I once happened upon my brothers attempting to fly an SUV off a cliff. This was years ago, when Grand Theft Auto III was still new, but it was already easy enough to search online for the cheat code to make cars fly. After about an hour of trying to glide across a river and into a football stadium, they finally cleared the edge of the wall, landed the car inside, and broke into proud laughter upon discovering the Easter egg inside: an image of fans spelling out the name of Liberty City's football team: "COCKS".

I often think back on this when I read various theories on why we find games "fun." Some of the most popular theories of engagement come down to offering an optimal level of challenge, establishing a pleasant "flow" state. Surely there was something like that going on here, but there was also so much more, from the thrill of intentionally messing with the laws of physics to the naughty humor in the final payoff.

Theories that account for a range of different types of players, meanwhile, have been useful in considering that games affect us on more levels than simply how challenging they are.

In the process of trying to simplify and codify how people think, however, these theories have trouble accounting for how a game can affect a person in different ways and in different contexts, or how to address appeals of games that don't fit quite as neatly into a carefully-structured model.

In this article, then, I offer five general categories of appeal (hence, "appeals") describing a host of different -- but not necessarily mutually exclusive -- ways that we engage with games.

This framework of appeals has been developed through research conducted between 2008 and 2011, including discourse analysis of online sources (e.g., collecting examples from public forum discussions and blog comments; see "'You are dead. Continue?'") and participant-observation ethnographic research (e.g., playing games with people in arcades; see "Arcadian Rhythms").

The appeals I'll offer aren't necessarily all "good" appeals -- this framework includes ways that games engage players that some designers have criticized as little more than manipulation -- but they may offer some broad ways to describe what makes games tick, and how to blend different kinds of appeals to encourage or even discourage different kinds of engagement.

Types of Players vs. Types of Appeals

I describe this theory quite purposely in terms of the characteristics of games and play instead of the characteristics of players themselves. Models of player personality and demographics are very attractive in their elegant simplicity, whether you're talking about the common-knowledge distinction between "casual" and "hardcore" or more scientific approaches drawing on social psychology. (See Bart Stewart's relatively recent Gamasutra feature for one such robust approach.)

Nevertheless, it may be more productive to describe engagement with games according to a variety of approaches to play itself, for at least three major reasons.

First, theories of "player types" often don't easily match up with empirical and anecdotal evidence of how people actually play games. We can display different "personalities" between different games, or even within a single game that offers a variety of different mechanics.

Take, for instance, the anecdote that began this article, in which my brothers continually flew a car off a cliff. What type of players are my brothers in this example -- say, in terms of Bartle's types?

Are they explorers, fiddling with the game systems and investigating its world? A large part of the reason they were attempting to get into that stadium was indeed that they wanted to know whether the game logic would allow them, and they wanted to discover whatever might be inside.

Are they achievers, looking to beat a rules-based challenge? It was a challenge of their own making, but they still had a distinct end condition, and even a sort of in-game reward in the Easter egg.

Are they socializers, playing side by side, telling a story together through their play? Certainly, playing cooperatively and sharing a laugh had something to do with the appeal.

Are they killers, going out of their way to subvert the rules of the game? They couldn't have played this way at all if it weren't for the fact that they entered a cheat code.

Does it change our answer if we find out that they also played the game separately from one another on other occasions, each following the rules and paying attention to the plot? Or does it change our answer if we find out that they approach other games completely differently -- say, eschewing any "cheating" or exploration in competitive sports games?

To be fair, Bartle originally suggested this typology not to describe all game players, but to describe MUD players. He even makes the point that three kinds of players aren't treating the MUD as a "game" at all, but as "pastime," "sport," and "entertainment," and acknowledges that "most players leaned at least a little to all four [types], but tended to have some particular overall preference."

The fact that game critics and designers have applied this typology more broadly may reflect an admirably progressive willingness to broaden our understanding of what a "game" can be, but it also extends this particular model well beyond the claims of the original 30-person study that brought it about.

My goal with this thought exercise, then, is to illustrate the problem with focusing on a small group of players or a single genre. Players exhibit different preferences and behaviors with different games or in different social contexts, which makes it problematic to claim that anything so fixed as personality or an inherent "type" is at the root of enjoyment. My brothers played the way they did not just because of who they were, but because of the context of the situation: Each was sharing the game with another player he knew very well, and they were playing a game whose design allowed them to play it in multiple ways.

This brings me to the second major issue with describing how we engage with games based on types of players instead of types of behaviors. By suggesting that we design games around categories of players, we run the risk of reifying our own top-down notions of what the player base is like.

This risk could be as innocuous as simply missing out on audiences that we didn't know existed -- a segment of players that requires more nuance to understand than "hardcore" or "casual," perhaps, or that can't be defined as any of killers, achievers, explorers, or socializers. More problematic, however, designing games with player typologies in mind opens the doorway to reinforce stereotypes of which games different people "should" be playing, and which play styles are more valid than others.

In his original article, Bartle didn't have much good to say about the "killers"; they were basically the Slytherin of player types, a category for those who don't play well with others. Bart Stewart's Unified Model goes some way toward legitimizing their activities as a valid play style that most games simply aren't designed to accommodate, but the fact remains that the original typology was constructed in such a way that essentially demonizes a segment of players. In the meantime, releases like Gears of War have demonstrated that there's a market for games that explicitly encourage "killer" play styles, such as by offering brutal and demoralizing ways to dispatch with opponents.

Even more problematic than missing out on audiences, however, is to unintentionally exclude audiences by assuming that certain games are only for certain "types" of players. This risk is probably less in overt categorizations than in implicit or easily inferred connections, like the common assumption that women are more likely to be "casual" gamers, and less interested in games like fast-paced, first person shooters.

To be fair, some studies have indeed observed different preferences between different demographics, and some have even attempted to explain such differences -- say, in terms of innate, cognitive differences between men and women (e.g., see John Sherry's "Flow and Media Enjoyment" [pdf link]). Again, however, it's important to consider the huge role that context can play in terms of what people will feel comfortable engaging in -- or, to put it another way, what they'll even bother to try.

Consider a study by Diane Carr, for instance, which found that when girls were given the chance to regularly play whatever games they wanted in a comfortable, non-judgmental atmosphere, expectations from both stereotypes and other empirical evidence practically disappeared. Yes, men might have a slight neuropsychological edge at navigating a 3D maze in a first person shooter, but that's probably not what's keeping more women from playing. The fact that a game is popularly considered more "meant for" some audiences than others is worth considering as a factor in who chooses to play, rather than who would be capable of enjoying it.

From the standpoint of simply designing more engaging games, however, the greatest reason I see for thinking in terms of "game appeals" is that it's a lot easier to contemplate how to blend different appeals than different personalities. Through design, we can make room for a number of different ways to enjoy a game, or we can purposefully pare down the number of appeals a game offers so that they don't conflict with one another in unintentional ways. Before I offer some more specific examples of how to think about this, however, I'd like to suggest what I see as some useful categories to think about appeals.