remember all "testing" is done with variable data on the other side of the field (different pitchers, different parks, different defenses) ... very difficult to specify why things trend one way or another, which is what makes it part research, part crapshoot (kind of like real baseball, honestly, which I think would be the point of the simulation)