Web 2.0 Expo Preview: Will Wright, Sims and Simulations

by Kurt Cagle | comments: 7

Will Wright has been the foundational genius behind a thirty year string of blockbuster games, from the early Raid on Bungeling Bay in 1984 to the first truly fun urban simulation Sim City, a game "universe" that let players create and manage their own cities, dealing with everything from balancing budgets and battling crime to dealing with the aftermath of alien attacks. This game was later expanded to SimCity Societies to better explore the larger social factors that shape society.

From there he delved deeper into the lives of the individual inhabitants of those cities with the Sims, a virtual "dollhouse" that gives players the ability to shape the eponymous sim-people, their houses, careers and relationships (and in subsequent installments, let them start businesses, party, go to college, have pets, and take vacations, among many other activities).

In 2008, Wright produced Spore, where the players can "play gods" - raising new life from Sim-ooze to intergalactic civilizations in a freeform multiplayer environment that's evolving nearly as fast as the spores themselves. Scheduled for June 2009, Wright will release the much awaited Sims 3, in which for the first time, the Sims world comes together in a full immersive environment, perhaps the full merger of Sims and Sim City.

Wright will be speaking on the Sims and games in general at the Web 2.0 Expo in San Francisco. O'Reilly editor Kurt Cagle caught up with Will Wright to ask a few questions.

Kurt Cagle: You've been doing this a long time. I can remember distinctly playing Raid on Bungeling Bay back on the old Commodores days back in the late '80s. The thing I find fascinating is every game that you've done in the last 25 years or so would more actually be considered a simulation rather than a game. What have you found most fascinating about simulations as games?

Will Wright: Well, as a kid, I spent a lot of time getting models and an inordinate amount of time dealing with the plastic and with models. That kind of got me into robotics which was kind of a different form of modeling. I bought my first computer, which was an Apple 2, to connect to my robots to control the programs on that. It wasn't too long before I started doing little simulations of the robots I was working on on the computer, and I started realizing this was kind of a new way to make the models that I'd kind of grown up making, except these models had dynamics underneath them rather than just static structure.

So I guess I've always been fascinated with modeling, even before I got a computer. For me, models are a way of coming to understand the world. I mean I found that as I built models, it would be around subjects that interested me and in building a model, it would eliminate sort of the aspects of the structure of something to me in a really deep way. On the computer as I started building these really simple little programs, I started realizing this was like the perfect place to build -- you know, it's basically your ultimate modeling tool.

Even robots in some sense are a model of human ability, whether they be physical or mental abilities that we have as humans, when you attempt to build a model of them, you start to realize how incredibly elaborate our abilities are and how hard they are to recreate, and it kind of makes you appreciate them in a fundamentally different way. It's the same thing when you're trying to simulate some aspect of the real world whether it's a biological system or an economic system, as you start building models of them, you start seeing kind of the unexpected levels of emergence, complexity, kind of subtle interplay development, et cetera.

Kurt Cagle: Perhaps one of the greatest models that you've done, and you've done some spectacular ones, has to be SimCity. It's gained a wide following not only among gamers, but I suspect there are probably a few city planners and economists that also have spent a little more time than they probably care to admit playing around with it. It's easily one of my favorite games. What went into the creation of that? What started the process and how did you go about modeling something as complex as a city?

Will Wright: Well, actually, it kind of came out of Raid on Bungeling Bay which was a super shoot 'em up where you bomb these little islands. I had to create an editor where I could scroll around and create these little worlds to bomb. I found in playing with the editor that I was having a lot more fun creating these worlds than I was blowing them up.

After I finished that game, I started kind of playing with the editor, and I started to add a little bit more dynamic to it. I wanted to keep traffic moving on the roads. I wanted to see kinds of dynamic systems operating. So I started reading about city planning. At first, I didn't find the subject that interesting, but as I started developing very simple simulations of the theories I was reading, it became fascinating to me, because I had a little guinea pig city that I could sit there and experiment and play with ... and that totally changed kind of the nature of my relationship to the subject.

Then I got very interested in the other work in simulation and all of the work with Jay Forrester, with system dynamics, who interestingly enough, was also probably one of the first people that actually had built a digital model of the city. He wrote a book called Urban Dynamics. Although, his model was not spatially located (it wasn't on a map). It was just a very simple kind of system dynamics model with population totals and things like that.

But cities as a subject, I found fascinating just because of all of the different layers of infrastructure and statistics. You have crime statistics. You have population, pollution, traffic. All of these things are overlaid and interrelate to each other, and they interrelate in really interesting ways across different timescales. When you have a toy city and you can kind of run it through a mini-decade, you start seeing the city as a very organic process. When you just drive around the city and look at it, it looks like a very static thing. But when you can put it in fast-forward, it starts looking very much like a living creature and in fact, shares a lot of the same kind of infrastructures, you know, communication systems and circulatory systems, waste disposal systems, et cetera. And so I just kind of fell into it in that sense.

Kurt Cagle: The games have become considerably more complex over the years. When you're modeling them, how much of that modeling, how much of that creation is basically independent agents essentially reacting with the rest of the world? How much of it is coming from external data about how cities work? And how do you update that? Obviously, things like what's happening right now on the financial markets and stuff like that, does that impact the way that you're putting together these worlds?

Will Wright: Well, there's a lot of different methods of simulation or prediction. One of them is kind of Sims analysis, or in stock terms, it'd be called technical forecasting, where you look at past data and you try to recapture that.

Then there's kind of fundamental analysis where you're trying to fundamentally recreate the behavior system by modeling its components. That's closer to what we do in our simulation work and so they're not very data-driven. It's more like we're building algorithms and playing with those algorithms a lot until we get something that looks reasonable. Depending on the type of simulation we're doing, we'll build test cases that we would expect that if A, B, and C happen, the simulator should react this way. And we were kind of playing with the underlying algorithms against these test patterns across a number of different dimensions of behavior.

But, for the most part, we're basically exploring an emergent system. And because it's emergent, by its very nature, you can't sit there and engineer it top-down. What we have to do is we have to sit there and kind of play with a wide variety of algorithms and structures. Turn them on. Observe the behavior. Then when it doesn't quite do what we want, we go back to the drawing board. We refine it a little bit more. But it feels much more like the process of exploration that is in engineering.

Kurt Cagle: Have you found with this kind of approach that it's something that you can take to policy makers and say, "Maybe you need to start thinking about using this kind of bottom up approach to development?"

Will Wright: I think that's kind of a realization that's happening across a number of fields. It's been happening over the last 10 years. We basically live in a bottom-up world. We might occasionally try to impose top-down structures on that world. The economy is a really good example of that. At the federal or government level, you can try all of these regulations and tax policy and interest rate adjustments, but at the end of the day, you have millions of independent agents out there basically making decisions based upon their confidence of the future and a number of other factors that are basically the tide. The primary behavior of the system is coming from the bottom-up. And what we thought were these high level controls and leans on the system are getting looser and looser. In fact, they never were that tight.

Typically, we would have a boom economy and they'd go back and say, "Well, the feds did a great job." But really it had nothing to do with the feds. It had to do with underlying cycles that were happening from the bottom-up. But I think that this is kind of interesting in that, and it depends on how deep you want to go down this.

I mean you can go all the way back to Aristotle who was trying to turn science into a reductionist method; the idea that you could understand a system by picking it apart and understanding its components, you know, it works in certain areas, but when you get to a sufficient level of complexity and you have emergence occurring, you really can't look at the individual agent and predict what their overall group behavior is going to be.

So this is a fundamental limitation that we're finding across all fields. And I think the fundamental realization that policy makers, modelers and forecasters, are seeing is that the world is also getting more emergent and obviously more interconnected in economics and in an environmental form and in a kind of personal social sense. All of these kinds of things are forming a more dense inter-web connection between these systems which makes the emergent factor even higher than it was before.

Kurt Cagle: Changing subjects a little bit, my nine-year-old daughter has become a hardcore Sims fan. I can barely keep her off the computer. It's kind of like, "Daddy, you're home. Can I play on the computer now? I want to play Sims." I think she sees it as this is the ultimate doll-house. Obviously, there's been a lot of thinking about psychology and human interaction with this. What challenges did you face when you were putting together the Sims? What were you trying to do with it? What have you learned from it?

Will Wright: Well, one of the first challenges was could we develop a really robust model of human behavior test so that we could put these little characters in the elements in any situation they would behave reasonably. A Sims user is kind of controlling the environment. They can put the Sims in a wide variety of potential environments and then the Sims have to act reasonable. So we kind of had to develop a very -- well, on the surface, it looks like an object-oriented programming model, but in fact, it's what's technically called a subject-oriented programming model. But I won't get into that detail. So developing this robust kind of behavior system -- in fact, it was environmentally distributed intelligence is the way we solved it.

But on the design side, there was a lot of thought about how much autonomy do these characters have versus how much reliance do they have on the player directing their actions. And there was also a whole dimension of thought around how much are we going to let the player read in to the simulation. In other words, how much of the simulation are we going to offer them to play of imagination versus make very clear and overt?

For instance, when you hear the Sims talk, they speak in this kind of gibberish language. We've actually recorded hundreds of lines of dialogue with a lot of different emotional nuances, but you never actually hear the words. We found that they seemed a lot more real when the players were hearing it as a foreign language. They were kind of able to read emotional content into this. "Oh, they're arguing. They don't look happy."

And rather than hearing the actual words, which would start sounding very repetitive and very robotic right off the bat, the players in their heads kind of fluidly will imagine the conversation.

It's actually amazing how far we were able to take that into different dimensions. On the original Sims, which was kind of isometric, the characters were only so big on screen. And you could just kind of make out their faces. And it was interesting how many players after having played for a long time, if I asked them whether the characters have facial expressions, they would all say yes. In fact, they didn't. They all had a very neutral expression. And it was just small enough to where people were imagining their facial expressions, you know, depending on their emotional state.

So we were able very carefully to rule in how much of the imaging we were leaving blank for the player to fill in.

There was also a whole question of the resolution, you know, of things like when you're creating a house and moving furniture around, were they working down to the last inch or the nearest meter. It was a real balance sort of between ease of use and expressiveness, all the way up to what are the perimeters for Sims psychology. How many dials do they have to kind of tweak to program a Sims character?

Kurt Cagle: What impact do you think Sims 3 will have?

Will Wright: It's hard to say. You know, the Sims I think for a lot of players that are really into it, it's more of a hobby than it is a game. They sit there and it's kind of a creative endeavor. Some of the players go off in different directions and some are very much into the architecture and they've done really cool houses and share them with other people. Other ones like young characters or putting their friends and family in and kind of replaying their life in a weird surreal dimension.

A lot of the players just kind of play it for several weeks and then put it aside. They'll buy an expansion pack. Pull it off the shelf. Play it again for a few more weeks. Put it aside. But they'll actually kind of end up playing it ongoing over several years. And I think the Sims 3 is kind of reflecting this universe that it's familiar to players that have bought the Sims 1 and maybe Sims 2 and maybe a few of the expansion packs. And with Sims 3, every generation of Sims we're kind of trying to bring the walls further out so they have more possible experiences in the game.

Because really, a game like the Sims is all about the player-driven story. The player's the storyteller. We're giving them a system where they can put anybody into it they want. They can put up all of these different situations and play out their own soap opera or drama or whatever. And we find what several of the players do with it, in certain areas where they're trying to pick their stories, but they hit a brick wall or the simulation doesn't support this in whatever dimension. Maybe the behavior of the Sims or something about the environment.

So the Sims 3 is trying to make that storytelling more seamless because we've found these players in a sense too were spending a lot more time having their Sims roam across the entire environment, you know, going to visit their neighbors. Going downtown. Doing this. Getting a job. Going to school. And so with Sims 3, it makes that environment much more seamless really and a smoother storytelling experience.

Kurt Cagle: The other game that you released last year was Spore. I find it interesting because you're now modeling evolutionary processes as well as societal ones, sometimes in a very amusing manner. I've thoroughly enjoyed seeing what I've seen coming out of Spore. How was the game received? What were you hoping to achieve with it besides selling lots and lots of copies?

Will Wright: Well, it's a game that I would say is still very much in development. You know a game like Spore where it's kind of leveraging off of what we saw happening with Sims and user generated content except going to the next level with it where much higher level tools for customizing content within the game is part of the game play.

So our only thought when we released Spore, the game, was that it was about halfway through development. The next step was to see what the fans do with it. And this is kind of what happened with the Sims. When we did the first version of the Sims, we watched what the fans did. And then we started doing the expansion packs relative to what we saw the players wanting. And we're starting to develop other experiences and expansions for Spore right now based upon what we've seen players do.

The amount of content that we've created totally blew us away. It was much more than we expected. I was hoping to see maybe around a million player-generated assets by the end of the past year. Instead, we were surpassing -- I haven't even looked in a while, but the last time I looked, we were passed 70 million. We have 30 million unique creatures in our database and there are only 7 million unique species on earth. So we out-populated Earth in a matter of a few months.

And what's really amazing is the quality of the assets. I mean some of these are far better than any of our internal artists were able to make even when they tried to use the tools for several years. It shows you kind of the value of having parallel exploration of a solution space.

So in some sense, the game is not just about biology, but the very mechanisms through which we're kind of generating these worlds for players is very much like a biological process where you have millions of people exploring this solution space in parallel. Now we're looking at ways to make better use of all the huge data in the worlds the players are creating and give them new experiences with that data. So I'd say Spore is still very much a work in progress.

Kurt Cagle: The educational aspects of everything you do from SimCity all the way through Spore and Sims 3 seems fairly evident to me. What are your thoughts about the way that we educate kids and ourselves for that matter? And what do you see in this ongoing process that you're doing, how do you see that possibly changing that?

Will Wright: Well, I think especially in today's world, it's changing so fast and we, in fact, have amazing educational technologies available to most kids, at least in this country that it's much more important to focus on motivating and inspiring the kids than it is on educating them.

I think if you can get a kid interested in a subject, they have plenty of easy opportunities to go out there and learn about it. The really hard part is getting them interested in it in the first place. And so for me as a kid, there were certain subjects that captured my imagination. This was before the Internet ... I was able to very easily go out and kind of find out about whatever subject I was very interested in, mostly in bookstores or libraries. Nowadays, every kid has the internet on their desktop and whatever they're interested in, they have incredible resources at their fingertips if they're just ready to pull them, and I think we're finding that having kids pull information rather than us pushing them.

Our model up to now for education has been have the kids sit in the classroom; have the teachers get up and talk at them for an hour whether or not the kid is really interested in that or not. So I think in terms of what I do and the way I see technology relating to education is I'm much more interested in trying to take subjects that might seem boring on the surface and turn them into toys, you know? Turn them into a fun experience so that kid goes up and you're not even really trying to preach to the kid, but somewhere inside that toy, even if it's a simulated version of a city or a planet, they start understanding subjects like economics or geology or biology in sort of a fun, engaging way and also in a way where they're an active participant. They're not just sitting back and listening. They have enough information wash over them that they're invited into a kind of very creative involvement.

Kurt Cagle: If you're a programmer thinking about getting into the simulation or games market, what skills should you have? What do you look for in a programmer or a designer?

Will Wright: Well, I think one of the really important skills that we're seeing in both programmers and designers is the ability to do rapid prototyping. And also, we're finding that people with mixed skill sets are far more valuable. So a designer that can do a little bit of programming or a programmer that has a good design sense are both a lot more valuable than a pure designer or pure programmer. And a designer at some level if you understand the way algorithms work, that programming works, it doesn't mean you have to be the one programming the game, but kind of like a carpenter understanding wood grains. If your designer is going against the grain of the machine the whole time, it's never going to feel as polished as it could.

Likewise, a programmer that has no sense of design, there are always a million ways to solve a good programming problem and if the solution is going kind of along the same path as the design goal, it's going to just work better in a number of different ways. It really is just kind of craftsmanship in interactive design. So I would say that's probably at the top of my list.

Kurt Cagle: Interesting. Final question. What's next? Are we going to see Sims' string theory anytime soon?

Will Wright: [Laughs] I'm starting to dive into new projects right now, but I'm just not ready to talk about them because they take me so long to work on and so I'm kind of in my zone right now thinking on the next one.

Kurt Cagle: Well, I want to thank you very much. This has been just absolutely engaging and intriguing. I look forward to what else is coming out of your brain. It should be very, very fascinating.

Will Wright: Thanks, Kurt, for the interview.