Taking Games Seriously

Pete Adey’s recent post on this blog about security-oriented video games has made me think more broadly about the role of “games” in how we think about geopolitics and security issues. While video games are all the rage now (in popular culture, but also in critical IR / critical geopolitics studies), there’s a much longer history of security studies people using “games” to think their way through situations, from the formal modelling techniques of game theory through the looser scenario modelling exercises practiced by international relations professionals and military planners.

In this post, I want to turn specifically to the latter kind of game: the simulation scenario. Geopolitical analysts are trained early on in the simulation scenario. In the US, for instance, the final component of the Foreign Service exam involves a real-time decision-making scenario game in which several applicants are placed in a room, told that they are the staff at an embassy, and given a crisis scenario to which they are collectively instructed to devise a response. The candidates are observed through a one-way mirror where they are graded for their ability to comprehend the scenario, devise creative solutions, listen flexibly to others, and persuade others to support their proposed plan of action.

I’ve recently become involved in an online international conflict scenario exercise managed by the security consulting firm Wikistrat. Wikistrat’s business model involves putting together a stable of “analysts”, giving them access to an online, proprietary discussion-board website, and asking them to design and model scenarios around a specific topic. Although Wikistrat makes its money from performing modelling exercises for paying clients, they populate their list of potential consultants by staging mock scenarios in which individuals with an interest in the topic can participate as volunteers. These mock scenarios (actually, mock scenarios of mock scenarios, to be precise) give Wikistrat a chance to test-drive its software while also providing potential consultants an opportunity to advertise their services to the Wikistrat managers who are looking on. You earn points for each comment that you make on the online discussion board, and I imagine that the Wikistrat managers make other, more qualitative notes as well, much as instructors do when they mark students who are required to participate in online discussion boards (or even in-class discussions).

To encourage wide-ranging thinking, Wikistrat lets analysts design their own scenarios within the modelling exercise. The current exercise is called How the Arctic Was Won, itself an interesting title since it assumes that there is a competition, that the current situation is not good, and that eventually there will be a winner (and presumably a loser). Just as indigenous people in the Americas 500 years ago were likely surprised to find out that their land needed to be “discovered,” I imagine that indigenous people of the Arctic will be surprised to find out that their land (and sea) needs to be “won.” Proposed scenarios have ranged from one in which the Arctic Council claims territory to one in which China sends an aircraft carrier group to the Arctic. So far as I can tell, only two of us (myself and Arctic law scholar Timo Koivurova) are from outside the security studies world and are unaccustomed to these sorts of “games.” We keep asking questions like, “How do you get from the situation today when the Arctic Council doesn’t have any security responsibilities to your proposed scenario where the Arctic Council is claiming territory?” and “Why has China chosen military action as the best way to achieve its objectives in the Arctic?”  The responses to our protests have been revealing, as they’ve ranged from indicators of ignorance regarding the details (some have implied that their expertise is in geopolitical scenario modelling with whatever data is given to them, and that lack of specific knowledge of the Arctic shouldn’t be an impediment to good Arctic modelling) to reminders that the whole point of a modelling exercise is to begin with acceptance of core assumptions about rational choice and the behaviour of individuals and states.

As an experiment, I played my own game with the gamers. When one suggested a scenario wherein Russian geoengineers developed global warming technologies and applied them to the Northern Sea Route, I responded by suggesting a scenario wherein the G-77 countered with a technology (developed out of a collaborative institute in Singapore) to refreeze the water, starting a new kind of cold/hot war of climate modification (I confess: I was inspired by Ice-9 in Kurt Vonnegut’s Cat’s Cradle). I earned points for my insightful suggestions. Yikes!

In case it’s not obvious, I find these kinds of exercises in what ultimately are applied microeconomic rational choice models deeply frustrating. If one establishes a framework wherein ANYTHING is possible, then the only constant which can be used to guide scenarios are pre-social things: namely “man’s” “naturally” aggressive instincts and the “tyranny” of landscape (the unchanging arena in which these aggressive instincts are applied). This is Robert Kaplan run amok. I also question whether such exercises could truly be useful to potential clients who want advice on how to respond to possible real-world situations. So, I am highly suspicious of the “game” as an exercise for gaining perspectives on reality. At some level, every game is dependent on reifying abstractions about choice and opportunity (after all, that’s what goes into the algorithms that limit/enable individuals’ responses). As a result, such attempts at appreciating the world’s complexity and open-endedness tend to end up just masking that complexity.

At the same time, we do need to think “outside the box,” and games can help us do this. In fact – somewhat ironically – Timo and I are preparing to engage in our own simulation exercise of sorts, designing a framework for a model law of ice, even though we acknowledge that the world currently is far from a political situation where such a law might be applied to the world’s frozen regions. The difference, I hope, is that we will never fully bracket the social conditions that make it possible (or, perhaps, impossible) to get from Point A to Point B. When you add a truly social element to scenario modelling the game might become a bit less fun….but it becomes much more useful for understanding reality and for imagining aspirational, but potentially realizable, futures.

Phil S.

5 thoughts on “Taking Games Seriously

  1. Phil,

    Thank you for the feedback. As one of the staff overseeing and designing these crowdsourced exercises at Wikistrat, I always appreciate feedback and criticism, which helps us constantly improve what we do.

    In terms of the methodology used, indeed crowdsourced scenario planning are about exploring the full spectrum of options and alternatives. We believe that through that wide exploration, bringing together experts from different disciplines and backgrounds, in transparent and collaborative-competitive fashion, you can uncover the outliers and black swans typically missed by traditional consultancies.

    Yes, you typically get some “out there” scenarios in the process, but I’d also point out that while we encourage analysts to think outside the box, we request that all scenarios are within a zone of plausibility. Through the discussion and peer criticism we are able to spot these and remove or flag them. The “how the Arctic was won” simulation is still in progress, and scenarios are continuously removed/adapted, in part under the guidance of our simulation supervisors.

    It is an iterative process, where at its end you get a range of plausible scenarios. Take the example of the Chinese Carrier scenario you mentioned- this scenario since been updated by more than 10 analysts, including Dr. Timo, to reflect on the criticisms and refine it. It now looks into China Trying to secure and expand its economic interests in the Arctic, and will likely keep evolve in the coming days. These ongoing efforts are constructed into a final report at the end of the exercise, synthesizing the main insights and weighing the various scenarios based on their merits. I would be interested on your thoughts on the report following the end of the simulation.

    In terms of client simulations, they are typically more comprehensive in terms of the analytic team composition, analytic modules applied (impacts analysis, policy options, resilience analysis etc), and the simulation’s length and complexity. These are also projects where analysts are paid for their time, in addition to an incentivizing prizes scheme, awarded to top contributors (the latter is also applied in internal simulations like the Arctic one, to further enhance the competitive element in the simulation).

    I’d welcome your thoughts on the simulation moving forward, in particular as when it is concluded and when the final report is released. Looking forward to see you on the wiki,

    Elad Schaffer
    Director, analytic community

  2. Dear Elad

    First, I truly do appreciate your response, I suspect that we may never reach complete agreement on several of the matters that underlie this discussion, but I know that the dialogue is giving me much to think about, and I hope that it will prove similarly useful on your end.

    Please understand that the heart of my comments are less about how Wikistrat runs simulation games and more about what I see as an underlying problem with simulation games in general. The point of my post was to raise these issues in the context of an ongoing discussion on the RHULGeopolitics blog about the relationship between international relations and video games.

    To be sure, there are certain aspects of the Wikistrat format that I would recommend be considered for revising. I’m not sure whether Wiki-type editing is better for facilitating discussion (especially when consensus is not absolutely necessary) than a more traditional discussion-board. And I also think that the discussion might be more fruitful if the leaders posited a scenario (or several scenarios) rather than letting the participants come up with their own, as many of these have turned out to be either poorly thought-out, overly fanciful, or repetitive of other scenarios that are already in play. I’ve seen your staff try to cope with these problems through repeated entreaties to participants to edit the wiki instead of commenting, to more directly address the framework parameters, to merge comments from one scenario into another scenario, etc. But all of this cross-talk while dozens of scenarios are in play keeps things confusing for all but the most committed game-players, and it allows a lot of ‘chaff’ to hang out there in the foreground before it gets separated from the ‘wheat’. That said, I DO see the logic for you organizing things as you do and how it may create a space for the ‘black swans’ to emerge. And I also can see how, over time, the dead-end scenarios simply fade away and get dropped from the report (or get merged into other scenarios). I suppose that organizing things the way you have requires more vigilant attention from both staff and analysts while the game is in play. But on the other hand it makes thing easier for the staff beforehand (you don’t need to put much effort into conceiving of scenarios) and afterwards (the final product of the scenario should already be close to a final report).

    While I’m discussing the specifics of the software, I do have one very small practical suggestion: On the HTAWW homepage, where you list each scenario (and present links for commenting on or editing that scenario), it would be useful if a ‘date last edited or commented on’ notification appeared. That way, casual analysts like me could easily see which scenarios are being actively engaged by the more committed analysts and which have effectively been abandoned. This would hasten the process by which some scenarios rise above the others, which I assume you want to occur over the course of the exercise.

    My bigger concern revolves around the whole project of scenario-modelling as a means for thinking through international relations and security questions. These concerns speak to much broader issues that reflect differences in disciplinary outlook, scholarly cultures, etc. And this was really the main point of my previous blogpost. Interestingly, I’m not sure if the problems that I identify are more severe in a highly unconstrained game system like yours (because, without any guiding limits on what is possible, the players tend to fall back on the least common denominator of isolated agents making primal rational-choice calculations) or in a more constrained system like a video game (where these calculations are embedded in algorithms). And to be sure, I get equally frustrated by excesses in the other direction (e.g. lawyers who have been trained to limit their thinking according to existing rules, without stopping to acknowledge that those rules are themselves of social origin and potentially could be changed if they were not reproduced through practice). So, these are big, open questions, but they speak more to ongoing debates in both how to promote practical problem solving and how to conceive of the arena of international politics than they do to the specifics of the Wikistrat business model, the Wikistrat software platform, or the Wikistrat crowdsourcing concept.

    And finally, please accept my apologies for my relatively low involvement with the simulation over the past week. Some of that reflects the frustrations noted above, but it also just reflects other time demands. I hope to make another round or two of contributions before the simulation concludes on the 15th.

    With kind regards
    Phil Steinberg

  3. Phil,

    I definitely welcome the discussion and feedback, thank you for that. In fact, we took your suggestion for adding a “last update” date and added it on the main simulation page. One of the benefits of an agile start-up is its ability to improve and adapt quickly, and we sure do try.

    In terms of the wiki format, we find that merely having a discussion board is less helpful when trying to have a proper scenario written “in the end of the day”. The discussion is good (hence we allow comments on all pages of the wiki), but should serve to and feed into the actual scenario. The wiki format allows us to have interactive pages that can be added collaboratively (hence iteratively improve and capitalize on the “crowd”), and not limit us to chatter only. It has its disadvantages, but for the purpose of content creation (rather than just discussion), you do need actual pages to be finalized. And you are right, it takes training, education and supervision to get the community members to that point, and to have people accept the notion that “their” original text is adapted. But that’s exactly the point of what we do.

    I do understand your more general concerns over the use of games in international relations. I won’t tire this blog’s readership with what I see as their merits, but am happy to further discuss this in private. I invite you to look at previous simulations on the wiki, and can give you a demo of one of our client simulations, just for the sake of seeing how a far more complex simulation than the Arctic one looks like in terms of structure, methodology and output.

    Thank you for the discussion and suggestions,


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s