Quants V. Critics: Who Was Better At Predicting The Oscars?
In our small sample, math models won out slightly over expert critics in predicting Oscar wins in our 10 favorite categories.
It looks like you can mathematically predict the Oscars, after all–but you won’t do much better than traditional critics.
Last week, I spotlighted four quantitative Oscar predictions that left expertise and intuition entirely out of the equation. (Since I wrote that post, stats demigod Nate Silver came out with his statistical predictions, too.) I spent this morning comparing all those predictions to the predictions of traditional critics from a few major news outlets, including the Los Angeles Times, Entertainment Weekly, Next Movie and the New York Times (which also hosts Nate Silver’s blog, so I guess the Grey Lady is hedging her bets).
My limited sampling found that the quants did better than the film critics, but only by a little. On average, the models I found were correct 78 percent of the time, while the critics were correct 73 percent of the time. Overall, such numbers pale in comparison to the accuracy of Nate Silver’s political predictions for the 2012 election season. It seems the Academy of Motion Picture Arts and Sciences is a bit more inscrutable than the American public.
You can find the whole table of comparisons here. I apologize; it’s pretty ugly, but I hope it does the job. For a stunningly thorough comparison of the quants versus the critics, see the right sidebar at Farsite Forecast, the Oscar-predicting website of a consulting company that did the best among the models I checked.