heavy lifting. You’ll have to be an immanent, hands-on deity.

A fictional example is Westworld. Everybody who comes there wants to cosplay in a wild west setting of the late 19th century. Think of the real world (our world, that is; the one the guests come from) as a typical big virtual world – it has an economy, assets are user-created and traded, laws are user-designed and agreed on. If you want to play a game in such a world, you design it. You might make or buy a ball or a set of pieces, explain the rules to others, and you can play whatever you want. It’s real life. Westworld is a boutique world. In effect it’s a game, a traditional one, whereas the larger virtual worlds are exactly that: worlds.

People who come to Westworld aren’t there to redesign the town, they’re there to play the game. (Admittedly, like roleplaying, it’s actually a set of overlapping games with personally defined goals.) There won’t be any Travis Scott concerts, although some boutique worlds might be designed specifically as music venues. With a boutique world, you have to dress the world as well as create the fundamentals. Why would you do that, given that you’ll get fewer users? Because most of the people who frequent the big, tabula rasa worlds are not going to be giving you a lot of money. Most of them, the ones not directly driving the internal economy, are casual visitors – tourists in your world, really – who can and will migrate to other virtual worlds whenever they like. But the aficionados of a themed world will pay a lot, and they’re likely to stay.

I haven't even touched on the ethics of enslaving intelligent AIs to play the NPCs in a theme park (real or virtual), a topic examined in Westworld and discussed here by Dr Richard Bartle.

Between the themed and the tabula rasa world there’s another option: the world specifically designed for tourists. Literally tourists, in fact, as these would be recreations of real-world locations that either no longer exist or that some pesky pandemic or travel restriction or lack of time stops people from visiting. Imagine you could walk through the ruins of Pompeii – and at any point you could dial back to 79 AD and see what it looked like before the lava rolled in. Or London before the Great Fire. Or Tenochtitlan before the Conquistadores razed it. Or Angkor Wat without the blooming jungle. You could climb Everest. You could explore the Mariana Trench. You could stroll around the Forbidden City or Jurassic Park.

Who pays for that? Many developers already have models of such locations. Museums too. You might do a deal with the Great Courses to feature some of their content in the audioguide. Hilton hotels might provide visitors with their only 50th-floor view of Lake Texcoco. Tickets for real-world tour trips might be hidden, Willy Wonka style, for visitors to find. Reuse the content as AR so that tourists in real Pompeii can still look into your reconstructed version through their phones.

Virtual worlds are going to be sprouting up alongside the real one in ever-greater numbers. People will spend more time in them. They will have real economies. Events will take place in them: concerts, political rallies, parades, riots. Some people will make their living there. What happens there is going to matter. The gods of those places had better be ready.

Analyze this

Maybe you’re still in the halcyon days of roleplaying. I mean the Goldilocks time when you’re safely past the juggernaut of finals and have yet to be distracted by career or kids. The days of dossing around, some might call it. You get to see your friends all the time and together you can slip into the parallel life of your roleplaying world whenever it suits you. Nothing beats it for immersion. Arguably it’s the only true way to roleplay.

For me it was back in the ‘80s. We’d have at least one evening’s gaming a week with the whole group, and usually two or three side sessions featuring one or two players who could then flesh out their characters’ extracurricular activities. And after the big Thursday night game, often we’d sit into the small hours (or even dawn) talking about the world of Tekumel, or chatting in or out of character. It was at one such post-mortem gathering, after our characters had disrupted the summoning of the tempest demon Kirikyagga, that the wind started to pick up and somebody mentioned that Kirikyagga was annoyed. That was October 1987.

But I digress. The reason I mention all this is that there was an interesting difference of approach in those post-game chats. Some players (eg Jamie Thomson and Mark Smith) liked to talk about their characters’ goals and personality and what we’d nowadays call story arc. “As Jadhak I used to be very cruel,” Jamie might say. “But since that Llyani curse caused me to lose my sense of fear I'm much mellower. I was cruel out of fear, you see, as a defence mechanism, and now my need for that has gone.”

This was a foreign language to players like me and Paul Mason. We threw ourselves into the role while playing, but I didn’t ever think about my character in an authorial way. On reflection, that might just be because I don’t think about myself in an authorial way. I would never map out how my character was going to develop, or even have any interest in analysing his behaviour. “You must have planned it that Drichansa is always kind to children,” Jamie might protest.

Drichansa was my character. All I'd started with was a mannerism: tugging at my earlobe when really trying to get to grips with a problem. Everything else about Drichansa I discovered as I played him. Jamie found the kindness to children surprising because Drichansa was otherwise notably lacking in tenderness.

“Kind to kids? I suppose I am,” I said. “I never thought about it.”

“You told Jadhak you were adopted. Could that be why?”

“Maybe. Want another whisky?”

You might think it’s odd that an author wouldn’t go in for that kind of character analysis, but I don’t tend to do it with the characters I write about either. Sitting at a keyboard making stuff up can get boring if the characters don’t surprise you from time to time.

This could explain why I’m not much interested in narrative mechanics for roleplaying. I don’t want to control my character like an author; I want to be them. I recently saw the latter method derided as the Actor Approach, and the person went on to say, “That’s not even how real actors do it.” Quite right. An actor has a script (most of the time) and even if they’re in a Mike Leigh or Christopher Guest movie they’ll have sat through extensive character workshops and discussions of the storyline first. But the attraction of roleplaying for me is to be neither author nor actor. It’s more like life: fielding stuff as it comes at you, and finding the story (or rather, stories) that emerge from all that noise only when you look back at it – and even then only if looking for story patterns is your thing.

But that style of playing is not so easy once you’re out of the sweet spot between college and adult life. We get fewer opportunities for gaming (my sessions are down to once a fortnight) and less time (no more playing till after midnight). No wonder that today’s games look for ways to jump-start inter-PC relationships and squeeze your fantasy life into the familiar shapes and tropes that stories take in creative writing courses.

And it occurs to me that’s what dungeons were, back in the dim mists of roleplaying history: a story shape, admittedly crude and built out of rooms and ten-foot corridors, that led you to a Big Bad at the end and allowed for campfire moments back at the town in between expeditions. A three-act structure in architectural form.

Modern games do a lot better – although arguably a physical environment is just as effective a way to shape a story as using plot points and scene breaks. Still, I gave up dungeons pretty early in my roleplaying career and I enjoy the emergent unpredictability of just-dive-in roleplaying stories too much to want to wrangle them with plot paradigms. Also, one of my day jobs is sitting with other writers planning characters’ story arcs. I enjoy that exercise of craft very much, the problem-solving and the personality construction, but I can’t see an evening spent doing pretty much the same thing as relaxation.

An example: not long ago I came very close to running a campaign from a published book complete with pre-planned adventure. The book begins by saying that each player should pick one of the other PCs as their closest friend, and another who they most trust, and so on. For me that should all happen in-game. I don’t want written backstories, I want players to forge those relationships out of their experiences as they play. Then they’ll really feel it. If somebody at my table says, “Out of game for a moment, I think my character would…” then I feel like I’ve failed. They should be leaving their everyday life behind. If they’re stopping to view the characters from outside then they’re distanced from the fantasy, and that means the game isn’t working.

By the way, this applies to writing too. If you start a novel or script with two characters already in love, that won’t have anything like the impact of having them fall in love in the course of the story. Games likewise. A year or two back I consulted on the design of a computer game that began with a long cutscene explaining how the player had a pet dog called Jack who was your best pal, and together you got stranded on a desert island. I threw out the cutscene. “Have the player get shipwrecked and then find Jack trapped under an overturned lifeboat," was my advice. "You get to free him -- that's the first time you've met him, and so the bonding between you happens in-game rather than before the game starts." That way the player will actually care, because they experienced it rather than just being told about it. (Game storytelling 101, that, but you'd be surprised how many developers don't know it.)

Some people enjoy being the author of their character’s life, and/or bringing a five-page backstory to the first session, or calling time out to explain (often in third person) how their character arc dovetails with something that's happening in the game. They are more comfortable with the distance that brings. Well, fine -- you should play the game whatever way lets you get most out of it. But given all the RPGs these days that are designed to conform characters to types and tailor events to an archetypal narrative, maybe you should try it at least one time without preconception, script, or safety net. Just put on the persona and be that character. The worst that can happen is you'll lose yourself in the game.

It's in the trees

I've shown you this before, sort of. While working on a book I like to print up prototype versions rather than read the text on-screen. The upside is it provides a different perspective. The downside is that by the time the print company gets the book to me, often I've changed most of it.

I prepared these two copies of the Jewelspider RPG (2nd edition Dragon Warriors, if you prefer) so that my group could start playtesting the rules. I'm sorry to say the finished book probably won't have Jon Hodgson art -- I don't have the money to pay him, and if I did I'd spend it on Mirabilis -- but for private use around the gaming table I can indulge my wildest dreams. And I really wanted to have a proper look at that gorgeous Players Guide artwork without the book title inexplicably covering up half the image.

Some people have asked about the new rules. Details are still changing week by week, but the core of the system seems pretty solid now. There are eight abilities, ranging from 2 to 18, which determine your chance of succeeding in any action. There are also four qualities, ranging from -3 to +3,which don't affect your chance of success but rather your degree of success. So if you attempt an action using Agility (ride, dodge, climb, etc) or Dexterity (shoot, cut a purse, pick a lock, etc) then having a positive score in the Graceful quality would make any successful roll more effective.

There are also masteries, ranging from 0 to 6, which give the character more control over how they use their abilities for actions relevant to that mastery. Mastery in swordplay, for example, lets you finesse your Dexterity rolls when attacking or parrying with a sword. The way a mastery works is that you can trade off chance of success against degree of success, up to your level in that mastery.

The system is designed for ad hoc play. Any action you want to attempt will be governed by one of the eight abilities, and masteries can be extemporized too.

That's not quite all. There are two very rare qualities, Holy and Fey, that can be unlocked and give access to actions that ordinary people can't attempt. You can't have both at once, of course, and Fey doesn't necessarily indicate faerie blood, it's just the Jewelspider equivalent of DW's Psychic Talent.

When will all this be available to the public? I'm currently running a short campaign with junkable characters. Then Oliver Johnson is planning to run a Jewelspider campaign through through the autumn, and Tim Harford will hopefully give the rules a spin in one of his eagerly-awaited Christmas specials, and then I'll go back and revise the whole caboodle in light of my players' comments. So not till next spring, at the earliest. But, as you know, nothing's forgotten and it's coming.

Get real

The realism versus playability debate has been going on for decades. Which is odd, because there aren’t really any games that err by being too playable. I can cite lots that are too realistic, though. There was the CRPG where you had to remember to restock on shoe leather. If you didn’t, your character would go “Ow!” every so often and lose a hit point. A long journey could kill you before you even got to the dungeon.

The same debate occurred long ago in movies and TV – although there it was “realism vs enjoyment”. Thankfully, the realists were beaten back into a tiny corner. Other than 12-hour Andy Warhol epics watching a flag flap on the side of the Empire State Building, visual narrative is free of realism. Arnie says, “Let’s go to Cairo,” and – alakazam! – there he is.

The guys at Pyro got it right when they talked about narrative games (like CRPGs) involving a contract with the player. It’s what happens all the time in movies when there’s a flashback. Sixty years ago, audiences needed a wash dissolve to believe it. Now you can play around with time using just an ordinary cut.

Why have realism at all? Well, take an example I used when designing my RTS Warrior Kings. Without any rules for supply in such a game, conquest works like infection. You can take a single worker behind enemy lines and build a massive base to attack from. That will lead to some pretty odd strategies if the game is set in the Trojan Wars.

But you don’t want real realism. Full-on true-to-life supply line rules can so easily lead to a player struggling against the game rather than against the other players. So you need to find a way that rewards the player if he does it right, but still allows him to ignore supply lines if he wants. One way to do that is to have injured characters automatically recover hit points if they’re in supply, for example, which is how I had it work in Warrior Kings. The player doesn't have to micromanage supplies, but they do get a bonus for not letting a force get cut off behind enemy lines.

Still, games aren’t movies. The whole point of a game is to give the player a hands-on experience. And sometimes that experience might be of inevitability. I played a wargame of the Cuban revolution. The government player couldn’t possibly win (Michael Corleone was right) but it was fun to see why they couldn't. Only games can do this. Which is why the debate will rage on. And there will always be a case to be made – even for shoe leather.

How to roleplay

Paul Mason is famous in roleplaying circles as one of the uber-fans involved with Dragonlords and as the editor of the superb if infrequent imazine, in which he treated us to a stellar series of articles and reviews in his inimitably trenchant and thought-provoking style. He was also for many years one of my Tekumel players and has written Outlaws, a great but so far unpublished RPG of the heroes of Liangshan Po, which I used as the basis of my (also unpublished) Heian Japan roleplaying game, Kwaidan.

These days Paul is too busy with his academic career in Japan to do much roleplaying, but the last time he was over in Britain I asked if he wouldn’t mind me running some of his articles as guest posts, and he gave a kind of oblique permission. That is, he looked at me with an expression that was more 'are you serious?' than 'don’t you dare'.

This piece might strike you as very basic stuff if you’re a roleplayer – but hey, I’ve been roleplaying since the mid-70s and I found it useful. Remember that once you reach 10th Dan you go back to wearing a white belt. Nobody should ever think they’re anything but a novice. Take it away, Paul...


In a role-playing game the rules are details: they are the trees from which part of the wood is composed. So let’s consider a different approach to writing rules for role-playing games. Let’s try to look at the wood.

goal
The purpose of this game is to take part in a story. The story isn’t told by anyone, but is built up from the improvised contributions of all the participants. See the sample for an idea of how this works.

how to play
The game creates a story. Participants in the game all play a part in creating the story, by making contributions. The goal of the game is to make it as easy as possible for participants to act or describe their improvised contributions to the game without spoiling the sense of immersion.

participants
There are two basic types of participants in the game. Players are a little like actors. They will usually act the life of a single person: their character. The referee is more like a director. The referee describes sensory information in the story, and may occasionally act other characters in the story, as needed.

action
A participant who contributes to the game by acting does so by saying what their character is trying to do. So in the sample, Fred says: ‘I climb up the gantry to the deck above.’ If you like, when this action is speech, the participant can act the speech by actually speaking as the character. So later in the sample, when Fred says ‘Set it to stun!’ he’s actually saying what his character is saying. In some cases you might need to check which it is, but usually it will be obvious. Two or more participants can thus act the roles of their characters, conducting a conversation which forms part of the story.

Anything which is acted by a participant takes place as described, unless it is challenged by another participant (usually this is the job of the referee, but other players may also challenge if they like). A participant whose action has been challenged must prove that the character could succeed. To do this, they need to use an agreed game mechanic (such as Outlaws Light, presented in imazine #33). An example of a game mechanic is that you must roll 9 or less on two dice to hit with your phaser. Really skilled characters like Worf need an 11 or less. Other Klingons need 7 or less.

Some complex interactions, such as fights, often involve continual implied challenges, and therefore may require a lot of use of mechanics. Other actions, if they seem reasonable given the character and the story, can pass unchallenged.

description
A participant who contributes to the game by describing does so by talking about something accessible to the senses of characters in the game. This is usually the job of the referee, but players may also occasionally describe things connected with their characters. So in the sample, Sam describes what the players can see once they have climbed the gantry, and what they can feel.

Descriptions, like actions, can be challenged. They shouldn’t be contradicted outright, but senses can be mistaken! A player who describes a scene is speaking only for their character, and other players, or the referee, may perceive things differently. Note that the referee is privileged in description: because they speak for ‘everybody’ a player who challenges a referee’s description is simply describing what their own character perceives, and not what anyone else does.

Obviously, not everything needs to be described, and referees should beware of trying to act events in the story in the guise of description! For example, if Sam in the sample goes on to say ‘When you walk on to the transporter pad, there is an explosion’ this is wrong, because the players haven’t yet said that they are acting by walking on to the transporter pad. Remember, you’re not telling a story by crafting it authorially, you’re creating one by inhabiting it.

contributing
There are no fixed rules governing how and when you can contribute to a story, but there are some obvious guidelines that should be followed. The most important is: take your cues from the story. If you act something your character is doing tomorrow, then everyone else’s actions today will have to be done in flashbacks. This will be difficult, and may even cause a contradiction with what you acted about tomorrow. Challenging other player characters, or getting into conflicts with them, is fine, but blocking the story itself is generally bad form.

A typical sequence of contributions will be:

• Referee describes the situation facing the player characters, and/or uses a character to act a stimulus.
• Players respond by acting their character’s reaction. There’s no fixed order to this, but if a player feels that their character should be able to act first, they always have recourse to a challenge.
• Participants respond to the actions. This may lead to further description—the referee, or a player, may describe the result of actions.
• Out of all these contributions, a sequence of events will soon be evident. This is the story.

Even in your own mind, separate Action from Description. At first it’s tempting to think that your character could do absolutely anything, but soon you find that the limitations are what create drama. Maybe you can’t leap that chasm, maybe you’re not fast enough to outrun the fireball. Maybe the Ferengi saw you pick his pocket. Sometimes you should challenge yourself, not wait for other players to do it.

timing
Time for the characters in the story does not pass at the same rate as it does for the players. At times, it will pass very slowly, if you’re working out something that doesn’t take long, but needs to be explained in detail. At other times, it will pass very quickly, as with a long journey in which nothing much happens. As with most things in the game, time can be skipped over, subject to challenge by any of the other participants.

winning
There are no rules to cover winning. Players can decide on their own ideas of what constitutes winning. However, they may find that other players don’t agree with them. So how do you win? Well, how does a character win in a story?

ending
The game takes place in game sessions. A game session is when the participants get together to play the game. It can end at any time that is convenient for the participants. The end of a game session doesn’t mean the end of a story. The story can continue in the next session. A story only ends when everyone agrees that it’s finished, and you start a new one, or when you stop playing the game entirely!

postscript
Thanks to Dave Morris for providing comments and useful examples based on Star Trek. In writing this, I’ve been particularly inspired by all those games which have started with some vague waffle about how role-playing is like improvisational radio theatre, have followed it with a sample dialogue, without any explanation as to how and why people said what they did, and then plunged straight into tables of character generation. I’m also indebted to my own players, half of whom were complete beginners.

- Paul Mason

Like an egg stain on your chin

Recently I’ve been reminiscing about our roleplaying days of yore. Not in order to wallow in nostalgia, but for the sake of some interviews and podcasts I was doing. I talked about the saturnine loner who achieved enlightenment and saved the people he realized were not lackeys but friends. The civil war that split our party when each player-character came to different conclusions about the right and honorable course. The subtle ways that characters within a legion, even at different ranks, could push their disagreements as far as military rules allowed.

I’m forced to the conclusion that the roleplaying was better back then – more immersive, more nuanced, more surprising – when we just took a Tirikelu character and developed them by playing. Now we mostly use GURPS, which encourages you to plot out every preposterous detail of the character before you start playing. It’s not a springboard for the imagination. More often it’s just a straitjacket.

And by the way, I'm just singling out GURPS because it's the game I've played most in the last ten years. Plenty of so-called narrative systems are just as bad, with their nannying insistence on each player writing down which other character they like, which one they have a grudge against, and so on. It's like being at infant school and being made to write about your weekend. The point of playing is to discover these things, not scribble the backstory to a bad novel.

I much prefer the approach taken by Stephen Dove for his Jewelspider Chronicles campaign. There you begin with a short "mission statement" for the character, clarifying some background details but leaving plenty of room for future development. As an example, here's my initial description of the character I played in Tim Harford's Company of Bronze campaign.

I already talked about why GURPS’s mental disadvantages don’t work but there’s a problem with character disadvantages in general. Say you cap disadvantages at -20 points. All the players will immediately take the maximum allowed. What's wrong with this picture? Simply that if the disadvantages were properly priced, you'd expect to see some players not bother with them at all.

“Ah, but character diamonds.” No, giving extra points for disadvantages is the junk food version of interesting characterisation. A lame epileptic drug-addicted albino with the regulation five quirks is not the slightest bit interesting. What makes a character compelling is in the gap between desire and duty, wants and needs, feelings and experience. And better by far if those internal conflicts are drawn with a subtle brush, not the cartoonish personality traits offered by the GURPS rules.

So I'd allow players one disadvantage. Just one. That's it. Not a mental one, either, because they're all anathema to good roleplaying. If you take the disadvantage, you can spend the points on an advantage. Again, just the one.

How are you going to get that interesting characterization? Do what good roleplayers manage without any of the personality-by-numbers stuff. As Laurence Olivier said: dear boy, just try acting.

A rules rutter

What are we looking at here? A good question. You know I was talking about having another crack at revising the Dragon Warriors system? More like completely rewriting it, in fact. It's a project I've returned to many times over the years, usually abandoned in short order as the need to actually crack on and run a fortnightly campaign gets in the way.

This time of going back to the well, I have a rules mechanic that I'm finding pretty neat. Those could be famous last words. I did remark to one of my gaming group that "designing a new set of rules is like doing a jigsaw. After early frustrating dead ends, everything seems to come together, gathers momentum, gets exciting – and then you see the gaps that the remaining pieces just won’t fit into. Rinse and repeat."

But I think I can punch through the doldrums of design to arrive at a workable set of streamlined rules that will fit any contingency. God knows we need it. The obscure rules lurking in the thousands of pages of GURPS books is starting to try the patience of most of my players. We only get a few hours' gaming every couple of weeks. We need something simpler.

So, that book in the picture. In order not to repeat the false starts I've made in the past, I took all the notes I've made on different versions of DW2 rules and collected them into one volume, which I then printed up on Lulu. I find having a physical book like that is easier than wading through multiple files on the computer. Just behind the rules book there you can see the homemade booklet I used to prep for writing a chapter in The Design Mechanism's upcoming Lyonesse RPG.

Just to give you a taste of all these notes, one of the briefest sections in the booklet is this overview I sent to Grenadier Models UK when we got to talking about collaborating on a new roleplaying game in the early '90s.

Everything is based on a skill system, so a character might be a Rank 3 Wizard and a Rank 8 Fighter, or whatever. Ranks are purchased with Improvement Points, which are acquired by training or experience. There are no "character classes". The cost to acquire ranks of different skills depends on the character's culture. So elves need fewer IPs to advance a rank of Wizardry, more to advance as Fighters. 

Combat is handled by comparing Attack and Defence values. In some ways it is similar to the Dragon Warriors system, but characters can exercise a degree of choice in how much they concentrate on attacking as opposed to defending. The range of choice reflects different styles of combat. When a hit is scored, damage is determined by a single dice roll which is modified by the weapon used and the attacker's rank as a Fighter. Armour works by absorbing some of the damage.

I am in two minds about whether to include hit location or not. It adds a certain colour to any combat system, but it does tend to slow things up - and you get into problems where non-humanoid creatures are involved. The alternative system uses "wound values" - any wound causes Attack and Defence penalties, depending on how much damage is inflicted in a single blow. Characters are more likely to pass out from cumulative wounds than to fight on until cut to ribbons. This means that combat is fairly ferocious and damaging, but as long as the players' side wins in the end they will generally be able to heal up their fallen companions.

Magic is divided into three types. The first is Wizardry. This uses up no spellpoints, but requires a skill roll to work properly. It is also quite difficult to learn. It is the way a magic-user would contrive most of his "special effects" - weird events that are not directly related to combat. About a hundred Wizardry cantrips allow the magic-user to pass through locked doors, go unnoticed, conceal a trail through woods, and so on. I dislike the idea that wizards in many systems have to use up their spellpoints for quite minor effects. I cannot imagine Merlin or Gandalf crossing off a couple of spellpoints for an illumination spell, for instance. The Wizardry rules are intended to represent the popular fictional concept of the magic user more accurately.

The second branch of magic is Thaumaturgy. This is combat related magic. Wizardry illusions do not do real damage, for instance, but Thaumaturgy illusions can. Thaumaturges expend psychic points to cast their spells. The number of points available increases only slightly with rank, but what does increase significantly is the number of spell-matrices the Thaumaturge can hold in his mind. When a spell is cast, the mental matrix for that spell "fatigues". It will defatigue with sleep, but a further casting of the spell when the matrix is still fatigued will cost double points. Higher ranking Thaumaturges therefore never get to the kind of artillery-level capability of a D&D magic-user, as their power really lies in the greater versatility they get from having more spell-matrices available.

The last magical skill is Theurgy. This involves the manipulation of campaign magic. Such things might include gathering information about a foe's army or creating an enchanted artifact. Theurgy is often done in conjunction with other magic-users, as it involves a permanent loss of psychic strength and it is better if this loss can be shared between several characters. It takes long periods of time to work (and must often be performed on specific astrologically-favourable days) so it is useless within the limited time-frame of one adventure.

The idea is to capture all the rules notes from over the years so I can sort the wheat from the chaff. So I'm not sure which of the ideas here (if any) will make it into Dragon Warriors 2 but we'll see. I certainly want magic to be more mysterious, less "artillery".

Oh, and while you're here -- did I mention my Kickstarter for the final Blood Sword book? It's going strong and there's still one day left to jump aboard.

How a surfeit of skills in an RPG stifles interesting stories

This is going to look like another Gripe About GURPS, but the fact is that I’ve been thinking (for the millionth time) of writing a new edition of Dragon Warriors, and so I’ve been taking a look at what I’ve liked and disliked about various roleplaying systems over the last forty years.

Early on we hardly had skills. Lots of people started out dungeon bashing, a form of tabletop skirmish wargame on rails. So, apart from hiding in shadows, opening chests and hitting things, they didn’t think about skills much. The more roleplaying broke out of the dungeon and became about the whole scope of a fantasy life, the greater the demand for rules that covered all the things a character might do.

My 1979 edition of Runequest lists about twenty skills. That felt like a great liberating leap forward. My 2010 edition of GURPS has more than twenty skills that begin with the letter A alone. There's maybe two hundred and fifty skills in all. And that doesn’t feel liberating, it feels like being tied in knots.

Having too many skills limits the narratives that will emerge, because not having a specific skill tends to block potentially interesting developments.

‘In the back room there’s a guy who’s tied up. “Thank God you came! They kidnapped me.”’
‘I free him. But as I do I’m taking a look at those ropes. Is it possible he could have tied himself up?’
‘Suspicious, huh? Have you got Knots skill?’
‘Er… no.’
‘Too bad. Nice idea, though.’

OK, Knot Tying defaults to Dexterity -4, so the player could still try to make the roll. But in practice DX-4 pretty much scotches it. With DX of 12, that nice idea gets squashed down to a 26% chance.

Obviously a good GM is going to find a workaround, maybe make it an IQ roll with a bonus if the character had had Knot Tying. But now we’re falling back on off-the-cuff rules, often a sign that the system isn’t fit for purpose.

In a much simpler system, with no rules for knots or ropes, you might just ask for an IQ roll. Some early RPGs didn’t drill down to the level of skills, but they did allow for character background, so you’d often hear an exchange like this:

‘…I free him. As I do I’m taking a look at those ropes. Is it possible he could have tied himself up?’
‘That’s going to be an IQ roll.’
‘I’m a sailor, too, so I’m familiar with knots.’
‘OK, I’ll give you a +1.’

That’s also making a ruling on the fly, but the difference in the second example is that the rulebook was probably 50 pages rather than 500. In a very granular system like GURPS 4e, skills are differentiated down to the level of Shadowing (following a person in a crowd, which could just as easily have been dealt with using Stealth and Observation rather than inventing a new skill) or Forced Entry (kicking a door in – yep, there really is a skill for that and it has no default).

The problem with all these skills is that PCs are unlikely to have most of them (Knot Tying, for instance), which will often block a course of action that would keep the game moving and be fun. Often they overlap, and inconsistently to boot, so the game degenerates into sophistry as players argue the case for why their obscure skill has a bearing on this situation. And of course, when you do decide to take a character with Forced Entry, the entire world starts looking like it’s made up of doors to kick in. These are all factors that straitjacket the kind of fluid improvisation that powers the best game sessions.

Instead of all that, you could design a system that lets players unpack the level of detail they want. So say I have Melee skill of 10. I can just roll that in any fight, whatever weapon I’m using. But if I prefer there's an option to specialize in one weapon – sabre, say. So now I get a +2 in my specialized weapon but -1 in everything else. Meleeing with a sabre, I now use a skill of 12, but with a club or spear I’m at 9.

And I can unpack further. Specialising in parry, I can now parry with a sabre at 14, but if I attack or dodge I’m at 11 – and at 8 with other weapons.

That’s not necessarily the way I’ll go with Dragon Warriors 2e. I’d like something that moves away from the kind of abstract number-crunching that accompanies character creation in something like GURPS and is instead based around the character’s life up to the time the game starts. Traveller began that trend back in the dawn of roleplaying, I used it in Tirikelu, and it’s in games like Warhammer too. The advantage is that you end up with a character with a history, a context in which his or her skills make sense, rather than just the best numbers you could wrangle using Excel.

The thought that breeds fear

I didn’t get to The Weirdstone of Brisingamen until my thirties, and only then because I started going out with a girl who grew up in Alderley Edge, which is the setting for that and most of Alan Garner’s stories. She had in fact lived in the house he identified as the home of the Morrigan.

“What did you think?”

“Those kids are going to suffer from PTSD. It’s traumatic enough just reading it.”

The trauma wasn't caused by Garner’s prose. That’s beautiful. It was the descriptions of narrowing lightless tunnels and wobbling planks propped across sheer drops; that’s what I thought would scar those characters for life.

Alan Garner wrote a sequel, The Moon of Gomrath, but left it nearly fifty years before completing the trilogy. If only other fantasy authors showed such restraint. What I loved about Boneland was what many fans hated. Garner didn’t give us jolly japes with elves and ginger pop, he returned to the main character to find him broken, his twin sister missing if she ever even existed at all. If you’ve played any Dragon Warriors you’ll understand why that was the sort of conclusion to the story that would really appeal to me.

Other books in a similar vein are:

• The Box of Delights by John Masefield
• The Dark is Rising by Susan Cooper
• Five Children and It by E Nesbit
• The Giant Under The Snow by John Gordon
• Inkheart by Cornelia Funke
• The Lion, The Witch & The Wardrobe by C S Lewis
• Moon Eyes by Josephine Poole
• The Weirdstone of Brisingamen, The Moon of Gomrath and Boneland by Alan Garner
• Wolf Star by Tanith Lee

How about a role-playing game that digs under the surface of a children’s cosy fantasy epic to see what crawls out? It turns out Becky Annison has done exactly that with her game When the Dark is Gone. She discusses it in this episode of Fictoplasm with Ralph Lovegrove and explains the design principles here.

In brief: the players are adult survivors of such an adventure, uncovering their repressed memories with nudging by a therapist character who’s the nearest the game gets to a referee. Minimal set-up, raw character interaction, no dice, emergent stories… What are you waiting for?