Third Helix is now Kickbomb Entertainment! Please visit kickbomb.us for all future updates.


15 Mar 2010

The last session on the conference schedule was "Metaphysics of Game Design", hosted by the pseudonymous "Phaedrus". I didn't think anything of it at first, planning instead to attend the concurrent "Designing Assassin's Creed 2" talk hosted by Ubisoft's Patrick Plourde. But late Friday I caught wind of a rumor that "Phaedrus" might actually be Will Wright and... well, sorry Patrick, I loved your game, but I have my priorities. ;)

The rumor turned out to be true, and Will Wright walked on-stage to thunderous applause. As it turned out, the title "Metaphysics of Game Design" seemed a bit misleading: Wright spent the bulk of his presentation talking about perspectives and models, networks and communities.

Much of the material was familiar: Wright has been presenting and refining his "conceptual models" thesis for a few years now, and it gets better each time he revisits it. From my perspective, at least, this idea is quickly becoming Wright's signature theory, but the familiarity of the subject in no way diminished the value of the session.

He opened with a discussion of perspectives. He referenced Zen and the Art of Motorcycle Maintenance -- which he described as one of the most influential books he read in his youth -- and its concept that one's mindset while repairing a motorcycle is very different from one's mindset while riding one. Both capture elements of the truth, Wright said, but neither is the full truth.

He asserted that "perspectives are more valuable than solutions." As a result, he explained, "I've gone through my life collecting lenses." He talked about digital photography and likened it to "training wheels for the eye". A photographer is always looking for new perspectives on the world; pretty soon, he said, you can actually throw away the camera and still see the world intuitively from all those perspectives.

He came back to his signature idea of conceptual models, suggesting that play *builds *models, while games (formalized play) *test *models. So how are conceptual models formed, and what use are they to us?

Wright outlined a process by which we abstract data into metaphors, identify patterns in metaphors to form schema, and aggregate schema to build models. In game terms -- and this is my own interpretation, here -- we start with some data, like our avatar's physical state or the amount of ammunition in our weapon. That data is abstracted into metaphors: a health bar and an ammo counter. We can identify a pattern to these metaphors to form a schema: the concept of a HUD that tracks our game state. And we can aggregate multiple schema -- the HUD, the current game mode, our clan or party -- into a conceptual model of the game itself. And that model is just one perspective on the objective reality of the game; my model and your model, like the mindsets of a motorcycle owner, are each partial truths.

Wright gave an example of a child playing with a Thomas the Train Engine toy. By playing with that toy, the child is building many models simultaneously. He's building a model of real-world physics: gravity, friction, the wheel and axle. He's also building a model of the Thomas character, which he can recognize from the TV show (and indeed, the child is aggregating multiple schema, such as a "toy" schema and a "character" schema). He's even building a conceptual model of real-world trains.

We're constantly refining our models, Wright said, and they get better and more accurate as we age. Learning is the process of improving models, of building new schema. He referred to a theory of psychology that suggests that people spend about the first four years of their lives constantly tearing down and rebuilding their conceptual models entirely from scratch. This may explain why most people lack memories of their early childhood. Memories, he suggested, only have meaning in reference to models: early childhood memories would have been formed against models that were subsequently rejected, thus invalidating the memory. In theory -- and this, again, is my own interpretation -- if we could somehow access our earliest memories today, they would probably be completely nonsensical because they wouldn't line up at all with our current models of the world. The animated TV show *Bobby's World *and the popular comic *Calvin & Hobbes *both seem to be built on this concept, or something very much like it.

We build up, tear down, and rebuild conceptual models of the world around us. Those models are perspectives, partial truths, ever-shifting and always subject to new information, new metaphors, new schema. This raises the question: if everybody's models are different, how can we design games that everybody understands?

Wright's answer seemed to be to design for networks, rather than individuals.

There are layers upon layers of networks. There's a network of technology: the television, the cellular phone, the game console. Within the game console is a network of platforms: Xbox Live, the Playstation Network, Steam. Within the Playstation Network is a network of games: Warhawk, Flower, Gran Turismo. Within Warhawk *is a network of systems: combat, game modes, community features. *Warhawk's community features link to a social network: your clan mates, friends, and nemeses. And on and on and on it goes.

Complex networks, Wright explained, tend to specialize: they ultimately become reflections of their communities. Communities, he said, are "fractal and nested". For instance: there are communities of gamers; within those, subcommunities devoted to particular genres; within those, subcommunities for specific games; within those, subcommunities for particular play styles. Korean professional *Starcraft *players are one example of specialized network.

Looking at networks affords us a way to address many models in aggregate. The more specialized a network is, the less variation there is likely to be among the models of the people within it. This may explain why it's so difficult to design successful mass-market entertainment: the mass-market is a very broad, unspecialized network where everyone sees things from radically different perspectives. It may also explain why players of very niche games -- like *Dwarf Fortress *-- seem to be so much more devoted to their games of choice: the niche market is a highly-specialized network likely composed of nearly-homogenous models.

Media itself is rapidly becoming a tightly-integrated, specialized network. "One of the interesting things about entertainment in general is that we have these dedicated silos," Wright said, referring to specific media forms like radio, television, film, and games, but "interdisciplinary entertainment" is the wave of the future. He argued that we need to become "entertainment designers", building experiences around the intersections and not the silos themselves.

My interpretation, again: reality itself can even be thought of as one enormous network, perhaps a network of experiences. Play is a subset of reality, and games a subset of play. Games in general are one conceptual model for understanding life, and at the same time, games encompass many specific models, many varied perspectives. Games are connected to a network of gamers, who belong to one or more social networks, which in turn belong to the internet, which itself is both a model for and a perspective on reality. And so the loop is closed.

Wright's thesis is, itself, a conceptual model of the world we live in, and this writeup is my conceptual model for his thesis. In the end, it turns out that perhaps he was talking about "Metaphysics In Game Design" much more than I realized. ;)


Will Wright is a fantastic speaker, but his ideas come fast and dense. Synthesyzing and analyzing his talk was incredibly intimidating and a significant challenge. This writeup -- moreso than any of my others from this year's GDC -- contains a liberal amount of my own interpretation, due in large part to the fact that I had a hard time keeping up with him while taking notes on the session. But to be completely honest, this subject matter is really inspiring, and I frankly couldn't help but examine patterns, correlations, and perspectives.

I only hope I've done the topic justice.

Posted In:

game-design gdc video-games