One of the most difficult parts of training troops (be they infantry, sailors or fighter pilots), is accurately showing how the enemy will operate. In the past, ignorance was the norm on both sides, so the disadvantages were roughly equal. And there was quite a shock when you first encountered a foe that fought differently than you did. In peacetime, troops often practice against each other, giving everyone the false impression that the enemy will fight the same way you do. Once a war is under way for a while, training is changed to show how the other guy will come at you. But there's obviously an advantage to training against a realistic opponent in peacetime. The United States air force and navy introduced a "realistic opponent" for their pilots in the 1970s, and the army did the same for its combat troops in the 1980s. But the problem today is that there are so many different types of opponents out there, how do you deal with that. Troops have found that commercial computer games have the answer. When computerized wargames appeared in the early 1980s, they were generally designed for a human to play against an "artificial opponent" (called the "AO"), which was created using programming techniques that used AI (Artificial Intelligence.) It turned out that, at the tactical (man to man) level, the options are not all that great for fighters, thus the "combat AI" didn't take up as much computer resources as was originally thought. This was one reason for the success of the "First Person Shooter" games (like DOOM) in the early 1990s. Tactical combat is relatively easy to model. A good example is fencing. Check out a fencing manual (although you have to add some historically accurate moves that don't make it into those manuals, like grabbing the other guys sword or knocking over stuff to distract your opponent.) Faster and cheaper PCs enable these ancient routines to run in real time with a lot more "actors" on the screen. Back in 1985, one computer modeling project was done (in 90 days) to test the effects of new "smart mines" being developed. These weapons are now in service under several names and configurations, one of them being SADARM. The 1985 combat model used a Symbolics workstation, which had computing power equal to a 486/33 PC (which was about one percent of the computing power that a present day 2-3 GHz PC has.) Thus today you can use these simple routines, and the ability to easily edit them, and realistically put hundreds of accurate "hostiles" out there for the troops to practice on. This use of the inherent simplicity of combat, enormous amounts of cheap computing power and the average recruits vast experience with computer games gives American troops a significant training advantage,