Conflict simulation, peacebuilding, and development

Tag Archives: Stephen Downes-Martin

Connection UK 2015: Day 1 AAR


The Connections UK 2015 interdisciplinary wargaming conference started today at King’s College London, with two of your PAXsims correspondents in attendance—Devin Ellis and myself. This is the largest Connections UK ever, with up to 130 registrants (and, I think, the second largest Connections conference ever).

We started off with lectures on “Wargaming 101.” Tom Moaut (Defence Academy of the UK) provided a general overview of basic gaming approaches, while Jim Wallman (Past Perspectives) discussed how to design a wargame to meet particular requirements. In the latter presentation Jim stressed the importance of determining game objectives and purposes at the outset, noting that the client may not always be clear exactly what they want. Next, he stressed, you also need to establish game constraints and boundaries: participants (numbers, skills, enthusiasm), time, space, level, game resolution, equipment needs, and so forth. Having done this, one can consider initial elements of structure: scope (what does the game explore), structure, time/scale, and how open or closed the game is (that is, whether information is public, or private with “fog of war” represented). Next are game mechanisms. He suggested that this was a somewhat easier step than those prior. A key challenge here is balancing complexity/detail/granularity with simplicity and design elegance. In terms of playtesting, he identified three stages: the “unbaked” session in which one brainstorms initial ideas’ “half-baked” when you have some of the initial ideas translated into game mechanics; and finally playtesting the “baked,” near-final game design.


Jim Wallman’s obligatory design diagram.

Jim correctly stressed that games needed to be assessed against their design aim, and care should be taken that this is not lost sight amid in the enthusiasm of design and play. He warned too that “defensiveness is the enemy of good wargame design”—that one has to be prepared to discard ideas, approaches, and mechanisms.

After coffee, we returned to hear Stephen-Downes-Martin (Naval War College) examined “How NOT to analyse wargames.” Stephen emphasized the importance of using professional analysts, and warned against the influence of senior officials who lack analytic expertise but who do have the power to press their views. Analysts need to be partners with the game designer, thinking from the outset about how they will extract the necessary data from the players and their interaction. Immediate hot-washes, conducted amid the continuing buzz of a recently-completed game, may be inadequate to collect impressions and feedback. He noted that the decisions made in the game are typically less important than the reasons behind those decisions. He stressed the analytic need to treat the White Cell/adjudicators as participants, and understand the rationale for their decisions too.

Stephen too had an obligatory wargaming diagram.

Stephen too had an obligatory wargaming diagram.

Stephen also highlighted the challenge of having the right players in the game. I’ll admit this is an increasing concern of mine, since I’m of the view that game outcomes are heavily shaped by the profile of participants (domain knowledge, risk aversion, interpersonal skills, etc.).

Following on from this there was an excellent panel of analysts from the UK Defence Science and Technology Laboratory (Dstl) on using data (and models) in wargames. Mark Pickering looked at trials and experimentation as a data source for wargames. Dan Ledwick explored systems and performance modeling, looking at lethality and effects. Stevie Ho addressed the use of historical analysis to generate wargame data. He noted that the Falklands War highlighted how different combat experience was from earlier field trial weapons and performance data. Finally, James King (who had also introduced the session) suggested ways of checking wargame data, and underscored the importance of doing so.

The presentation panel from Dstl. Being Dstl, they had many diagrams.

The presentation panel from Dstl. Being Dstl, they had many diagrams.

After lunch we all engaged in a large participatory wargame, New Dover Patrol. This revolved around a vicious separatist insurgency against the rightful government of Silvania by Kippist religious extremists. Faced terrorist gangs seizing parts of the southern city of New Dover, the government had been forced to call upon the United Nations and the powerful country of Freedonia to assist. (Of course, my perspective in all of this may have been distorted somewhat by playing the role of the President of Silivania.)

Jim Wallman presents the game.

Jim Wallman and Tom Mouat present the game.

My government was anxious to get as many Freedonian troops on the ground as quickly as possible, both to combat the extremist menace and to assure their continued commitment. To this end, Freedonian marines seized the port district while our own battered forces performed gallantly and liberated the airport from Kippist terror gangs. This allowed the rapid follow-on of additional forces.

As evidence of their treachery, Kippist extremists tried to assassinate me. Although bloody, I was unbowed, and called upon the country to redouble its efforts. At the same time I held out a hand of reconciliation to moderate rebels who might wish to abandon violence and seek a political settlement. Sadly our efforts were rebuffed by the fanatics.

The enemy was steadily pushed back, but not without heavy collateral damage that began to eat away at Freedonian political support for intervention. This was compounded by the sometimes less-than-cautious activities of the Freedonian air contingent, as well as a second amphibious landing to the west that captured the area around the New Dover water treatment plant—but at the cost of damaging the facility and risking an outbreak of disease. We pressed for the UN to address the humanitarian emergency, and as the game ended we had also called for a local ceasefire in the area around the water treatment plant so that we could effect repairs.

Following the game we then had an analysis session in which we discussed both how the campaign had been fought as well as the strengths and weaknesses of the game design. I thought the game went very well indeed, despite the difficulty of having scores of players all in one tiered lecture theatre.

For the penultimate session of the first day we broke into three groups: Tom Moaut discussed “wargaming effects;’ Jim Wallman explored “developing insights from wargaming;” and Graham Longley-Brown (LBS Consultancy) led a session on “fully engaging the player.” I took part in the latter. Graham stressed the many ways in which truly engaging play supported games-based learning. He did an especially good job of suggesting how we ought to frame the gaming experience so that students remain in both the flow of game play and the “bubble” (or the “magic circle”) of narrative setting. I absolutely agree, and in my Brynania simulation I spent a great deal of effort immersing students in a fictional conflict in a way that generates enthusiasm and emotional commitment to role and interaction. However, I raised the concern that engagement ought not be allowed to substitute for clarity about learning objectives—after all, it is possible to be enthusiastic about learning the wrong lessons. This, I think, was a problem with the Jane McGonigal/World Bank EVOKE social entrepreneurship game. It is, in a somewhat different way, a problem also explored by Anders Frank, who has written about military cadets entering “gamer mode” wherein they are so motivated to win that they exploit game mechanics in ways that undermine realism.

Graham's impressive diagram.

Graham’s impressive diagram.

Finally, we had a hot-wash discussion of how the day had gone.

Both here and at the US Connections conference these first day lecture/course/introduction sessions face a couple of challenges. The first is how to pitch them: although they are intended to aid relatively new wargamers develop their skills and knowledge, a great many of the people in the room are actually very experienced gamers. That may skew the discussion. Second there is the risk that we all tend to discuss the approaches we habitually use, which may mean that some techniques receive more emphasis than others. Nevertheless I thought it was all very well done.

Tomorrow the main session program starts in earnest, with discussion of wargaming developments in the UK and around the world, as well as a games fair (including a demonstration of AFTERSHOCK: A humanitarian Crisis Game).

simulations miscellany, 30 April 2013


Some recent items on serious games and simulations that may be of interest to PAXsims readers:


A revised version of Stephen Downes-Martin’s presentation on “Adjudication: The Diabolus in Machina of War Gaming” (presented at the 2011 Connections conference, and also featured here in PAXsims) has now been published in the Summer 2013 edition of the Naval War College Review.


The University of Minnesota will be holding a humanitarian crisis simulation and field exercise from 31 May to 2 June 2013. This seeks to train participants to:

  • Understand common good practice, minimum standards, and how to improve the effectiveness and accountability of humanitarian program implementation.
  • Achieve humanitarian-based outcomes by using resources efficiently and effectively.
  • Develop collaborative skills, coordinating people and organizations at times of heightened complexity and risk.
  • Operate safely and securely in a pressured and changing environment.
  • Develop personal management and leadership skills.

You’ll find further information and registration details here.


The Public International Law & Policy Group will be holding a training session on “negotiating and drafting provisions related to natural resources in a peace agreement” for legal and policy practitioners on May 3 in Washington DC. The day-long session will include a negotiation simulation. Prior sessions in the series have included ceasefires, power sharing arrangements, transitional justice, and security sector reform. For more information, visit the PILPG website.

Syria, intervention, and the limits of wargaming

Today’s PAXsims post is something that I’ve been meaning to write about in a general way for a while, but is triggered this week by the confluence of several events. The first was a question raised by Michael Peck on the milgames email list, in which he asked how policy planners might usefully wargame the current civil war in Syria and the political and military complexities of possible external intervention. The second was a presentation by Stephen Downes-Martin at the recent Connections conference which focused, in part, on the ethics of wargame design, coupled with a follow-up point that he made in a comment on the Wargaming Connection blog about knowing the limits of our craft. Third, the Saban Center at Brookings has just released a report on a recent policy game they conducted on Syria, entitled “Unravelling the Syria Mess: A Crisis Simulation of Spillover from the Syrian Civil War.” Finally, I’ve been spending much of the week working on Syria-related research.

My answer to Michael’s question about wargaming Syria was fairly straightforward: while I felt that gaming could offer insight into the military challenges of intervention in Syria, I didn’t think it had much to offer in exploring the possible first, second, and third order political effects of intervention in such a complex and dynamic environment—especially given how many “known and unknown unknowns” lurk in the situation there. On the contrary, if I wanted to illuminate the Syria question I would prefer to do it through a workshop or well-run BOGSAT  (“bunch of guys/gals sitting around a table”) discussion, with a team of participants and moderators who knew how to move the discussion along in interesting and thought-provoking ways.

The advantages of this are two-fold, I think. First, a wargame usually (although not always) embeds a single model of a situation, and unless there are opportunities for multiple plays (a rarity in policy games) there is little  scope to explore a variety of possible relationships between key elements of the scenario. A workshop or BOGSAT discussion, on the other hand, allows you to unpack critical assumptions, debate them, consider alternatives, and talk through the policy implications of all that. This is especially important in a case where there is genuine disagreement about the causal processes and relationships at work. Such a discussion can also be very agile, allowing you to rapidly explore new directions when interesting ideas are put forward. That can be much more difficult to do in a wargame, especially the heavily-scripted three-move seminar games of the sort that often predominate in policy settings.

The report of the Syria wargame undertaken at Brookings, unfortunately, rather illustrates my point. The report itself is rather badly done: it isn’t at all clear how large the teams were, how they were formed, how much experience and expertise they had, what range of policy options they considered (or were allowed to considered), or how the white cell and game adjudication operated. The write-up itself is also extraordinarily vague in recounting who actually did what, when, and with what effects. Even the initial scenario is poorly described, as are any injects that might have been introduced during the game. Only three actors were represented by active players in the crisis game: the US, Turkey, and Saudi Arabia. The Syrian government wasn’t represented, nor were the many Syrian opposition actors. With no Syrian actors represented, it isn’t clear that any creative countermoves were taken by Damascus to deter or offset the creeping  intervention that took place in the game, nor that the locals could act to manipulate foreign engagement to their own local advantage. Many of the games findings seem to be self-evident, and hardly needed multiple players to give up a day to identify. The game report notes, for example, that Turkey is important in what happens in Syria—a fact that ought to be evident to anyone who has read the news, or who can even read a map.

Even more serious than this, however, is the extent to which many of the “findings” of the game seem to be as much presuppositions of the game model as they are actual findings. For example, the report notes:

Nevertheless, despite Turkey’s significance, American power still went a long way. For instance, the Saudi team evinced interest in the battles for the arms‐ supply routes through Lebanon which the scenario depicted as escalating—and in the deteriorating situation in Lebanon more generally. But the Saudi team again found itself relatively unable to explore options there without support from the U.S. team. The Saudi team also considered working with Jordan as an alternative (or supplement) to Turkey, but the other teams (principally the U.S. team) showed little interest in pursuing the feasibility of that option.

The paragraph, and the game, presumes that there are “battles for the arms supply routes in Lebanon.” It isn’t clear what exactly that means, nor does it sound especially realistic to me. The report also suggests the game was configured in such a way that the Saudis couldn’t influence events in Lebanon without US assistance—which I also think is fundamentally wrong.

To take another example:

…the Saudi team found that Saudi Arabia had only a modest impact on events in Syria itself, and made little headway with either the American or Turkish teams until the Americans and Turks had decided on their own—and for reasons having nothing to do with Saudi efforts—to intervene in Syria. At that point, Saudi/Arab help became extremely useful, but even then it was not decisive: the American and Turkish teams had made up their minds to do so based on their own interests, and would have intervened (and felt they could have intervened) with or without Arab support, although the Arab support was certainly welcome.

Potentially, Saudi and Gulf influence on events in Syria is (I would argue) much greater than the report, and the crisis game model, seem to suggest—especially regarding financial and material support to the Syrian opposition.

Overall, one gets the sense that the game was rigged to tilt the process towards certain policy conclusions, either because of the policy preferences of the game sponsors and designers or because of the particular worldview that it was built on.

In his comments at Wargaming Connection, Stephen warns of the dangers of “trying to sell war gaming as a solution when other solutions might be better.”  He’s right, of course. It is often the case that a wargame or policy game is not the best way of exploring an issue; indeed, I would argue that it is very often not the best way of exploring complex political issues.

On the other hand, if (say) you’re a think tank in Washington and you want to (say) influence policy on Syria, running a game and putting “crisis simulation” in the title of your report is one way of making it seem somehow more weighty and special. It is, after all a crisis—you know, important and urgent—and a simulation—which, of course, must mean that it bears some relationship to reality. Which means it’s got to be correct, right?


Downes-Martin: Adjudication: the diabolus in machina of wargaming

In our after action review of the recent Connections 2011 professional wargaming conference, we mentioned the excellent presentation by Stephen Downes-Martin (Naval War College) on “Adjudication: the diabolus in machina of wargaming.” In it, he highlighted the challenges of adjudicating wargaming on novel topics that, by their very nature, involve a large number of political, social, or technological unknowns:

Traditional adjudication methods break down when wargaming operational or strategic problems, especially in novel situations for which we have little experience; for example information war in a regional nuclear conflict. The primary causes of this breakdown are the myth that player decisions during a wargame are useful and the failure to recognize the role of adjudicators as dominant players in the game.

Newtonian physics and the statistics of small unit actions provide rules for determining the possible outcomes of interacting player decisions when wargaming tactical level attrition warfare. The adjudicators either “roll dice” to pick one of those possible outcomes randomly as the one that actually occurred, or decide themselves which one occurred to force the players into a situation that best addresses the sponsor’s objectives for the game. However, we do not have the equivalent adjudication rules for wargaming novel operational and strategic problems. In these cases the adjudicators (who are no better informed than the players about how “the world works”) first have to decide the possible outcomes of interacting player decisions and then decide which one occurred. However, psychology research demonstrates that people cannot predict the decisions they would make under different information circumstances, and so decisions made during a game by players are unreliable predictors of decisions that would be made in the real world situation the game is attempting to reproduce.

When wargaming novel situations using traditional approaches the adjudicators not only decide how the world works but also decide what information is given to the players. They become dominant players whose actions and beliefs drive the game –diabolus in machina – resulting in game results which are seductively compelling but ultimately unreliable.

The solution is twofold. First treat the adjudicators as players whose behavior it is critical to analyze. Second, psychology research indicates that human beliefs are robust even in the face of contradictory evidence, and so focus the game design onto the beliefs of the players and how those beliefs drive their decisions, not on the actual decisions….

Stephen has been kind enough to provide his full speaker’s notes for the presentation, which you’ll find here.

The piece even includes an after-the-fact footnote to the Jon Compton/Yuna Wong debate at Connections over the predictive value of formal social science models—although I must say it fails to capture the full colour of Yuna’s descriptive  language (she’s clearly been hanging around Marines a lot). The picture at the top right is, of course, the diabolus in machina, not Stephen.

%d bloggers like this: