James Wallis levels with you

Things to do in game design #1: cheat

(This post is a follow-up to “Things not to do in game design #2: cheat“)

Because it’s not about fairness or balance or level playing fields. It’s about the user experience.

Let me tell you about a game you’ve never played. It’s called StarPower.

StarPower was created in 1969 by R. Garry Shirts and published by Simulation Training Systems as an educational game for use in schools, colleges and training situations. It’s a face-to-face turn-based game for 18-35 people that models a simple market-based society. Players draw random counters from a bag and must trade them in order to make combinations that score points. High-scoring players get Square badges, mid-scorers get Circles, low-scorers get Triangles, And from the third turn onwards the Squares can change the rules, altering how the game functions and what players can and can’t do.

What happens, of course, is that the three groups stratify and insularise quickly and the Squares make rules that keep them in the lead and force the other players to continue to lose. The Circles work to consolidate their position in the middle and appease the Squares, and the Triangles begin to feel oppressed, apathetic—or angry.

StarPower is often used to teach historical and sociological principles like class stratification, the abuse of power, and the roots of popular revolutions.

Here’s the big secret-that’s-not-a-secret-at-all: the game is fixed. The draw of the counters is not random. Once the Squares have started to win they will continue to win. Once you’re a Triangle it’s almost impossible to move up a group. StarPower is, by all conventional standards of game design, broken.

In educational-game circles StarPower is a legend.

No matter how liberal the players, the Squares always take advantage of their position, and when the Triangles gets angry, often they get really angry. Though I’ve been unable to confirm it, a senior member of ISAGA once told me about a game of StarPower in an American university where a member of the faculty who’d become a Triangle became so overwrought about the state of play and the abuse of power he was witnessing that he walked up to a member of the Squares and stabbed him.

He didn’t attack the people running the game. He didn’t walk out. He became so involved in the experience that he actually stabbed another player.

How’s that for an emotional response to a game?

I admit that frustration, hatred and homicidal rage are not necessarily the emotions that we as games designers want to create in our players. But still, the designer of StarPower created a game that cheats and managed to focus the players’ resulting feelings not at the game but at the other players. I know I said that two paragraphs up. I’ll keep saying it until the implications sink in1.

Of course games cheat. Computer games have cheated since the original single-player Pong. Of course it’s possible for the computer to play a perfect game of Pong, implacably batting the pill back at you until you falter and miss. But nobody would have played it if it did. So it cheated and missed a few shots on purpose, and voila there’s the games industry.

Because it’s not whether you win or lose, it’s how you experience the game. Good experience means repeated plays, more quarters in the slot, whatever. I mean, if I only bought boardgames that I can win I’d have a much smaller collection than I do.

The most obvious contemporary example of cheating is rubber-banding in racing games: letting the AI opponents exceed their supposed top speed and performance so they can catch up with the player. Because racing on your own around a track isn’t nearly as enjoyable, particularly if your vehicle has forward-mounted weapons. Rubber-banding is occasionally frustrating but crucially the player doesn’t feel that the game is cheating. You may suspect it is, or treat the reappearance of an opponent with momentary disbelief, but did the game cheat or did you over-brake on that last corner?

On Friday I played Journey to the End of Night, an urban chase-game that’s essentially British Bulldogs with checkpoints. Players had to race across London via six checkpoints. If they were caught by a Chaser, they became a Chaser. I was a Chaser. Here’s the thing: we started chasing from the get-go, but we deliberately didn’t catch any players till checkpoint 3. Two reasons: too many chasers early on unbalances the game; and secondly it’s not as fun for the players to be caught. Of course, the players didn’t know they were initially safe: they still ran like rabbits (or in one case froze like rabbits) as we swept down on them, yelling. We were—or rather the game was—cheating, but it was a better game for it.

Back in the day when tabletop RPGs were still fresh one of the first big debates to rage was whether rules should emphasise realism or playability. The argument wasn’t just simplicity vs complexity, though that’s often the way it worked out: the perception was that the more complex the game, the closer to verisimilitude it was. There was even one game, The World of Synnibarr, that was so keen on fairness and transparency it insisted that after a scenario was over the GM should give the players his notes, and if they found any discrepancy between the game as written and the game as played they could claim extra experience points for each deviation.

Twenty years on and the argument’s dead, as is the realistic end of the market and thankfully The World of Synnibarr too. Nobody wants realism or player/GM balance in an RPG any more. They want something that models its genre in an interesting and amusing way, and that’s fun whether to play or to GM.

I’m not going to pretend that the roots of all good game design lie in tabletop RPGs—

Bwah hah hah ha ha ha ha ha ha ha ha ha ha ha

oh, I’m sorry…

hah ha ha ha ha ha


—but every genre of games has lessons to teach and lessons to learn from every other genre. And tabletop RPGs, at least some of them, are very good at using rules to model their genre. So if you’re designing a racing game, what the player wants is racing, and racing needs opponents. If the game has to bend the rules to get them to catch up with you, the good outweighs the bad. Same reason that the player always starts at the back of the grid: it makes for a better game. Not necessarily a more realistic or believable one, but more fun in the long run.

…course, if you’re taking a long run then there’s no need for rubber-banding. Slowcoach.

1 Boardgames, or some of them, draw players in and elicit emotional responses possibly better than any other type of game. Kids kick boards over. I’ve seen a friendship end over a game of Junta. Cat’s parents, diplomats on a foreign posting, once played Diplomacy with the Russian ambassador and his wife: a furious row broke out between the two and they got divorced shortly afterwards. I suspect that the bloom was already off the rose there and the stresses of the game just exacerbated an already damaged relationship, but for the sake of the anecdote I’ll pretend it was down to the ever-tricky problem of controlling the Mediterranean.


Categorised as: game design | street games


  1. Pope Guilty says:

    I have to disagree with you on racing games. Maybe it’s just because I don’t play a lot of them, but it viscerally angers me in Mario Kart 64 or F-Zero when I manage to establish a commanding lead and have it taken away by fiat. It’s utterly maddening to feel like you’re winning only for the game to just say “nope, sorry, fuck you.”

  2. james says:

    I think if you’re playing an arcade racer then you have to accept the rules of the arcade, whether that means nitro, massive drifts, power-ups involving bananas, faking the collision detection to favour you, or rubber-banding. If you want a realistic driving experience there’s always Gotham, Forza, Gran Turismo, F355 Challenge, or actually getting in a car.

    Someone on a mailing list pointed out that rubber-banding is simply dynamic handicapping, it’s just more obvious in a race game than in, say, a golf game.

  3. C says:

    It’s interesting that most of the definitions of “cheating” do not specify whether or not the cheater actually gains. To “cheat” to lose, like in Pong, seems like the wrong word to me, but then I routinely make up my own words. Uncheat. See?

    Diplomacy: I am also a bad board game player when I become bored … one of my friends passed me a note during the game: “TAKE ALASKA” and I waited awhile then passed back my own: “TAKE IT WHERE [RIM SHOT].”

  4. james says:

    Yeah… does cheating have to benefit the cheater? Earlier I was hunting through Shared Fantasy by Gary Allen Fine (University of Chicago Press, 1983) for his piece on dice superstitions and came across this:

    “Lueschen defined cheating in sport as ‘the act through which the manifestly or latently agreed upon conditions for winning such a contest are changed in favor of one side. As a result the principle of equality of chance beyond differences in skill and strategy is violated’ (1976:67)”


    I have no idea who Lueschen is, but on googling I discover his first name is Guenther, and he may or may not be the person, according to Wikipedia, who named his daughter Sonja after Red Sonja.

    The internest. It rocks.

  5. […] rubberbanding is a game designer’s meta cheat to keep the player in the game and the learner hooked to the […]

  6. I note that Guest of Honour at this year’s World Boardgaming Championships was one Dr. Nick Palmer MP (Lab, Broxtowe) – who wrote one of the very few books about board war gaming back in the 1970s and edited Flagship magazine about postal games for many a year – and that he also won the Diplomacy tournament there after having apparently been away from the Dip scene for a decade or two. Surely just the sort of public servant we require?

  7. […] rubberbanding is a game designer’s meta cheat to keep the player in the game and the learner hooked to the […]

Leave a Reply