I think Canabalt is probably the best example I could give of what a game, at it’s core, really is.

We take a player, give them a quick context to orient themselves as to what we’re wanting them to do. Then we force them forward towards arbitrary barriers that can be overcome with a test of skill. Give the player a long term choice, in this case hit box to slow down or jump over box to keep speeding up. Use only as much simulation as you need and keep a clear and easily understood score count. And one final bit that isn’t required but that I think is something we’ve really lost, make failure not only inevitable, but the only possible outcome.

I noticed in one of the blogs that I followed the other day that someone bandied about a statement that seemed utterly absurd as soon as I saw it. “This is a game, you shouldn’t be forced to do [stuff the main discussion is about].” The problem is, not only should game designers force you to do stuff, it’s literally a core tenant of game design. In fact if I were asked what separates a game from a puzzle it would be that the game is forcing the player to come to a solution, while the puzzle doesn’t especially care if it’s solved or not. Sure we’ve gone to great lengths to find ways to hide, or beautify the fact that we’re forcing the player forward, and we’ve gone from being a lot less push them forward to more pull them forward. That hasn’t invalidated the fact that we’re dragging the player along somehow.

Of course the barriers we’re dragging them towards are basically arbitrary, but that’s not really such a bad situation. There has been a lot of focus recently on making the barriers into an elegant curve, easy to hard, but I don’t especially subscribe to that methodology. We’re not trying to find some optimum method of learning for learning’s sake. Our goal is to deal with the player’s psychology, and how they feel about their own abilities. Sometimes you need to put them in front of a barrier that is much higher than they “should” be able to overcome according to the curve. If they CAN succeed then they know that they’ve really grown and have new limits, if they can’t succeed then they actually get some useful information on what their limits really are. Likewise running into something easy, especially if it’s something they struggled against before, can help them see just how far they’ve grown and gain some appreciation for the obstacles they’ve faced. Just like a narrative, we’re not attracted to an abstract act structure, we’re attracted to a character’s story, our very memory works in stories. In order to keep people coming back to your game, the goal is not to make every minute fun, it’s to make the story they remember one of personal growth and enjoyment. Also a good thing Canabalt specifically does well is to make a situation where success is a combination of skill and luck. This means that after any given run, you can reason that it may be possible to go even further if you just got lucky, however it also allows how far you have gotten to matter since it really did take skill to get out that far at all.

It’s important though that these arbitrary barriers are testing a skill. It has to be something we can get better at, and something with a clearly defined success and failure case. Jumping from rooftop to rooftop mostly requires good timing, and in the case of Canabalt they specifically removed other concerns from the table. You cannot use buttons to change the vector of the jump, or power up for super long jumps, or jump off the screen, any of that stuff. Timing is the only bit they allow you to control and the only bit they test for. On top of this they add the sole bit of long term planning, speed, so that it takes much longer to feel bored or to train yourself to the point where you are reflexively capable of overcoming any challenge.

I could keep talking for a while about simulation, but really the part that I most want to address is that last bit, the untenable scenario. We see plenty of survival modes, but personally I’ve always thought that simply measuring time the player stayed alive was somewhat redundant. The player knows how long they lasted, not in minutes and seconds, but comparatively, next to previous lives. Also a simple time mechanic means that the player is always facing the same situation and being scored the same whether they are always pushing as hard as they can or simply drawing everything out. A scored survival mode, where time is deemphasized and score is counted is much more satisfying as it’s an ever growing challenge, but worth getting better at every stage of. Also, we’ve moved from the nintendo hard paradigm and unfortunately we lost something very valuable with it… that losing is okay. It’s fine to not hit the end and get the super cool ending, just get as far as you can then challenge it again. It’s an ethic of personal growth, rather than the more common practice of waiting for someone to hand you “Achievements”. I’ve never personally liked achievements, it’s just someone else who knows nothing about you giving you a pat on the head for doing something they like. There are things which I find trivially simple which for others around me would be major achievements, and vice versa, how anyone without any personal knowledge of you could have any idea what would be an achievement for you to accomplish simply boggles my mind. But then, there is a reason why I’m indie.