Unveiling the Possibility Space

Revealing the scope of a game should be, I think, a seductive act by the game for the player. Not that I know anything about seduction, but I think I've seen it done properly in movies at least, and read a thing or two about it. Seduction means leaving a lot to the imagination while drip-feeding reality in tantalizing doses. There is a sexuality there usually but it doesn't need to be there. A car can be seductive. So can a game. Even an E-rated game, get your mind out of the gutter.

The game I've been working on is now out there, most of the reviews are in, the first-week sales have happened, all that. I can talk about it freely, though if you haven't played it yet, please don't read on unless you're OK with having aspects of it spoiled. Thankfully the game got a generally good response from players, and I'd like to talk here about one of the major reasons why I think it worked for those people. Sure, we worked to craft the discrete elements of the game to a good level of quality, everything from the tuning of the game systems to the art and music and writing and so forth. But I think the essential structure of Bastion is very important to its potential to have an impact on the player.

We said during development that we wanted for the game to present to the player an ever-growing series of gameplay-expanding choices. Part of this involved keeping the player in the dark about the extents of the scope of those choices, culminating in a series of climactic narrative choices designed to feed back on the entirety of the player's experience up to that point. We had to not only continue introducing new elements of play from beginning to end, we also had to do it at slightly irregular intervals such that the structure itself resisted becoming predictable in a negative way. The desired effect is for the player to feel a sense of wonder and intrigue. When you fully understand something, you cannot wonder about it or feel intrigued by it any longer.

Concluding the game with a pair of purely expressive choices, whose gameplay impact was implicit but not overt, was to me the ultimate way of subverting – in a hopefully interesting way – all the gameplay choices that had come before. Up until that point, you'd been deciding what to build, which Spirits to drink, which weapons to use, which upgrades to buy, whether to invoke any of the Shrine idols, and so on. Hopefully, then, the last thing you'd come to expect at the end of the game – especially in a game that appeared to be driving toward one specific outcome – is to have to make an expressive choice about what to do with the world you've been playing in. It's a world you've either grown attached to or haven't grown attached to, and the choices at stake are meant to encompass that entire range of experience.

Bastion is built on this idea of the gameworld slowly unraveling. Every aspect of it. The world unravels almost literally. The story unravels. The game systems come online one by one. There's no telling how many game systems there are in total when you begin play, and in fact, we deliberately mislead you several times about the extent of the game's scope. We make you think you're almost done with the story after several hours, then we introduce another system around upgrading. And when you're almost done with that, we introduce the endgame act. Only when you're about to reach the endgame do we explicitly tell you that, yes, you're about to go into the final area. But even there, the final area is substantially larger than previous areas and has several new kinds of gameplay beats in store. To top it off, once you've finished the game, then we unlock a whole second play-through with more new content. We structured the game this way to keep the experience feeling fresh within the constraints of our scope.

While I think this type of slow-and-steady-burn worked for us, I'm not about to suggest it's beyond reproach. A game needs to prove to its player as quickly as possible why it's worth playing. One way to go about this is to reveal great depth straightaway, such as by rapidly exposing complex game systems. A classic example of this is the character creation screen in a role-playing game. In Icewind Dale, an excellent old computer RPG, I spent probably a good two hours just making my party of characters before ever beginning play. The character creation system was just so rich with possibility. These days I think it's more fashionable to keep stuff hidden and not scare away the player with too much information up front. That's fine, but finding the right pace at which to reveal new elements of play becomes all the more important in those cases. I'm not sure that proper pacing can be taught, because it's resistant by its own nature to being reduced to a formula.

So then, if you're a game, keeping some of your best ideas hidden away for your later stages is arguably a risky proposition. In fact, by doing so you are implicitly accepting that some probably rather large percentage of your players will never see that content. In Bastion's case, we invested heavily in the ending, by scheduling a bunch of time for a bunch of unscheduled stuff, because we wanted to do everything possible to make sure players who finished the game felt rewarded for their time and effort. From a clinical production perspective maybe this was a bad decision. One could argue that we should have disproportionately focused on only the early levels in this fashion, because more players would see them. The reason this mindset is wrong, to me, is because it ignores who the game's audience really is.

Say you're an author writing a novel. The idea that you'd short-change the ending because not many readers would get that far is deplorable. You need to have faith that your readers will get there, ought to be focused on providing every reason for them to get there. Then you save the best for last for these people, because you owe them. They're the ones you're writing for. As for the ones who don't make it, sure it's probably your fault they gave up, but you can't just go in assuming they won't stay interested because that would make you a hack.