Relevant Commentary From The Old Place:
To explain the bolded part some more, wargames in the military aren't just giant laser tag matches. Most of what happens at a high-level is adjudicated by computers and by actual people (OCs or observer-controllers).
You can think of this in terms of a roleplaying game, like D&D. The computers handle the "rules" of the game, but those rules don't cover every scenario. When someone wants to do something unconventional, it is the role of the OCs to adjudicate that. Conceivably, those OCs should be subject matter experts and able to make informed decisions as to practicality of the ideas.
In the case of the Millennium Challenge, the real failing was with the OCs. Once Van Riper started to do things that worked outside the bounds of the computer model, they needed to intervene to adjudicate. They either fail to intervene or they lacked the understanding to do so in a reasonable way. The two most commonly cited failings were:
Van Riper famously used "motorcycle couriers" to get around the destruction of his C2 network. The problem was that this was handled as simple handwaving, with him saying that he would use motorcycle messengers to handle all the message traffic. Despite supposedly doing so, he continued to relay messages as if he doing so with a normal communications network. Essentially, his motorcycle messengers were treated as being just as efficient as an electronic communications network. Clearly, that wouldn't be possible, hence the folks joking about "light speed motorcycles".
Van Riper also supposedly destroyed the US fleet with swarm tactics. Attempting to do so isn't unreasonable, and evaluating these tactics is exactly the kind of thing wargames are meant to do. The trouble was that Van Riper again got too creative for the computer model and the OCs failed to impose reasonable restrictions on Van Riper. AShM aren't small and you need a robust ship to launch them; you can't just strap an Exocet onto a Boston Whaler and expect it to actually remain seaworthy. The problem is that Van Riper did exactly that. He mounted AShM onto boats that couldn't have reasonably fitted them, and he further took advantage of the limited operational boundaries of the wargame to essentially have his swarm fleet close to effectively point blank range without having to transit any intervening distance.
What Van Riper did was less of outsmarting the US military and more a case of outsmarting the boundaries of the game. This would be like me playing Monopoly with Warren Buffet, exploiting some weird flaw in the rules to win, and then going on a press tour afterwards saying that I had proven that Buffet was a financial idiot and that I had invented some new financial investment paradigm.
relevant edit:
I once participated in a wargame that simulated an ABCT attempting to repel a Russian armored division. We were given a limited time, the scope of our resources, and a narrow AO in which the exercise was supposedly taking place (our flanks were secured by other armored brigades).
We set up a defense-in-depth across our entire AO, and then started the exercise. As soon as we did, the OPFOR literally just drove outside the map, through the conceptual AO of our supporting brigades, emerged behind all our forces and then obliterated us. It was stupid.
The way that exercise was handled reminded me of MC. Basically, the OPFOR just ignored the conceptual boundaries of the wargame and the OCs failed to step in to stop them. Of course the enemy isn't bound by our AO definitions, but on the other hand, if they had actually attacked along that axis they would have ended up fighting two brigades instead of one. Ignoring the bounds of an exercise doesn't mean you've outwitted the exercise; it just means you agreed to a set of constraints and then chose to ignore it (i.e. you cheated).