One challenge in wargaming, and especially political-military (POL-MIL) games, is how to best model the behavior of unpredictable, even apparently irrational, foes. Is the mercurial behavior of North Korea’s “Supreme Leader” Kim Jong-un, Boko Haram in northern Nigeria, or Joseph Kony of the Lord’s Resistance Army truly irrational, or is it a simply the product of a very different set of interests and objectives sustained by a very different world-view? To what extent do seemingly erratic aspects of their strategic behavior derive instead from factors we don’t understand well, such as internal politics or decision-making process? It has been well established since the POL-MIL wargaming of the 1950s and 1960s that actions that one actor believes to be rational signals of intent or deterrence are often entirely misunderstood by their intended recipient, in large part because they are deeply shaped by internal decision making processes that opponents fail to appreciate or understand.[i] How do we incorporate this into wargames when, almost by definition, we do not fully understand what is going on?
This ongoing methodological challenge has acquired greater significance in the context of recent political changes in the United States. Leaving issues of political partisanship aside, it is clear that many US allies find the new Administration of President Donald Trump to be unpredictable—to the point of posing a potential threat to their countries’ core national interests.[ii] Harsh campaign rhetoric, a seemingly chaotic foreign policy making process, mixed signals, and the propensity of the President to express his thoughts in provocative tweets have left many allied policymakers scrambling to develop contingency plans in case long-established US positions or commitments are no longer credible.[iii] Indeed, even those members of the US State Department charged with reassuring nervous US partners express frustration that they are often unclear as to what American policy is on any given day.[iv] The result has been an increasing interest in some allied countries in gaming the US as a potentially unreliable military-diplomatic ally, or even—on some non-military issues, like trade or climate change—as a political adversary.
...even those members of the US State Department charged with reassuring nervous US partners express frustration that they are often unclear as to what American policy is on any given day.
To start with, it is important to note that while the behavior of unpredictable adversaries or unreliable allies is difficult to forecast with precision, statements and actions usually fall within a range that can be broadly anticipated. True “Black Swan” events are, by their very nature, very rare indeed.[v] We thus tend to be considering a bell curve of plausible behaviors, bounded at the extremes by low probability/high impact actions.
Whether adversary or ally, the mechanisms with which behavioral unpredictability might be introduced into a wargame or crisis simulation vary depending on the basic type of game, be it educational or analytical, as well as the particular objectives of the game. It also varies depending on the importance of the apparently erratic actor concerned in shaping game outcomes, and the extent of knowledge as to what factors drive decision-making and account for unpredictability.
In many hobby wargames, a degree of unpredictability is built into gameplay via the roll of dice or other common random event mechanisms (for example, drawing from a deck of cards that each describe a possible occurrence). Mechanics of randomness in such games are intended to capture the causal effects of the unknown, as well as elements of Clausewitzian fog and friction, rather than to suggest that there is no actual linkage between cause and effect.[vi] The game mechanic is also intended to complicate planning and decision making without unduly influencing outcomes: after all, who wants to play a game in which the results are decided less by player decisions than they are by the draw of an event card or the roll of a die?
In professional wargames, however, truly random events might subvert the educational purposes or research design of a game. Instead, if the actor concerned is of secondary importance, the game designers can script apparently erratic behavior into the scenario from the outset, or have such injects on standby to use if their introduction would serve the game objectives. In an educational or training game, for example, scripting surprising actions by non-player actors is a valuable way of encouraging agile and adaptive thinking among participants.
Scripted actions may thus seem surprising or “unpredictable” to the players but not to the WHITE cell, the team officiating the event. One drawback to such an approach arises when players feel that they are being railroaded along a preconceived scenario, and that nothing is truly unpredictable because the “unpredictable” behavior of an ally or adversary has been preprogrammed from the outset. In this case, some apparent element of genuine unpredictability (whether real or not) may help to sustain player engagement in the scenario, and help thereby promote the role-playing and dramatic immersion in the game that is so important to effective wargaming.[vii] It is also important that the unpredictable not be completely implausible, or the credibility of the game will be compromised among participants.
Scripted or randomly generated events are only useful, however, when the behavior of an unpredictable adversary or unreliable ally is secondary to wargame outcomes. The more important their actions are in shaping outcomes, the more sophistication is needed in modeling their behavior. One mechanism is to use a modified random event system, whereby players within the BLUE or RED team are presented with a series of potential policy positions, decisions, or declarations by their “supreme leader” and must expend limited internal political or bureaucratic resources to tip the final position in the way they think would be most productive. Another approach would be to try to represent the internal tug-of-war that shapes an actor’s final policy positions. Interagency games often do this in an effort to reproduce the sort of semi-cooperative dynamics that shape so many collective and coalition efforts.[viii] However, what is suggested here is that the struggle be represented in political and ideological terms too, and not simply institutional ones.
The result is what can be described as two- (or even three- or more) level games.[ix] In such games, players are not only interacting with adversaries and allies at an international level, but are also seeking advantage in domestic political and bureaucratic struggles too. Policy positions are thus the outcome not only of geopolitics, but also domestic interests, institutional process, and ideological struggles within a government or a non-state actor. One or both these sorts of approaches are likely to figure in representation of US and North Korean policy making in a South/East Asia crisis game being planned for the Connections UK professional wargaming conference in September of this year.[x] The question of modeling messy and complex policy-making dynamics is also likely to come up at the Connections US wargaming conference, which will held in Quantico, VA in August.[xi]
...in games examining Iranian behavior some players assume a sort of cartoonish Islamist revolutionary fanaticism drives behavior...
One challenge in designing such a game-within-a-game is to know that you have the model right, when the real-world process itself is confused or opaque. To some extent game designers will necessarily be designing-for-effect (that is, creating a game system that produces the right kind of outcomes) rather than establishing causal linkages in a way that can be fully validated and verified. A second challenge is to assure that, given the game dynamics and procedures that have been established, most outcomes fall within the bell curve of plausibility described earlier, while retaining at least some possibility of the highly improbable occurring. As with all serious games, play testing is essential. Third, there is the danger of either mirror imaging or caricaturing the worldview and ideological behavior of others. In the first case, players import their own views when playing others, failing to fully understand important differences in perspective. In the second case, players treat the views of the other as far more exotic and unusual than it actually is. For example, in games examining Iranian behavior some players assume a sort of cartoonish Islamist revolutionary fanaticism drives behavior, while actual policy making in Tehran is “rational” but also shaped by a complex and sometimes dysfunctional decision making process, political rivalries, domestic politics, and often poorly-coordinated policy implementation.
In educational wargames, modeling apparently erratic behavior through these sorts of game devices serves to underscore the fundamental importance of messy, even dysfunctional, decision making processes in strategic behavior—an important antidote to narrowly “realist” and rationalist images of international relations. In analytical games, the process whereby in-game actions are determined can help to illuminate the conditions under which such erratic behaviors might emerge, how they might be influenced, and what early warning indicators might be relevant.
When it comes to the modeling of seemingly “unpredictable” or “erratic” actors, it is important that the game model not be treated as sacrosanct—after all, it is only an approximation of something that is poorly understood. Instead, the discussions players have, their own experiences within the process, and the perspectives they develop are likely to be the most valuable output.
Rex Brynen is Professor of Political Science at McGill University, and senior editor of the PAXsims blog on conflict simulation and serious games. He has previously worked as a member of the Policy Staff of the Canadian Department of Foreign Affairs, as an intelligence analyst, and as a consultant for various other governments, the World Bank and United Nations agencies. Follow Rex on Twitter @RexBrynen.
Have a response or an idea for your own article? Follow the logo below, and you too can contribute to The Bridge:
Enjoy what you just read? Please help spread the word to new readers by sharing it on social media.
[i] Thomas C. Schelling, “Red vs. Blue,” in Pat Harrigan and Matthew Kirshenbaum, ed., Zones of Control: Perspectives on Wargaming (Cambridge MA: The MIT Press, 2016).
[ii] Although far from scientific, one survey of ninety NATO and FVEY colleagues working in the diplomatic, defence, and intelligence fields found that a striking 73% rated the new Administration as a larger threat to Western economic and security interests than Russia, China, North Korea, or Iran.
[iii] Dan De Luce and John Hudson, “U.S. Allies Are Learning that Trump’s America Is Not the ‘Indispensable Nation’,” Foreign Policy, 27 February 2017. Accessed at: http://foreignpolicy.com/2017/02/27/u-s-allies-are-learning-that-trumps-america-is-not-the-indispensable-nation
[iv] Julia Ioffe, “The State of Trump's State Department,” The Atlantic, 1 March 2017. Accessed at https://www.theatlantic.com/international/archive/2017/03/state-department-trump/517965/
[v] Nassim Nicholas Taleb, The Black Swan: The Impact of the Highly Improbable 2nd ed. New York: Random House, 2010.
[vi] For a discussion, see Philip Sabin, Simulating War: Studying Conflict Through Simulation Games (London: Continuum Books, 2012).
[vii] Peter P. Perla and ED McGrady, “ Why Wargaming Works,” Naval War College Review, 64, 3 (September 2011). Accessed at https://www.usnwc.edu/getattachment/413220cd-a209-4447-bac2-af6cdaeabd4f/Why-Wargaming-Works
[viii] Rex Brynen, “Gaming the semi-cooperative,” PAXsims blog, 2 February 2016. Accessed at: https://paxsims.wordpress.com/2016/02/02/gaming-the-semi-cooperative/
[ix] Robert Putnam, “Diplomacy and Domestic Politics: The Logic of Two-Level Games,” International Organization 42, 3 (Summer 1988). Accessed at https://www.jstor.org/stable/2706785?seq=1 - page_scan_tab_contents