Paul Vebber, a gameplay instructor in the navy, says that in the past decade the government has started using strategy board games much more often. They do not help predict outcomes. For that, the Pentagon has forecasting software, which it feeds with data on thousands of variables such as weather and weaponry, supply lines, training and morale. The software is pretty accurate for "tight, sterile" battles, such as those involving tanks in deserts, says an intelligence official. Board games are useful in a different way. They foster the critical but creative thinking needed to win (or avoid) a complex battle or campaign, he says.
The article goes on to explain that board games are advantageous over computer-based games for what is essentially a simulation:
...you can constantly tweak the rules to take account of new insights, says Timothy Wilkie of the National Defence University in Washington, DC. With computer games, this is much harder. Board games can also illuminate the most complex conflicts.
That's a strawman argument. I'm arguing that simulating the horrors of war aren't all that helpful to a future general than learning how to deal with cunning enemies and shifting conditions, not about overconfidence issues.
Wargames are simply one method of training available. There are in no way perfect, but they are also one of the few methods available. Field experience aside, there are no obviously superior ways to practice such skills. It certainly makes sense to include a multitude of learning methods, but in the end, it's better to have some poor experience than none at all.
So no, you won't understand warfare from playing wargames. But you sure as hell will understand it better than someone whose only military experience is Call of Duty. Skills are not binary states.
but offer you the illusion you are "prepared for it"This is a very interesting (read:insubstantial) claim. Do you have any evidence that mock training leads to significant overconfidence? Do you have any evidence that such overconfidence leads to more errors? Do these errors out-weight the errors committed due to lack of training? And most importantly, how are you sure that this cannot be alleviated through proper training techniques?