Emergence (and some devastation) in Sim City

I'll admit. I was one of those kids. I never had the patience to start from scratch in Sim City, to take a humble town to a bustling metropolis, paying constant attention to the state of my citizens and my budget. No, I would eagerly wait through the painfully long time it took my dilapidated eMachines desktop to "reticulate splines" so I could load up the game, punch in cheat codes to give myself an effectively infinite pool of cash, quickly build a behemoth of a city, and then proceed to the only sensible next step: I blew it up.

I can't imagine I was the only one to do this. It was just so fun to see how many disasters I could throw at my unsuspecting citizens at once. (A tornado, earthquake, and alien invasion all at once? There's one for the history books.)

Yikes. I hope that doesn't sound too morbid. Maybe things would be different if I were a ten year old playing Sim City today, though. See, in previous versions of the Sim City games, your city didn't really have any people living in it. Sure, you had statistics telling you about crime levels, employment rates, how happy your denizens were, and so on, but at the end of the day these numbers were just that, numbers. The game ran what amounted to a variety of macroeconomic calculations to give the player an aggregate view of the state of the city and its citizenry, but these values weren't really arising from the actions of simulated people living and working in your simulated city. But with the latest iteration in the Sim City franchise, simply title "Sim City" (EA seems as prone to title recycling as other game studios lately) and released earlier this week, that all changes. Quoting the game's creative and art director, Ocean Quigley, as reported in Farhad Manjoo's excellent review of the game for Slate:

"We now have every car and every person and every garbage truck and every criminal represented by an autonomous agent in the environment. Then we give each of them simple rules about how they should behave, so a criminal goes around looking for a place to commit crime, and a policeman goes about looking for places where crimes are being committed. What you get is a city built out of the emergent interactions between all these agents."

So perhaps ten year old me would have been a little less inclined to level a city if I knew there were actual simulated people running around in there, going about their simulated lives. Or perhaps not. But let's not get caught up with my youthful lack of empathy for my Sims, and instead dig into this notion of "emergent interactions" a bit more.

Emergence is a tricky concept, the precise definition of which is often hard to pin down. The basic idea, though, is that in a complex system with many interacting components, there can arise system-level effects that are not simple additive outcomes of the constituent parts, and are not easily reducible to those parts. In other word, the whole is somehow greater, or at least different than, the sume of its parts. A compelling example of emergence is the flocking of birds, where many individuals, each following a relatively simple set of rules, generate unpredictable group-level patterns of immense complexity. Researchers often use the tools of multi-agent modeling to study emergent phenomena, simulating individual agents endowed with simple behavioral rules in a computer, and observing how the group behaves when many such agents interact. In the case of flocking, a common technique is to have each agent adjust its speed and heading following three simple motivators: attraction (if you're too far away from the agents in the vicinity, move towards them), repulsion (if you get too close to another agent, move away from it), and alignment (try to match the heading of agents near you). With just these simple rules, agents will move in groups amazingly similar to real flocks of birds.

The are plenty of other examples of emergent phenomena, but the key idea is that interacting agents, even when operating under very simple rules, can generate interesting, unpredictable effects. And this is what makes the new Sim City so exciting and different from its predecessors. To again quote Ocean Quigley, "In the old model, the simulation would never do something you wouldn’t expect. That’s because the simulation designer had bounded the potential states that the system could get into, so there was no novelty, no emergent behavior that could come out of it. Everything that happened in the simulation was already defined upfront." But not any more. Now, individual agents, governed by simple but reasonable rules, interact within the structure of the city that you create, lending a new level of complexity and realism to gameplay.

Game reviewers will of course tout the the implications of these changes for the realism and difficulty of the game, but as social science researcher I'm more excited about the fact that EA has created a giant, interactive multi-agent simulator, an amazing sandbox for teaching players about emergent phenomena. Having a realistic city simulator such as this creates opportunities for kids (and adults) to learn about things like energy consumption and conservation, government structure and dynamics, and geometry, to name but a few. Sim City Edu is leading the charge along these lines, and for each example I mentioned there exists a Sim Ciy-based lesson plan on their website. The project is a joint venture between EA and the Institute of Play's Glasslab, and shows great promise for harnessing the power of the new Sim City for use a classroom tool.

The project is still young (they only have four lesson plans up on the site at this point, all aimed at grades 6-8), but it's heartening to see that the educational potential of this game has been recognized from the outset. What I will be excited to see is if educators can develop lessons for more advanced students to teach abstract topics around emergence. As I mentioned, emergence can be tricky to define, let alone conceptualize, but with a game like Sim City players get to see emergent phenomena in action. Importantly, these phenomena are not presented as abstract models or descriptions, but experienced in the context of a relatable system. Of course, it's entirely possible that the game is too complex in its normal state to be an ideal teaching tool, so it will take work to develop mods of the game for specific teaching scenarios. Still though, there are many possibilities here, and I'm excited to see how what kind of teaching tools are eventually born of the game.

Emergence (and some devastation) in Sim City by Jared Lorince, unless otherwise expressly stated, is licensed under a Creative Commons Attribution-ShareAlike 3.0 Unported License.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>