What can go wrong with war game simulations?

Toward a History-Based Doctrine for Wargaming by LT COL MATTHEW CAFFREY JR., USAFR, published in April 2000, is an exceptionally interesting study of how war games and simulations have developed, and more to the point, how they have been used (or misused) in actual events.
Ive tried to isolate some of the problems that have arisen – not to criticise war games as such, but because these problems can presumably also arise with any other sort of simulation. The military have just put more time and effort into finding them.

Lt Col Caffreys article records perhaps eight types of problem.


1. Not addressing all the issues. (What Id call getting the wrong boundaries to your subset of reality.) Before the First World War, the much vaunted German simulations of an invasion of France and Belgium did not address the diplomatic and political consequences of these actions. The invasion of Belgium brought the British Empire into the war, with the short term consequence that the German advance was halted at Mons, by a British army which the Germans had not anticipated (and of which they were unaware until they met it), and longer term consequences that contributed to ultimate German defeat.

2. Having the wrong objectives. In 1934, in US Army field manoeuvres, air power was not allowed to attack enemy forces before, during, or after amphibious landings, but was only used in close support after trench lines had formed. When it was pointed out that this was unrealistic, the Army’s response was that their learning objective was to practice trench warfare. If air power were used too soon, the trench lines might not form. Clearly the simulation should not be skewed to make something happen for exogenous reasons: otherwise the rest of it is rendered largely useless. (ie if you want to test trench warfare, test trench warfare, but be clear that this is (all) that you are doing.)

3. too much complexity. As the rules of war games became more and more complex, embracing not only the exact capabilities of weapons systems on each side, but also logistics, supplies, maintenance, infrastructure, etc., the games became so complex and time-consuming that they grew cumbersome and unpopular. Feedback was slow and directing staff overloaded. This is possibly one area where computers can make a real change: complexity is not an issue for them.


4. The “command problem”: when a senior player overrides the rules. The best example of this is the Japanese wargame prior to the Battle of Midway. During the game, the Japanese players representing the US forces sank two Japanese aircraft carriers. Rear Adm Ukagi Matome, commander of the Japanese carrier force, uniltaterally reversed this decision. With the carriers restored to the game, the Japanese side went on to victory. Some weeks later, the real battle of Midway took place, and US forces sank the same two carriers, plus two more. Alas for the Admiral, this time they could not be replaced on the board. Caffrey points out that the Admiral was right in one sense: in the game, the carriers were sunk by B17s, which sank no Japanese ships in the real battle. Because of his rank he could overrule the game umpires decisions. (This does not always lead to military disaster – Hitler did the same thing with timid German wargamers, and the early successes of the Blitzkrieg were the result. After all, senior leaders usually have shown strategic skills and judgement: thats why they are there. So the natural tendency to defer to senior commanders is not entirely foolish.) The problem is that one person senior enough to break the rules negates much of the value of the game: which is to explore in a structured way, not to restate existing assumptions.

5. Using simulations to prove things: Caffrey cites several examples where simulations and war games were used to prove or demonstrate that particular projects or military doctrines were correct. Mostly this was not deliberate falsification; it was due to competition betwen services, over-optimistic setting of the rules, or attempts to gain funding for pet projects. However in some of these cases, it does seem that even the flawed work was of value. At least ideas were developed in such a way that, when real war exposed the errors, it was easier to correct them than if no simualtions had taken place at all.

6. Ignoring the unthinkable: Caffrey quotes Admiral Nimitz: “The war with Japan had been [enacted> in the game room here by so many people in so many different ways that nothing that happened during the war was a surprise—absolutely nothing except the Kamikaze . .” Clearly, American or British planners would not naturally anticipate the fanaticism of the Kamikaze. It is very difficult to put oneself entirely into the mindset of an enemy, especially where there are large cultural differences. (Lessons for the war on terrorism here?). A variation of this problem is the tendency to assume that the enemy will do the most intelligent thing that you would do under the circumstances. He may do the wrong thing: will this catch you out?


7. Ignoring the lessons learned. Caffrey cites the war games in Saint Petersburg, in April 1914, when two Russian armies advanced into East Prussia. When they entered an area of lakes that made cooperation between the armies difficult, the players for the German side placed only a token force against one Russian army, and used most of their forces to destroy the other Russian army. When real war broke out, the real Germans did exactly the same thing. The lessons of the wargame had been forgotten or ignored by the Russians, and Russian defeat at Tannenberg followed.

8. Intelligence failures. The Nazi wargame Operation Otto predicted that German forces invading Russia would destory 240 Soviet divisions, leaving only 60 remaining. When the real invasion began, German forces advanced almost exactly as scheduled, reaching the objectives predicted, and destroying 248 Soviet divisions. However, there were 220 more Russian divisions left, not 60, with disastrous results for the Germans. (Also this war game seems to have failed to anticipate the arrival of winter…) However, lack of intelligence information does not render war games useless. Caffrey cites one example – US Naval wargaming in the 1930s – where instructors gradually allowed students to realise that available intelligence about the Japanese was poor. Instead of backing one assessment rather than another, the students developed a strategy of learning by experience before committing to any decisive engagements. In other words, says Caffrey, they were “learning how to learn.”


My argument in listing these faiures is not to say that military wargaming is a waste of time. Lt-Col Caffery produces a lot of evidence to show that it makes forces and strategies far more effective, prevents mistakes, and helps to stimulate sensible planning and preparation.

I just want to identify a list of issues which all simulators ought perhaps to consider. The military have had a lot of experience, and are often very frank about recording and analysing their mistakes as well as their successes. In addition, it is usually all too clear if they got it right or not: they either won the battle, or lost it.

These issues or problems must apply just as much to business or economic simulations, or to world models, models of political behaviour, and so on. In many of these areas, the arts of simulation may be less well developed, and the simulators less open with others.

Leave a Reply

Your email address will not be published. Required fields are marked *