I understand that he’s a statistician by profession. I still believe that there is a great amount of science to be applied to many of the statistics. Take for example, baseball. Every team had a great deal of statistics, and the players with the most RBI, HR’s and hits were most likely to sought by all the richest teams. Billy Beane came around and decided that while the use of those statistics were a decent indicator of a players talents, they were highly flawed in regards to the overall contribution of a team. He built a team built around OBP and was highly competitive with extremely low salaries paid out. (I never watched Moneyball but I heard that as the basic idea).
So while I agree that his statistics may be the best ones out there today, I’m not extremely sold on his use of averaging unscientific polls. He has to un-skew the polls, which would based on the average turnout predicted in polls, and average a demographic of the Likely Voters. I believe his accuracy can’t stay perfect if he keeps his same methods for next election. Again, the Redskins called almost every election correctly without any use of science. Hopefully he uses a little more science and not just numbers next time.