As some of you may have seen, I have a feature in this week’s New York Magazine in which I use a database of the last thirty years of Oscar history to predict the recipient’s of this Sunday’s Academy Awards.
This is fundamentally not all that difficult to do, since the winners of other awards such as the BAFTAs and the Golden Globes are quite strongly predictive of success in the Oscars. When the other major awards are split between two or more contenders, we can look at other sorts of tiebreakers: The Academy really does not take kindly to comedies or action films, for instance. And there is such a thing as “sympathy points”: if an actress or actor has been nominated for an award several times without winning (such as Kate Winslet for Best Actress), she becomes more likely to collect the hardware. (From a technical standpoint, the challenge is really just to build a reasonably reliable model without overfitting).
Spoilers follow below the fold. One of these, by the way, I’m almost certain that I’m going to get wrong, although I have a pretty good excuse. For the supporting detail, please see the original copy.
Best Supporting Actor: Heath Ledger, The Dark Knight (86% chance of victory)
Best Supporting Actress: Taraji P. Henson, The Curious Case of Benjamin Button (51% chance of victory)
Best Actor: Mickey Rourke, The Wrestler (71% chance of victory)
Best Actress: Kate Winslet, The Reader (68% chance of victory)
Best Director: Danny Boyle, Slumdog Millionaire (99.7% chance of victory)
Best Picture: Slumdog Millionaire (99.0% chance of victory)