Anatomy of a Game Design

Gaming’s Social Contract


Low on health, I crouch silently in Far Cry‘s tropical undergrowth. Pulling out my sniper rifle, I edge forward until I can just make out the guard tower in the distance. Suddenly, I get shot between the eyes by a grunt with a pistol half a mile away. “Stupid cheating piece of crap!” I cry, turning off my 360 in disgust. My problem is not just that I’m a sore loser, it’s that Far Cry‘s creators have (unintentionally or not) broken the social contract between gamer and developer.

A social contract can be seen as a set of unwritten rules which governs our interactions with each other. Always tip your waiter? That’s a social contract. Don’t steal bases in a blowout in baseball? Social contract. Swearing in front of your buddy’s mom? Don’t even think about it.

Social contracts between players are obvious in multiplayer games. Achievement grinding servers for Team Fortress 2 and an agreement amongst players not to use certain tactics, such as grenade tagging in Gears of War, are both social contracts.


In single player games, though, there is a sort of social contract between the game’s designers and the player which defines what we expect from games and what those games expect from us in return. Many expectations are obvious, such as giving the player a fair chance at winning or ensuring that bosses have weak spots, while some are more subtle, such as how we expect our character to grow during the game, but they all influence the design of the games that we play. A well-designed game will allow us to forget that these contracts exist. A poorly designed one will rub them in our faces.

The most notable instance of gaming’s social contract is expectation of “fairness.” As gamers, we both expect and demand a fair chance to win the game. (Note that this is not the same as a game being “easy.” A game can be hard but fair.) We expect that the game will not ask us to do seven impossible things before breakfast, nor will it perform those impossible things against us.

This is, of course, why I was so pissed off while playing Far Cry. There is no explicit reason why the NPC shouldn’t be able to shoot me from half a mile away. A human player isn’t squinting at a monitor, trying to make that shot; the game’s AI knows exactly where I am at all times and could shoot me at will. The only reason that superhuman sentries are poor game design is because such behavior doesn’t feel “fair.” I can’t pull off that shot and so the computer should not be able to either.

There are other situations in which “fairness” constrains design. Consider Half-Life. Did you get eaten by the first barnacle that you encountered? Me neither. That’s because we saw a Barney get eaten by one first. (All guards were known as Barneys in the original game.) The same thing happens with a scientist and the tentacle monster. It occurs, in fact, with any insta-kill situation in any of the Half-Life games. The designers at Valve are scrupulously fair. If there is something new that can kill you in one shot, they will warn you about it first, ensuring that when you do get eaten by a barnacle (and we’ve all tried to climb that “rope” at least once), it’s a fair death.

From a design perspective, showing your hand like that is limiting. Imagine that you’ve devised the best trap ever, full of malicious humor and beauty. No one would ever avoid it the first time around, but the social contract demands fairness. You need to warn the player, while not breaking their immersion. The warning signs in Portal are excellent examples of getting this right. They convey just enough detail, while, at the same time, such warnings add to the game’s environment, not detract from it.

Another point where we see our “fair” chance influencing design is in the construction of boss encounters. Usually, boss behavior is utterly ludicrous. If you had a bunch of attacks, only one of which left you exhausted and your weak spot exposed, would you ever use that one attack? Of course not, but bosses do so all the time. Like an adult playing cards with a five-year-old, bosses intentionally let us win.


Once you realize this phenomenon, experiencing it can be rather jarring. A well-designed boss will provide some reason for their weakness. The more animalistic a boss, the easier its behavior is to explain. Giant animals inevitably want to eat us and thus expose their gullet, for example. The more intelligent a boss appears, however, the harder such behavior is to justify. In the various Zelda games, Gannondorf is forever helping out Link by firing magic tennis balls at him that can be smashed back in his face and I can’t help but wonder why. The moment that occurs is the moment I stop believing that the boss isn’t letting me win. In contrast, Gears of War manages bosses with surprising elegance during both the Berserker and General Raam fights. In each case, the player must expose the bosses’ weak spots through cunning (luring the boss outside) or skill (using explosives to drive off the kryll). This shifts the cause of weak spot exposure from boss to player, reducing loss of immersion while still fulfilling the social contract.

Fairness is not the only aspect of the social contract that game designers must take into account. A game must have room for player skill to make a difference. This might seem astonishingly basic, but it still needs to be considered. A rock, paper, scissors or a Tower of Hanoi mini-game is unsatisfying because our skill is irrelevant. Any battle which relies on the AI’s choice of behavior more than your skill has a similar problem. Dragon Age: Origins would often destroy my party in ten seconds through careful use of stuns then, on the second, third or twentieth time around, neglect to use those stuns. It made a mockery of the social contract that stated that my skill would be the deciding factor. Rather, it was the randomness of the AI’s behavior that determined the outcome. Either the AI should never have used those abilities, or it should have always used them and required me to find some solution. Granting me a win based on randomness is an unsatisfying result. Some variation in NPC behavior is obviously required to avoid staleness, but those behavioral variations should not result in difficulty variations.

Then there are genre-specific social contracts. Players of RPGs expect that their character will be more powerful at the end of the game than at the beginning. At the start of a game, a giant rat might prove to be a challenge. At the end, that rat should be less of a threat than a mosquito is to a mammoth. This is no doubt the reason that many people find the auto-scaling foes of Oblivion so underwhelming. In Oblivion, all enemies level up with you, meaning your relative power has not increased and, in doing so, the social contract of RPGs has been broken.

FPSs have a similar expectation in weapon advancement. At the dawn of the genre, players started with the weakest weapon in the game and went on to collect more powerful guns as the game progressed. Over time, it has become much more common in FPSs to start with an average weapon. This can feel more realistic, but comes at a cost, as looted weapons may turn out to be worse than the player’s current one, breaking the social contract of advancement. Other advancement mechanics must replace it, such as RPG-like elements of character advancement. BioShock‘s plasmids and Sam Fisher’s increasingly sophisticated equipment in the Splinter Cell games are two good examples.


Social contracts in gaming are not a one way agreement from designers to players. We gamers must uphold our end of the bargain as well. Gamers possess a slightly more generous suspension of disbelief than in many other media. We forgive the fact that the boss is wearing “plot armor.” I saw my sniper round plug Zakhaev in the head in Call of Duty 4: Modern Warfare, but the plot called for him to lose an arm, so he did. And as for all those chest high walls that I can’t jump over? Well, I won’t mention them if you don’t mention how many guns I’m carrying. Deal?

For game designers, it’s important to recognize and adhere to the various social contracts in games, but doing so can be limiting. It requires design choices which, at worst, feel ridiculous like Gannondorf, or, at best, go unnoticed by the general populace. On the other hand, the best designs play with what gamers expect from the medium in novel, intelligent ways, such as the insta-deaths in Half-Life. Most consumers of games won’t recognize that a social contract is even in place. They will only notice that something is wrong when their expectations are not met and then lambaste the designers for breaking the social contract.

Just like saying “Shit” in front of Mrs. Swensen garners the response: “Dude, not cool.”

Andrew Bell is a sociologist, gamer and new dad. He never did finish Far Cry.

About the author