Nate Silver has been just about spot-on at predicting elections and he has an impressive track record in sports…but as Daniel D'Addario notes, Silver's unlikely to ever be able to do that with the Academy Awards. I think there's a simpler explanation than what Mr. D'Addario presents. Silver's successful projections all evolve from taking polls and past data (voting records for elections, raw scoring for sports events) and crunching and analyzing it. But when he turns to the Oscars, there are no polls and the data is largely devoid of details. He can look at what won Best Picture every year but he doesn't know who voted for a film or by what margin it won. I don't think he even knows how many people voted in total. We know that Daniel Day-Lewis won for Best Actor this year but we don't know if he got 40% of the vote or 99%. Maybe he only beat out the second-place finisher by one vote, which would make his victory statistically insignificant when compared to the loss of that runner-up…and of course, we don't know who that runner-up was.
I"ll pay rapt attention to what Silver has to say about who'll win political offices. But his guess at the Oscars next year will be no better than anyone else's.