This is a familiar and successful way of thinking that we are taught in school and even as scientists. The challenge is that some decisions and domains are simp

This is a familiar and successful way of thinking that we are taught in school and even as scientists. The challenge is that some decisions and domains are simply not amenable to statistics and concepts like correlation break down in the face of infinite payoffs (like patients getting to go to a grandchilds wedding) and infinite penalties (like the death of a child). How does one account for a situation in which 10,000 good decisions are for nothing in the face of one mistake?