PVEBBER makes a lot of interesting observations and comments which deserve some responses, so I thought I'd take a stab at some of them. He writes:

1) Linear, or at least analytically tractable relationships between cause and effect - the whole "metrics mania" that even Congress is getting into. Part of the snake-oil being sold as "network-centric assement theory" applies only to 'complicated' not truly 'complex' systems, which differ fundamentally by the very fact that cause and effect are discernable in complicated systems, but are not in complex systems - there are two many feedback driven interactions to know where the output needle will swing when you "twiddle the dials."
I see a lot of this--the idea being that we can use wargaming to forecast outcomes in very complex, messy situations and use them as a way to "test" or "validate" commander courses of action. Lots of money is being poured into such efforts, and I am not convinced that such are based in solid theoretical foundations, as PVEBBER writes of above.

Usually when I run across such simulations being used for these purposes, I immediate want to dive into the algorithms to see if it's indeed a complicated versus complex system that is being simulated. If it's the former, then it's easy to attack the simulation design. If it's the latter and somebody is trying to prove that it can validate courses of action, then I start asking how many "runs" were done and with what differences in variables...a good complex system will usually have wildly different outcomes even with the same variables. Typically when somebody is trying to peddle this kind of system, they've only done one run for the buyer who is golly gee so impressed with this latest technology.