April 5, 2004

Whither Game Research

by Michael Mateas · , 11:32 am

I had a great time at the GDC last week. The positive buzz around Façade led me to consider once again the issue of whether the phrase “game research” makes sense. To cut to the chase:

So neither industry nor academia will do the non-incremental work necessary to explore these hard to reach regions in design space. Who will? To put a finer point on it, how do I fund the next Façade?

“Game research” can mean many different things. For the purposes of this post, I’m not referring so much to game studies, that is, theoretical analysis of games, but to work at the intersection of design and technology in which new game genres are invented. In this style of research there is of necessity a feedback loop between design and technology; design suggests new directions for technical research, while new engines and infrastructures suggest new directions for design.

Interactive drama is an example of a non-incrementally reachable region of design space. Trying to reach one of these regions through an incremental series of shippable products is like trying to get to the moon by climbing trees. When you climb a tree it does get you closer to the moon. As your tree-climbing skills improve you can climb even taller trees. But no matter how good your tree climbing skills are, only a radically different approach, like building a rocket, will get you to the moon. Of course, the first few times you build a rocket, it will explode on the launch pad, or dive into the ocean, or engage in other such shenanigans that so disturb funders and publishers. But if no one builds rockets, then nobody gets to go to the moon. Academia is a good, relatively cheap place to house this high risk, high payoff research.

These hard-to-reach regions of design space can’t be reached by design-only approaches, necessitating integrated work in design and technology. As we’ve discussed here before, interactive drama is an example of this; there is no magic design perspective that suddenly makes interactive drama doable without significant technical research in socially and emotionally believable characters, drama management, and natural language processing. While any one of these research areas is enough to keep you busy, the only way to understand how the three major components integrate, and how these integrated components push and are pushed on by story and gameplay design considerations, is to try and build an integrated infrastructure and a complete experience. This is game research.

Maybe by this point the reader is willing to entertain the notion that it is good for the future of gaming to have people doing game research as I’ve defined it. So now the ugly question where this all falls apart – who funds this activity? I’ve participated in a number of academia/industry panels and discussions over the years trying to figure out what kind of productive relationships might exist between the game industry and academia. In my experience, these discussions have never moved beyond industry saying “if you teach students what we want you to teach, we’ll be happy to hire them when they graduate.” Though I didn’t attend the panel “Towards Relevant Research: Collaboration 101” at GDC this year, from what I hear it was the same old story, with academics falling over themselves to make their students as sweet and juicy as possible (anyone who attended the panel, feel free to post a comment contradicting me; but this is the buzz I heard, and based on my past experiences in such discussions, very believable). While preparing students to work in industry is one function of academia, it is certainly not the sole function. If industry believed “game research” wasn’t an oxymoron, perhaps genuine collaboration on problems (including funding) could happen. At GDC a couple of people mentioned to me that their companies are starting or have recently started a research group; I take this as a good sign for the future of research collaboration. Note that such a research group is different from an engines group. While an engines group may be involved in advanced engineering work on, say, the latest physics engine, this is still work in the service of an established design approach (genre or set of genres). A game research group would create new genres, engaging in a combination of design research and technology research.

As far as academic funding, the battle is still ongoing to define game research at the intersection of design and technology as a legitimate, fundable activity. My own work in game research focuses on AI – as I’ve argued before, AI-based art and entertainment (which includes game AI) really is a new research agenda, not a simple application of existing AI research. In order to fund this work through traditional academic funding venues thus means translating an Expressive AI research agenda into terms that funders of traditional AI research can understand, certainly an uphill battle. To make matters more challenging, the funding must support interdisciplinary teams of artists, designers and computer scientists (ideally, these roles are fulfilled by the same people, but the issue of artist/programmers (1 2 3) is an uphill battle for another day).

I don’t yet know a workable model is for this kind of research. But I hope we find one – otherwise we’ll never get to experience those non-incrementally reachable genres, like interactive drama, that lie waiting to be invented.