AGDC: Raph Koster On Designing For Everywhere

Areae president Raph Koster, designer of Ultima Online and previous CCO of SOE, gave his talk at GDC Austin in front of a full-capacity crowd, all of whom were eager to catch a shred of what he’s been talking about for the last year or so: how the web is destroying games in terms of revenue and access. “As you might know,” he began, “I’ve got this new startup thing going. I want to talk about some of what we’re doing… but I can’t." "For the last couple of years now I’ve been paying a lot of attention to how the web works, and how the web is eating the game industry’s lunch. Despite WoW we’re rapidly approaching a point where out of the top 10 MMOs, 7 or 8 of them are on the web, don’t sell on a CD and have a million users.” Areae's Plan - 'Play Anywhere' Koster wants his company, Areae, to follow this model, and have a concept of ‘play anywhere.’ This means being on the web, being on all platforms, and dissecting why those models are successful. He then got into a technical design discussion of how you go about that. Koster subscribes to the idea that games can be broken down into elements of game grammar, like a (post) structuralist reading of a film. He showed a slide with a bunch of networking sites, and pseudo-games by the traditional definition, such as ARGs, Line Rider, Hotornot, Club Penguin, and fantasy football. “All of these,” he says “are more popular than EverQuest. Why? After pondering what these things have in common, I think we as an industry have really gotten it wrong about what games are.” Koster went on to say that games are not the interface, and they are not the display. So you should cut those things out of your design, if you want to design for everywhere. It’s really hard, he admits, but all those other things do it. “If you’re like me, you’re really tired of hearing about Web 2.0,” says Koster – but he maintains that the elements of the concept behind the buzzword are sound. “The thing is we don’t trust the users to modify our [game] assets,” he says. Things like MySpace allow people to modify the source code, and determine the experience they’ll have. New Paradigms For The Game Industry Koster mentioned a recent statement by the co-chair of Columbia records, essentially saying “our industry is dead.” People are filling iPods with free content, and nobody wants to pay, and suggesting that the game industry is getting to be the same way. The Areae co-founder suggested that the industry monetizes people who even just want to try a game, and of course people want to try before they buy, which is spoken to by the success of XBLA and PSN downloadable demos. He also noted: "In the game industry we try to have the blockbuster openings – those are easy to ruin. The web is about word of mouth." “The value in web products is not in the content,” he asserts, “but in the data of how people use the content. Amazon’s real value is in what books you’ve already bought, because that’s how they sell you more books. The whole web is based on metadata.” Koster says that all these companies are trying to figure out how to get away from opening big, from high production values, and from monetizing trials. Game Grammar On the lower technical end of his grammar discussion, Koster says “There are very few game systems – we’re usually plugging together things we’ve done before.” This means things like how fast users can press the button, traversing spaces and weighted graphs, determining trajectory, etc. Odds calculation is the hardest one, he says, because the human brain is really bad at it. He says games are inherently social. All games, even single player ones, are two player. The other player is an algorithm, or the computer. When you play against Space Invaders, they’re playing against you. There are games where the other player is playing the same game, he says as in Space Invaders (goal is shoot the enemy), or where they’re doing something different, like in Pac-Man or Donkey Kong. There are parallel games, which are like in a footrace, where you’re playing the same game (against physics in this case) where you’re rated at the end for how well you did. Competitive games are playing the same game against each other. Koster suggested that parallel games are infinitely scalable – but competitive games are not. Team based games always support limited capacity – you only need so many healers in your group, and there’s always a point where adding more roles just breaks down. Bridging the Gap He continued that World of Warcraft has massively parallel teams, working against an asymmetric thing, which is a monster, and lots of people can do it at the same time. This is a step. Koster continued by suggesting that the hot platform is the net. The net says the platform can be anything - there aren’t real hardware requirements or interface problems. The hot topic right now is the non-gamer. The hot feature is other people (as in YouTube), not the systems we write. The hot technology is connectivity and simultaneity. He added: "The hot game is a mini-game. Really small games." The big thing in the traditional game industry now is haptics – but those aren’t designed for everywhere. “Look at the challenges people have putting a game on the DS and... anything else,” he says. “When you look at the kinds of problems we ask people to solve, and the things we assume them to do, it’s like we’ve given them a PhD in mathematics. No wonder you sit mom down and she asks 'how do I move?'” So how do I run everywhere? Interfaces are a huge barrier - they just get much more complicated, Koster suggested. There’s information overload, and to prove it, he showed a screen from WoW, with scads of boxes and statistics on it. “If I look at that WoW screenshot,” says Koster, “I see a user interface begging to be simplified.” He calls for something along the lines of just showing the most pertinent information – and already there are hacks to do this. “Every time you make an assumption about inputs or output, you’re shrinking your user base. This is really the secret behind the DS and the Wii – it’s mapped to stuff we already know, which reduces the learning curve.” Here’s what works in this new model, according to Koster: - the system is the game, not the interface, not the presentation. - any button will do. - long phases take your time – response time is rough. - be done fast, once you’ve made a decision. - do it side by side. Has to be massively parallel. - extended accumulated state – save your profile. - no roles – classless – teams are deterministic. - representation agnostic – draw it however. - open data – change it however. Things that don’t work: - twitch games. - Inputs that are locked to commands – dance mats, styluses. - Models that rely on specific representations (ie 3D). - Models reliant on prior art – if you haven’t played every RTS you’re screwed. - Narrative lock – if you tamper with our story, it won’t be good! Parallel models: - Badges (achievements) - Ratings (skill or social) - Rankings (high scores) - Reviews (and tagging) - Gifting - Networks - Leagues (segmentation) The grammatical elements of those successful ventures he mentioned hit a lot of those grammar points. “The games we’re making today are really bad at hitting those points,” he says, and as such limit their audiences. “There’s no reason why WoW couldn’t be represented by anything other than an RSS feed, and if you could, it’d probably be doubled in users.”