March 12, 2008

EP 8.2: Understanding Simulations

by Noah Wardrip-Fruin · , 6:20 am

The concerns about work such as Wright’s get to the heart of what is involved when we use computer models to make non-abstract media. As Ian Bogost puts it in Unit Operations, “the relationship or feedback loop between the simulation game and its player are bound up with a set of values; no simulation can escape some ideological context” (2006, 99). Or, as Ted Nelson put it succinctly two years before SimCity’s release, “All simulations are political” (1987).

In the case of SimCity, Wright’s inspirations included some decidedly political simulations: the “system dynamics” work of MIT’s Jay Forrester. When applied to urban planning in the late 1960s, these simulations produced anger from many quarters. As Forrester reports in “The Beginning of System Dynamics,” his conclusions included the idea that “low-cost housing was a double-edged sword for making urban conditions worse. Such housing used up space where jobs could be created, while drawing in people who needed jobs. Constructing low-cost housing was a powerful process for creating poverty, not alleviating it” (1989). Forrester’s critics were quick to point out that his models were based on assumptions far from verifiable (e.g., that housing is a stronger attraction than jobs) and the workings of his simulated city were not like those of contemporary U.S. urban centers (e.g., Forrester’s city was of a fixed size, and commuting into the city was impossible).

Of course, any simulation requires simplifying assumptions. A map as large and detailed as the territory is no map at all. And for many things one might like to simulate — such as human behavior — no one knows how they actually operate. So any simulation is actually an encoding of a set of choices. As I discussed earlier in relation to symbolic AI techniques, these choices can be made based on current beliefs in cognitive science. Or, as I discussed in terms of systems such as Terminal Time and F.E.A.R., they can be authorial choices, made for purposes of shaping audience experience and authoring opportunities.

Whatever the motivation behind the choices, there is inevitably a politics to how the world is simulated. This is perhaps most obvious in a genre of political games, many of which operate as very simple simulations. For example, the widely-debated game September 12th actually bills itself as “a simulation” (Frasca et al, 2003). The player is presented with an isometric view of a middle eastern village, around which civilians and terrorists walk in somewhat-unpredictable paths. The player’s cursor is a targeting reticle, which can be used to call in airstrikes. But these take time to arrive, during which the village’s inhabitants continue to move, and in many cases innocent civilians are killed in even a carefully-chosen strike. This sets the stage for the most controversial aspect of the simulation. For a few moments after each explosion, when other civilians come upon the dead bodies of civilians killed in airstrikes, they stop and mourn — and then straighten up and transform into terrorists (figure 1.2).

PIC

Figure 1.2: In September 12th mourners of innocent civilians killed by airstrikes (wearing blue and green) transform into terrorists (wearing white and black). I cropped this series of images from larger screenshots.

The transformation is animated and accompanied with a cartoonish sound. Certainly it is meant as an iconic action, rather than as a realistic portrayal of the nuances how one decides to become a terrorist. But the point of the simulation is a broad one about airstrikes as a tool for fighting terrorism — suggesting, through its model, that the tactics of the U.S. may be part of what recruits anti-U.S. terrorists, rather than simply that they “hate our freedom” or something similar. Gonzalo Frasca, the game’s lead designer, defines the act of simulation as “to model a (source) system through a different system which maintains to somebody some of the behaviors of the original system” (2003). Obviously, the key phrase here is “to somebody.” This is also the missing element in the conception of systems such as Abelson’s ideology machine and Meehan’s Tale-Spin as simulations. They are not correct models of how people think — but they are someone’s representations, in process and data, of an idea about part of human life.

Critiques of SimCity and similar games have tended to focus on the fact that players seem to be required to become part of the “somebody” that buys into their simulation models. After all, successful play is impossible without coming to understand — and act in accordance with — the model. In general, this is tied to a wider tendency to accept the veracity of simulation rules and results when, instead, they should be questioned. For example, Paul Starr, a professor of sociology and public affairs at Princeton University (who served as an adviser to U.S. president Clinton’s administration in 1993) offers this:

While playing SimCity with my eleven-year-old daughter, I railed against what I thought was a built-in bias of the program against mixed-use development. “It’s just the way the game works,” she said a bit impatiently.

My daughter’s words seemed oddly familiar. A few months earlier someone had said virtually the same thing to me, but where? It suddenly flashed back: the earlier conversation had taken place while I was working at the White House on the development of the Clinton health plan. We were discussing the simulation model likely to be used by the Congressional Budget Office (CBO) to “score” proposals for health care reform. . . .

[W]hen policymakers depend on simulations to guide present choices —especially when legislators put government on “automatic pilot,” binding policy to numerical indicators of projected trends — they cede power to those who define the models that generate the forecasts. This is happening in America today, most notably with the rise of the CBO as a power center in national policy. In a sense, Washington is already Sim city. (1994)

Obviously, this is a matter for some concern, given the history of assumptions built into simulations such as Forrester’s. But is it an appropriate basis on which to critique SimCity and similar media experiences? It would be unconvincing to argue that, being media, such simulations should be exempted from close examination. Arguably, what we learn from media representations has as profound an influence on our culture as the decisions of government. This is one of the motivations for Simon Penny’s call for an ethics of simulation media (2004).

Rather, I think the key here is in the comment of Starr’s daughter: “It’s just the way the game works.” Playing SimCity she already understands how the simulation operates, what its underlying assumptions are. In other words, SimCity is separated from simulations such as Forrester’s and the CBO’s in the same way that the SimCity effect is different from the limited-interaction version of the Eliza effect. When interaction with a software model is severely restricted — when we see only Forrester’s “conclusion” that low-income housing is harmful, or the “scores” released by the CBO — the shape of the underlying system is not understood and cannot be effectively questioned. On the other hand, the result of the SimCity effect is precisely the development of system understanding.2

This difference, between the Eliza and SimCity effects, is in turn what underlies Bogost’s response to Sherry Turkle’s critique of SimCity. In Persuasive Games (2007) he writes:

“Opening the box,” in Turkle’s opinion, would allow players to see how the simulation runs, providing better ability to critique. The problem with this objection is that the player can see how the simulation runs: this is, in no trivial way, what it means to play the game. Turkle’s real beef is not with Sim City, but with the players: they do not know how to play the game critically. (63)

Learning to interpret computer models critically is a vitally important ability for us to foster. Contributing to it is a major goal of this book — and an important motivation for the field of software studies as a whole. But, as Bogost and Turkle both agree, having an understanding of how software operates is the essential foundation for developing this ability. This, I believe, is the major unacknowledged contribution of systems like SimCity.

Authors such as James Paul Gee have made arguments that computer games are sources of important learning. But Gee is focused on learning in general, and so his examples (in books such as What Video Games have to Teach Us About Learning and Literacy, 2004) range to include “shooter games” (10–11) with models as simple as the earlier-discussed quest flag and dialogue tree logics. These may make contributions to learning, considered widely, but they do little to help develop understanding of complex software systems. Lev Manovich may seem to make a similarly-broad point when he argues that “As the player proceeds through the game, she gradually discovers . . . its algorithm” (2001, 222). But his examples (patterns of enemy AI behavior, Wright’s design approach) come closer to the issue: The example of SimCity is important to our culture, I believe, precisely because it demonstrates a way of helping millions of people understand the operations of complex software models. Only from this point does it become possible to develop a deeper, critical engagement with software operations.

From the perspective of digital media authors, SimCity also has another importance. When creating Raid on Bungeling Bay, Wright realized that interacting with his terrain editor was more interesting than interacting with its outputs. In a way this is quite similar to the insight offered by the Tale-Spin effect: let the audience experience the most interesting parts of the system. SimCity illustrates a specific strategy for accomplishing this. Years later, when this strategy was applied to characters, Wright created one of the most successful games of all time.

Notes

2Though it is technically true that models used by the CBO are open to scrutiny, as are most of those used by non-government researchers such as Forrester, Starr tells us that “to most participants in policy debates as well as the public at large, the models are opaque. Only a few can penetrate the black box and understand what is inside.” SimCity, on the other hand, shows that carefully-designed playability can be an effective tool for helping much larger groups understand how simulations operate For more on the CBO’s simulation work, see “Overview of the Congressional Budget Office Long-Term (CBOLT) Policy Simulation Model” (O’Harra, Sabelhaus, and Simpson,2004).