Hmm... lets take a simpler example.
Take 5 horses, All their up-to-date data on each leg of each horse,
each horse & rider's health, after w / L, factor the track , % chance of injury, weather etc. etc.
(all public info)
then pick 1 horse to win & THEN run a Sim.
How is that any different to actually betting on that horse race?
Both events, should invariably fall within the same variance of results, no?