July 29, 2013: The German armed forces recently received the latest version of CAE GESI. This is a computerized wargame that allows commanders and their staffs to train realistically. CAE GESI has been used by the German armed forces since the 1990s and is constantly updated to reflect new military developments as well as new tech available to commercial game developers. The current version is a networked system operating on PCs and able to represent 32,000 “actors” (individual weapons, vehicles, aircraft, small infantry units, and major items of equipment) in real time or compressed (sped up) time. The most recent versions have concentrated on supporting brigade (a unit of three or more battalions and over 3,000-6,000 troops).

The developer and manufacturer of this simulation is CAE (Canadian Aviation Electronics), a Canadian firm founded in 1947, to build aircraft electronics and flight simulators. CAE got larger and larger and began acquiring related companies and by the 1980s was developing other kinds of combat simulators (tanks and, eventually, things like GESI). While the U.S. armed forces was the leader in adapting commercial gaming technology for military use this applied mostly to low-level (individual troops, vehicles, and small units) operations. For the “operational” level games (battalion, brigade, division, and up) the U.S. military developed most of the game components. That has not worked out so well and non-U.S. military simulation/wargame developers simply adopted commercial wargame technology for professional products.

In addition to CAE (who does most of their business in Europe), a French firm also supplies military users with operational level wargames (MASA SWORD) and a civil version for disaster operations. CAE GESI is also available in a disaster relief version. The French publisher got started in 2001, with a commercial wargame (Conflict Zone) and then moved on to professional (as opposed to entertainment) games, models, and simulations.

Both CAE and MASA use a lot of civilian wargame tech, which can be bought for any kind of game or simulation the buyer wants to use for. This is also common for a lot of professional wargames produced by American firms, but not so much for operational level games. The result is that CAE and MASA have succeeded where the U.S. has failed in creating workable wargames for operational level commanders. What is going on here is different approaches to wargame development in the United States and elsewhere. In America developing wargames was always seen as a government function, even though civilian firms were often contracted to do most of the work. In the rest of the world the military never really got into the civilian combat simulation field and when civilian firms began creating products like GESI and SWORD there were many nations eager to just buy what they needed.

In the United States a lot more money has been spent to obtain a lot less. For example, four years ago the U.S. Department of Defense shut down its JWARS (Joint Warfare Simulation) project after a billion dollars and fifteen years of effort had been spent trying to build a joint (air, land, sea) strategic combat simulation. This effort began after the 1991 Gulf War, when it was discovered that there was no accurate simulation systems (wargames to civilians) that senior commanders could use to practice and tinker with large scale conflicts, or those that required, as the 1991 war did, gathering forces from all over the world.

The subsequent JWARS effort eventually collapsed because no one could control the development effort and too many people got a say in what JWARS should try to do. Moreover, each service (army, navy, air force, marines) was to build a simulation of their own that would mesh, where necessary, with those of other services. The Department of Defense would create the software to make all this run together. This ran into problems when it was discovered, as many old hands at Pentagon simulation efforts predicted, that the services would disagree on how certain combat operations should work. For example, the air force had a different idea of how ground combat worked than did the army. Even the army and the marines found they had some misunderstandings about ground combat. This sort of thing was useful, but the lack of firm leadership and development discipline led to the project spinning out of control. Thus, in late 2010, the Department of Defense pulled the plug. Not all was lost, as many of the efforts by individual services to simulate their own operations were useful, for them at least. So the actual loss was a few hundred million dollars (in wasted effort and software and other stuff that ended up being useless) not over a billion.

The Department of Defense still needed some kind of joint simulation and the air force stepped in with its Storm software. This was an operational/strategic level war simulation created to support the planning for air force operations. But Storm could more easily plug into simulations from other services and provide a workable joint war simulation. The navy has enthusiastically bought into using Storm, the army less so. Outside the U.S. the army tends to take the lead in deciding how war will be portrayed in simulations. But the U.S. has, since World War II, had a much larger (in absolute and relative terms) air force and navy than in any other nation. That means the navy and air force leadership are better able to impose their views.

People involved in the 15 year effort to create JWARS did get a lasting lesson about the need for firm leadership and well defined goals for large scale and multi-service (joint) war simulations. Although Storm wasn’t designed as a JWARS replacement, it was the best available substitute. Another lesson learned from JWARS (and not much talked about in public) is that the services are not eager to surrender their ideas about how wars are fought by other services. The army has long had problems with the air force and navy, in peacetime, using all sorts of inaccurate ideas in their training simulations about how the army actually operates. This continues to be a problem and every time there is a war there is a period of adjustment to reality.

The U.S. Department of Defense efforts to create simulations and wargames still run into a lot of other old problems. A major one involves history versus science. This began right after World War II, when the Department of Defense sought to simulate war via scientific principles. This meant collecting vast amounts of data and developing algorithms that would accurately simulate every aspect of combat. The alternative was the historical approach, where less precise algorithms were created, based on less complete data, and refined using historical experience, to produce a "good enough" simulation of combat.

An example of how the two approaches work in practice occurred on August 2nd, 1990, as Iraq was invading Kuwait. The U.S. Department of Defense swung into action. One of the first things the Pentagon officials did was order wargaming likely options for the unfolding situation. All of the Pentagons computerized wargames were too slow off the mark for this job. So later that day, the first wargame used at the Pentagon was a commercial one. With all the billions spent on computerized wargames since 1945, Americas most efficient military operation in this century was initially planned using a manual wargame (Gulf Strike) that could be bought by anyone in most hobby stores for under fifty dollars.

This was not a unique situation. During World War II, the German army regularly wargamed operations in much the same way that Iraq’s Kuwait invasion war gamed out in the Pentagon. The manual wargames used by the Germans were very similar in style to manual commercial wargames developed in the United States during the 1960s and 1970s. The main difference was that the Germans considered their games military secrets and not available to civilians. When the allies invaded France on June 6th, 1944, the Germans were in the middle of a wargame dealing with just such a possibility. As reality had overtaken the games hypothetical premise, the German commander ordered the game to proceed, not as a game but as a command tool. Wargaming had been a common practice in Germany for a century before the Nazis came along. Other nations copied this practice and the British, Russians, and Japanese also wargamed every major operation during World War II. In the United States, only the navy used wargames for such planning.

However, for many years after World War II, traditional wargames got little respect from the professional military wargamers because these older, history based, manual wargames were not considered precise enough for current needs. The development and application of computers, operations analysis (OR), and systems analysis during World War II firmly implanted the idea that simulations of war had to, and could, provide precise and unambiguous answers. It was ignored that history based wargames, despite being relatively imprecise and ambiguous, had usually been accurate enough to be useful. So, as the decades rolled by after 1945, even the slow learners realized that the scientific approach could not match the usefulness, reliability, and timely accuracy of the older style wargames. Finally, in the 1980s, the Department of Defense, without abandoning its science approach to wargaming, allowed the older historical games (both manual and computerized) to have a place at the table.

Several military wargamers summed up the problem very succinctly by pointing out that there is a tradeoff between the accuracy of your wargame and how much time and resources you have available. In wartime commanders have discovered that they can prepare a "good enough" analysis (or wargame) of a situation in a few hours. A few days, a few months, or a few years can create more accurate wargames. But these don't count if they are not available when needed.

Until the 1990s, professional wargaming tended to go after the most accuracy possible and consume years and millions of dollars in the process. Indeed, the process usually overcame the search for a solution, often leading to a lost, and failed, project. That happened most spectacularly in 2010, when the billion dollar effort to create JWARS collapsed. The experience with the $50 Gulf Strike wargame in 1990 had been forgotten, or ignored, by the JWARS backers but not by many other officers and especially non-U.S. commanders. Commercial manual wargames can generally achieve a "good enough solution." Put a wargame on a PC and you get a more accurate, and just as quick, solution. But it took the military a while to catch on to this cheaper solution sitting on the shelves of a local game store.

There are other problems military wargame developers ignore. For example, the best way to simulate irregular warfare is to accept the fact that this involves multiple games. One of these would involve self-protection (defending your troops, both in bases and outside them). Another game would involve collecting intelligence. This would involve raids and searches, establishing informant networks, and success and speed at analyzing the data collected. Another game would deal with local politics and how to persuade the local population to get along with each other, and your forces.

Obviously, all these games are interrelated. But historical wargame design always deals with that (aspects of the situation that are there but not manipulated in the game). Finally, there is the task of validating your wargame. Historically, this is done by setting up historical situations, playing them out using historical decisions, and measuring how close the outcome is to the historical one. Once you achieve the historical result, you engage in "free play" and see how the game holds up. You keep tweaking (making minor changes in numerical values and procedures) until you have a wargame you would bet your life on. This has been done successfully for over a century, so don't be afraid. An irregular warfare game has many historical situations to use for validation. If you can predict the past, you can predict the future.

Commercial games like GESI and SWORD are based more on the historical model than the bean counting approach. This is particularly necessary for the disaster relief versions, which need to be constantly updated with new experiences and must be very practical to be successful. Wars are infrequent compared to disaster relief situations, and in both cases accuracy rather than perfection is what is needed.