Originally Posted by smurgerburger View Post
I was referring to your assertions about AP sports betting rather than your position on AP generally.
Apparently it is impossible to gather a large amount of sports data, withhold some of the most recent data, build a model with the rest of it, test the model on the withheld recent data, if successful (beating the vig), test the model on some upcoming games without betting and then,if that is successful, bet a small amount of money on some upcoming games. Refining the previous steps as necessary.