Sure, I am probably exaggerating a bit the difficulty in doing this, but it is a fact in my mind that a simulation model could produce a schedule that is substantially complete, that with a few tweaks by the humans at the top could become a final product.
So could WMATA with the Metro and that's in a system with very few variables (train breakdowns, weather, EMS stops) and we see how well that works. It's not that I don't agree with you, because I do, it's just that when people make comments like "well, with a couple of R simulations and a couple of tweaks this could be done in two weeks tops" gets my eyes rolling. I would tend to think that the MLB schedule would have more complex variables and characteristics similar to setting up a major conference and less issues like scheduling cashiers or, to a certain extent, trains.
