So in accordance with my New Years resolution, I’ve removed all political and news blogs from my news reader — we’re down to all people I know in real life, plus a couple of Internet acquaintances, plus a few more total strangers who are nevertheless chock-full of crunchy awesomeness . This is all well and good in terms of productivity. But right now, the evening after the Iowa caucuses, it’s striking me as a particularly boneheaded move. Go ahead, laugh at the addict pathetically flopping around in the throes of withdrawal; I would love to read what all the chattering chatterheads are chattering about. Must… be… strong!

Anyway, one of the aforementioned strangers who made the cut is Timothy Burke, a professor of History at Swarthmore College (where I spent a semester on exchange back in the mid-90s, an experience that almost certainly saved me from transferring or dropping out). One of Burke’s areas of interest is games and, of course, the history of games. In a recent post, Burke talks about how even the more acclaimed recently-released games offer little in the way of storytelling and open-endedness:

I am more pessimistic about storytelling, however. The truth is that almost none of the current generation of game designers are good storytellers. Even the best of games rarely rise to the level of being proficiently derivative narrative engines. Look at Mass Effect, a game whose storytelling has been widely complimented. The main plot is pretty much a Science Fiction 101 space-operatic mash-up. Galactic civilization, many alien races, ancient progenitors, even more ancient menace which periodically swats down galactic civilization, humanity struggling to claim its place in the stars. It’s more a platform for character development and for interactive participation, which is what the plot mostly needs to be in a game of this type. As such, it’s great. Write it out as a novel and it seems like fairly thin gruel.

And:

The other direction where there could be some kind of evolution in games-for-gamers would be towards more emergent or “sandbox” kinds of gameplay. Designers like to claim that their games already accomodate this kind of design, but that’s largely wrong or misleading most of the time. This is one reason that Assassin’s Creed disappointed a lot of gamers. They expected it to be a very open-ended environment filled with NPCs who had autonomous-agent AI, where the player decided when and how to carry out his objectives. In the end, it was a fairly scripted game with a lot of repetition. Bioshock seems to me to be a good example of where this kind of element is really lacking. It’s set in a huge, interesting world, but the player is riding the amusement-park rails the entire time. Any time you might want to get out and explore, there’s a conveniently impassable obstacle.

These are just snippets from Burke’s piece; it’s worth reading the whole thing to get the entire context. But the basic arguments are the familiar ones, that most game plots are weak and constraining.

I think these familiar arguments are pretty obviously right. Sure, we can point to an exception that really blew us away for one reason or another. Grand Theft Auto, I Have No Mouth And Must Scream, and so on. Burke himself cites Deus Ex. But these sorts of games are few and far between. The gaming experience is not, right now, about superior story or freedom.

At the end, Burke is still optimistic that games could be much more than they are. I am less sanguine. As Burke acknowledges, the reason designers “don’t let their players climb off the amusement park ride and peek behind the scenery” is because it would be hideously expensive to implement.

But there are more worrisome issues than mundane cost concerns. Simply put, I’m not sure our culture knows how to tell the stories that Burke would like us to tell.
I blame Laura J. Mixon for planting this bug in my head. Laura agrees with Burke that games need to have better stories and offer more freedom. Laura is a passionate storyteller, and she firmly believes that professional storytellers are essential for games to really start making progress.

She’s right of course, and yet the core problem is that our storytelling forms are particularly ill-suited for the kinds of innovative games we all want. Short stories, novels, plays, movies — all force you to stay on the rails of the amusement park ride, as Burke puts it. That’s just how we’ve been trained to tell stories for the last few tens of thousands of years. Other than the Choose-Your-Own-Adventure novel, nothing that storytellers produce is anything like what game critics are asking for. (It’s a sad sign that most games praised for their innovative, free-form plots are simply Choose-Your-Own-Adventures, except with far fewer decision points than the paperback versions.)

And no, the answer isn’t “just invent better AI.” I think we know enough about game AIs to populate game worlds with interesting characters. Not novelistically realistic characters, but characters who react to our avatars in interesting ways.

What we don’t know how to do yet is fully unleash those AIs. That is, we can’t populate a town with simple AIs, drop the player in, and let the story evolve. Players want freedom to act, but they also wants to achieve goals, participate in the main plotline, be the hero. How do you make sure that key supporting characters don’t get killed or taken out of the story too early? How do you prevent the game from falling into a state where the player, through no fault of her own, can’t ever win? Or worse, can’t do anything interesting?

I believe that we can’t construct satisfying stories in this manner until we can develop a new breed of storytellers. People who don’t just understand story, but have a deep mathematical understanding of the web of the story’s possibilities, and can tune the myriad AIs accordingly. Not just yarn spinners, but actual story weavers.

In short, we have the computational power, but we don’t quite have the theory — and certainly not the practice. I’m guessing it’ll be twenty to forty years until we get people who are recognizably good at this, fifty to one hundred until they start showing up en masse. In the meantime, while we’re waiting for those people show up, we might as well keep playing around with shaders.