One flaw in the book that I think is major is that he uses teams to calculate his player % probabilities. There are too many possible combinations, and incoming and exiting players (and injuries) only complicate the issue. Basically the sample sizes are to small for each team.
