I have been thinking about combat in RPG games lately and wondering if I could come up with a way to model combat that I could use for machine learning. I'd like to be able to make decisions about the best weapon to wield and the best armor to wear …

I've never really liked the Monty Hall problem. The soution is counterintuative and my brain kind of breaks when I try to wrap my head around the conditional probabilities. In situations like this it can be beneficial to look at the problem empirically and estimate the probabilities of the outcomes …