Oscar Forecast RecapPosted: February 25, 2013
Sunday evening was a fantastic night if you’re a fan of Argo, Seth MacFarlane, James Bond, or movie musicals. But it was an equally great night if you’re a fan of movie math. In my second year predicting the Oscars with only math, my model went 17-for-21, including correct calls in the tense races for Best Picture and Best Director.
Let’s take a look at what went right and what went wrong for statistics in Hollywood. First, we’ll turn our attention to the closest races which Ben’s Oscar Forecast picked correctly.
Best Director: Ang Lee – Life of Pi. This was a very controversial category. Many “experts,” including some people endeavoring to use statistics, picked Steven Spielberg for Lincoln. Mathematically speaking, though, my model found it wasn’t that close. When I crunched the numbers, Spielberg was down at 10%, even behind David O. Russell for Silver Linings Playbook and Benh Zeitlin for Beasts of the Southern Wild. Why? There are a few reasons, but the biggest is that Mr. Spielberg did not receive a BAFTA nomination, which hurts him severely in a model using data from the past 15 years.
Best Picture: Argo. Many correctly called this award, while a few picked Lincoln as the favorite. While it is true that Argo was not nominated for Best Director, that is not nearly as important as the plethora of predictors Argo had going for it. For that reason, my model put Lincoln way down at 9%, with Argo up at 60%.
Best Writing – Adapted Screenplay: Argo. Others had Silver Linings Playbook, but anyone who put enough weight on the Writers’ Guild for this category got this one correct.
Best Animated Feature: Brave. The other popular pick was Wreck-It Ralph. This shows why it’s a dangerous pitfall to put too much weight on the Annies and the Producers Guild. The other factors, the BAFTAs and the VES awards, made Brave was the clear frontrunner.
There were also plenty of categories that most other statistical prognosticators did not attempt. My model correctly predicted Life of Pi for Best Original Score and Best Visual Effects, Argo for Best Film Editing, Anna Karenina for Best Costume Design, and Les Miserables for Best Makeup/Hairstyling and Best Sound Mixing. It also called Skyfall for Best Sound Editing, which split the award with Zero Dark Thirty, the first tie in 18 years.
Almost every predictor, including my model, picked Daniel Day-Lewis in Lincoln for Best Actor, Jennifer Lawrence in Silver Linings Playbook for Best Actress, Anne Hathaway in Les Miserables for Best Supporting Actress, Amour for Best Foreign Film, Searching for Sugar Man for Best Documentary – Feature, and “Skyfall” for Best Original Song.
In 4 of the 21 categories, the mathematical frontrunner which I calculated and the winner did not correspond. Three of them were simply extremely tight races. Christoph Waltz, who won Best Supporting Actor for his role in Django Unchained, was only 9% behind leader Tommy Lee Jones. Lincoln took Best Production Design despite falling 8% below Anna Karenina. Best Cinematography went to Life of Pi, though the math put it 3% behind Skyfall. From a mathematical perspective, these results are all entirely reasonable and expected, since the differences were so slight.
The only category with a larger difference was Best Writing – Original Screenplay, where second-place Django Unchained beat first-place Zero Dark Thirty. The mathematical suggestion partially rested on Zero Dark Thirty’s win at the Writer’s Guild, but Quentin Tarantino was ineligible for a Writer’s Guild award for his Django Unchained script under the WGA’s restrictive rules.
The overall lesson: math is an extremely powerful tool for prediction, working at an 81% clip on this occasion. But absolute perfection is unlikely to be found on a calculator; I’d prefer to look for it in a great movie.