Genre, Metaphor, and why Levelheaded sucks.

A few weeks ago one of my flatmates picked up my iPhone and started playing with it. She flicked through a couple of the games, but was totally stymied by whatever clever touch/accelerometer/virtual gamepad control scheme each used. Saucerlifter? Nope. Levelheaded? Not a chance. There were only two games she was actually able to play: Robot Reaction and Canabalt.

This was insightful for me because, as a developer and as a gamer, I really enjoy exactly the sort of games that she finds frustrating. Over years of play I’ve become (like any gamer) well-versed in the vocabulary of games — the metaphors, control schemes, and whatnot that define how players interact with the game. If I go out and buy Call of Duty N I can be playing (and enjoying) it in a matter of minutes. In terms of a tutorial all I need, as a gamer, is something that explains how the game differs from the Abstract First Person Shooter that I already understand. Gears of War is immediately accessible to me because I’ve played Halo, which was easy to pick up because I’d grown up playing Marathon, and so on. Dwarf Fortress is learnable (accessible might be a bit of a stretch) because I’ve played roguelikes and city-building games — but it’s difficult because the mental diff between these games and DF is enormous.

Non-gamers don’t have these internalized abstract games, though, which makes it incredibly difficult (I think) to design complex games for a traditionally non-gamer audience. For certain games you can get away with it; Smiles works because many of its players are already familiar with the match-three mechanic (thanks, PopCap), and it develops that core idea in interesting ways. If you haven’t played Bejeweled or Tetris Attack, though, it’s qualitatively no easier for a non-gamer to pick up than, say, Starcraft. Quantitatively, of course, it’s way simpler — I think it’s fair to say that match-three is a fundamentally simpler idea than the smorgasbord of metaphors that is real-time strategy. My non-gamer flatmate, for instance, has never played a match-three game and was unwilling to invest the effort to learn the metaphor in order to enjoy Smiles. So what’s different about Robot Reaction and Canabalt?

If you think in terms of background knowledge the answer is really obvious: neither one makes any assumptions about their players’ gaming history. Canabalt may be, in terms of genre, a platformer — but you don’t need to have played Super Mario Bros. in order to get it, and it won’t really help if you have. It takes a single idea (tap to jump) and explores that idea in interesting ways. Similarly with Robot Reaction, whose mechanic is so simple that it’s difficult to assign a genre label to it. Non-gamers can pick up these games and learn the metaphors they need to enjoy them, and they can do it in a single play. This last bit is important.

This is, I realize, was my biggest mistake in Levelheaded. To enjoy the game you need to understand several metaphors (using seed orbs to spread color, how orbs affect the water level, how orbs interact with each other), none of which is easy to learn in a single play. A glance at the leaderboards for the game shows that most people play it exactly once, fail miserably, and then never touch the game again. Looking at the OpenFeint profiles for these players shows that they’re a mix of gamers and non-gamers — why do both groups have the same difficulty understanding the game?

The answer, of course, is that in this case they’re the same group. The metaphors you need for Levelheaded don’t exist in most players gaming vocabulary, so when a new player picks up the game it doesn’t matter if they’ve been playing games since they were seven or if it’s the first game they’ve ever played — everyone’s a non-gamer when the metaphor of the game is novel. In order to get Levelheaded you really needed to have played Jarhead, which gives it a global audience of about twelve people.

The take-away lesson? Keep in mind what metaphors you’re expecting the player to bring to your game.