Modelling systems in simulations

We’re gonna need a bigger seminar room

As I’ve noted before, simulations are really good for integrating students’ understanding of large, complex events: the active approach required forces students to apply their knowledge and adapt it to a wide range of elements and dynamics.

However, one area were this breaks down is in the area of systems, in the sense of larger-scale interactions. The two most obvious examples are the international system in toto and the various stages of a legislative procedure from start to finish, both things we find rather interesting in politics/IR.

The difficulty comes in the sheer volume of stuff that has to be modelled by the simulation: lots of actors and of variables, usually without the requisite time, space or number of players to make it work.

So what to do?

There are a number of ways to address this, each with merits and shortcomings.

The first way is just to pitch in and try to model it all: certainly that’s what sims like Statecraft try to do (it’s also my approach here). You take the chance that you cover all the key variables, you plonk the players in some initial condition and then let them go. As I’ve noted with my game, the breadth of possibilities is both a blessing and a curse: players can do a huge range of things, but this also often inhibits them as they get cramped by indecision. Moreover, it’s possible to get too hung up on the range of variables: what goes in, what stays out, how do they interact with each other? In short, you can do it, but it’s easy to end up with either a very, very complicated game or one that produces some perverse incentives.

The second way is to break down the big thing into some smaller things. For an international system, you might create a world of actors, who then meet in some specific locations (a UN meeting, a security challenge, regional cooperation, etc.) that can be simulated more easily. For comparativists, you might get actors contesting an election, then forming a government, then falling into (and out of) civil war. The appreciation of the larger picture then emerges from the shadow of the past games (and the shadow of the future ones, if you tell students what’s coming), and the reflection about how that shapes action. The tricky part of this is creating a situation where those different elements hang together in a sensible way: this might well be a case of having to try to build up the sims over a number of iterations.

The third way is a more networked approach. I wrote about this a couple of years ago, suggesting a set of legislative documents that different institutions could pick up and advance through a full legislative process. This would let everyone focus on a specific element, while also gaining a sense of a bigger picture/process. The detailing could be very impressive, but the downside is the one that should be evident: namely, it’s a very complicated thing to create and maintain (which is the major reason I never got around to trying to make it happen).

The final way is go go down the stylised route and produce games that capture the spirit, without all the detail (like this). This is certainly the simplest to run and manage, but one necessarily loses the specifics. It’s also worth saying that it’s often quite tricky to capture the key mechanism, not least because it’s often a normative choice about what you think is important to capture: One can say that the European Parliament is about how committees dominate procedure, rather than about how transnational groups form, depending on what you want to tell students.

As always, there’s no right answer in all of this, but as you approach the bigger things, it’s really worth spending some time thinking about how you tackle them.