Michael Mateas, director of the University of California, Santa Cruz's Center for Games and Playable Media opened up the center's Gamasutra-attended symposium, Inventing the Future of Games, in Milpitas, California today.
In his speech he discussed the drive of USCS's program, and talked about how research drives innovation in other industries -- and how he believes research should also drive games. He also gave examples of how current work at the Center for Games and Playable Media is doing that today.
"What should be the role of the university in games and playable media?" Mateas asked, before launching into a discussion of both the challenges and opportunities facing game developers at present, as well as outlining the work being done at the university.
"If you look at the mission of the research university, what we're supposed to be doing is engaging in cutting-edge research," he said. The goal is "most fundamentally, impacting society, including industry, by bringing groundbreaking innovations."
According to Mateas, "in many other fields this has become routine" -- the internet, biotech, and green energy are shaped at deep levels by research done at universities.
"Unfortunately games research is sometimes considered an oxymoron," he said, since games are media -- and must be experienced by audiences. This is in tension with his mission -- to create "new experiences that can't be done incrementally."
Of course, all developers and gamers are aware with the current state of incremental improvements to existing game technology and design.
Like Trying To Carve Michelangelo's David With Broken Crayons
While at a university level, game studies and design research are well-established, Mateas said, his goal is to bring forward computational media, which he says is "underrepresented."
Computational media "creates technologies, and frameworks, and idioms that fundamentally change the computational and material conditions for the creation of games," he said. In other words, they are "new tools and approaches that allow you to make games you couldn't even think of making."
Mateaes likened current design approaches to trying to create masterpieces with a "cup of broken crayons" -- not only would you not be able to create a Da Vinci, you couldn't even conceive of creating a Michelangelo sculpture with tools unsuited for carving stone.
"There are potential games in this design space that we don't know how to reach today," he said. Examples include interactive drama, dynamically game-mastered MMO, or the Young Lady's Illustrated Primer from Neal Stephenson's novel The Diamond Age -- a fully interactive book that both learns and teaches.
These are all "conceptual games that are somewhere in this design space. How will we get to them? They aren't incrementally reachable. It's like trying to carve Michelangelo's David with broken crayons."
When it comes to commercial games, said Mateas, "The most obvious technical advancements in game have been in graphics." Current work is based around "optimizing the hard-won techniques of artists in the Renaissance."
Scaling Beyond The Choose Your Own Adventure Book
A lot is being left behind, in his view. "What about communicating everything else? There's a whole lot more we want to say in games besides 'Hey, look. Things move in the world?' What about human nature?"
Of course this is done through storytelling in games, but the distinction Mateas makes is doing it technically. "How do we go about computationally representing this huge space of everything it means to be human?"
Mateas observed that "the current conditions of content authoring are lovingly handcrafting content. It has trouble scaling. Why does scaling matter? As you try to make the content that you're representing even more open to variability through player interaction and agency, you have to produce an exponential amount of this content."
In his view, current games are "really not even a step up from the Choose Your Own Adventure book."
He wants to "move toward procedural everything", and approach solutions to games in the way current titles simulate physics.
"What makes [video games] different than all forms of media before," is that they function through systems. "That's when a game is operating at its best, when it's engaging in the unique and powerful communication games can engage in."
Creating Deep Social Interaction
Mateas' program at UCSC is working on a game called Prom Week presently, which is an experiment in "social physics."
"As Angry Birds does not pre-script a specific solution to the puzzle, we want to present the player with social puzzles that do not have pre-scripted solutions," he said.
"Can we create something that has pretty deep social interaction?" One way the game does that is by remembering every social interaction the player has ever had. "It's raw material for dialogue realization. We want the player to feel that this game is really listening to what they do."
Mateas showed several examples the way the game could play out -- when given a goal, the player can approach it through different angles. "You just can't do that by scripting. This is going to exponentially blow up on you. You can't make social interaction deeply playable." This social architecture "is a new tool" in his view.
And though multiplayer games have other players to interact with, they're still based on systems, so this research is still valuable, Mateas argued.
Board games are complex systems -- Arkham Horror "is a cardboard computer," he said. Sports are designed to "leverage the laws of physics."
Even in multiplayer games that engender "subversion and meta-play", this form of player interaction "is not completely free. It's not unconditioned by the rules, it's actually conditioned by the rule sets and systems the designer has set up."
And in "the digital multiplayer space, the primary form of social interaction is kind of the speaking tube. You have a game, and you allow people to text and voice chat on top of that game... But everything passing through that tube is outside the rule system. It's not being taken into consideration. It's a huge missed opportunity."
Mateas sees that "it would be a huge advantage in a game like WoW to listen in on player text chat to learn, say, when two guilds are feuding. A dynamic quest generator could create a quest that the only way to accomplish it is for them to cooperate... Or to bring them into close physical proximity and see what happens."
Taking Handcrafting To A Meta-Level
Another project underway at the school is data-mining of human-to-human gameplay -- it analyzes pro StarCraft matches to learn tactics. This wouldn't just create a great computer player. "What this would enable would be an AI player that tracks the overall metagame." AI could change its tactics "as human-to-human play explores the strategic possibility space of the game."
The school is also developing systems to allow designers to model rules, because, Mateas said, "there are emergent dynamics that come out of your rule choices, and currently game designers have no tools or techniques for thinking or reasoning about emergent dynamics. There's no computational support. There's nothing like CAD... which helps architects think through the possibilities."
The team has devised a "simple, quick" prototyping tool called BIPED. [PDF Link] This allows the designer to input all the game's rules and then query them to find what the player can do... "helping you much more rapidly explore the emergent dynamics of your ruleset."
The notion of authoring fundamentally changes in the context of procedural storytelling, and the team is addressing this. "If you're going to move toward procedural everything, how do you enable human authors to create games in these hyper-procedural frameworks?"
Said Mateas, "I'm not against loving handcrafting; that is what makes a game special. But now that handcrafting has to move to a meta-level."
One tool being worked on is Story Canvas -- a tool for "non-technical designers to create generative interactive stories."
In short, the goal is to move "more and more smarts into the authoring tool, to facilitate the designer in doing more meta-level authoring."
Conclusion
When it comes to research into games, "I believe the role should be to engage in game studies, so that we can deeply understand the cultural and aesthetic possibilities and implications of the games we're creating, and then deeply informed by that understanding, to engage in the intertwined and intermingled computational media and design research -- to enable people to create new kinds of games that we can't even conceive of today," Mateas said.
We are living in a "current historical moment as games are going exponential." He believes this will have "as profound an impact on society as the invention of the internet" and other massively important, academic research-driven technologies.
To that end, he believes, "academia and industry should be working hand in hand." Games should "fundamentally enrich the human condition," and help people become "smarter, more insightful, and better human beings."
In the brief question and answer period after the presentation concluded, academic and developer Ian Bogost asked if Mateas was not possibly overstating the reach of computational media -- it could be method or solution, but not the only solution for reaching the heights he described.
Said Mateas, "I do believe that our underlying computational tools and frameworks and idioms do fundamentally condition the kinds of games we can create. So there are regions of that design space that are unreachable."
"Interactive drama is one of the undiscovered countries we're trying to move toward... there are others. My mission in life is to enable those spaces... I still remain full steam ahead on trying to enable and annex" these areas of design.
internet, biotech, and green energy are academic - game design is creative and entertainment. As far as I can tell, universities don't set creative and entertainment trends at an academic research level. Those projects all sound interesting but to determine how to engineer them, design them, and then art direct them - you have a pretty large full time project on your hands that might as well be funded by a publisher. And unlike scientific research it is not a progressive buildup of knowledge, could easily be wasted time and become quickly obsolete.
Maybe interesting to note that I saw Mateas at Bruce Wilcox's GDC poster session. Bruce was talking about his work on chatbot technology (deciphering what the player is saying. or doing a good job of faking it).
A lot of play testing is about surveying the players, to see what they liked. wouldn't it be better to do this DURING the game? perhaps by having an NPC ask you about your intent? in a game like Mass Effect, wouldn't it be more deeply immersive if your assassin companion asked WHY you're going back to the citadel? (maybe he'd even seem to care about what you're doing? creating a deeper bond between you both?). and if the game used chatbot tricks to decipher your intent (" I wanna buy armor upgrade X from a certain shop"), then it could adjust the ruleset to make your trip more interesting (that shop is the midst of a hold up, by armor thieves).
And if you add technology like Kinect on top of this (voice recognition), this whole interaction could play out very casually (with the user muttering responses out loud, rather than choosing from a dialogue tree wheel). this could lead to some very exciting advances in what we think games are capable of.
currently games are "about making meaningful choices". but they don't really feel like communication. they feel like a set system we go into and mess with. they don't seem to be reacting to our choices. they don't care "why" we're doing something. if we can create the illusion that game cares WHY we're playing it, it'd probably crazy scifi leap forward territory.
...
might also be interesting to note that the Limbo talk at GDC shows that their editor had a built in character simulator that seemed pretty advanced. as soon as the developer tweaked a level, he didn't just run it in engine to see how it looked. he ran it in engine to see how it'd probably be played. That really struck me as a key part of why that game turned out so well.
I am definitely interested to see where you guys get with this. Allot of game dev seems limited by the tools they have, it will be interesting tom see how the academic world plays in the development of narrative and procedural technologies.
It sounds like great fun though.
A lot of play testing is about surveying the players, to see what they liked. wouldn't it be better to do this DURING the game? perhaps by having an NPC ask you about your intent? in a game like Mass Effect, wouldn't it be more deeply immersive if your assassin companion asked WHY you're going back to the citadel? (maybe he'd even seem to care about what you're doing? creating a deeper bond between you both?). and if the game used chatbot tricks to decipher your intent (" I wanna buy armor upgrade X from a certain shop"), then it could adjust the ruleset to make your trip more interesting (that shop is the midst of a hold up, by armor thieves).
And if you add technology like Kinect on top of this (voice recognition), this whole interaction could play out very casually (with the user muttering responses out loud, rather than choosing from a dialogue tree wheel). this could lead to some very exciting advances in what we think games are capable of.
currently games are "about making meaningful choices". but they don't really feel like communication. they feel like a set system we go into and mess with. they don't seem to be reacting to our choices. they don't care "why" we're doing something. if we can create the illusion that game cares WHY we're playing it, it'd probably crazy scifi leap forward territory.
...
might also be interesting to note that the Limbo talk at GDC shows that their editor had a built in character simulator that seemed pretty advanced. as soon as the developer tweaked a level, he didn't just run it in engine to see how it looked. he ran it in engine to see how it'd probably be played. That really struck me as a key part of why that game turned out so well.
It's a joke right?
Watch Mateas's full presentation here: http://youtu.be/CLh_6V5b0OQ