Considering that my computer (which isn't exactly crappy but not super powerful either) can barely even manage huge maps (on the vanilla patch my game was crashing almost consistently each turn on only large maps), it wouldn't surprise me if their test systems really did have trouble with these sorts of test scenarios. I suspect they are frustrated by the fact that they pretty much have to leave a test system on overnight just to generate one hugemap game and even then it mightn't be one that creates a particular problem, which will tie into my response to the next reply..
This is true. It's still making use of system resources beyond their own testing facilities, which is clever/practical. The reasons for doing so we can only speculate on, as you said. My speculation is that it's more to do with it being difficult to generate these situations (huge amount of time, not to mention energy costs of running computers to do that sort of thing, and so resources taken up from doing other testing that might be necessary). I think they're trying to hone in on the performance/compatibility issues rather than the gameplay balance type stuff, but it's possible they're going for both of course.
Whatever the case, I don't criticise them for taking this step. Nothing wrong with asking for help when you need it. I just find this whole situation a bit amusing, is all.