Predicting the Oscars is hard, if not impossible. You have a voting body that consists of thousands of people who are notoriously discreet and can in no way be polled. The best picture prize is given out via a voting algorithm that involves ranking and consensus building, so even if you could poll the voters, your predictions would be based only on a snapshot of people’s ordered preferences. And with just one event annually, a changing electorate and frequent rule changes, any predictive model you’d build would be obsolete in a few years’ time.
So how can we take a stab at divining the Oscars with these limitations? In short, by seeing who wins other awards — like the Golden Globes or the Screen Actors Guild Awards — that take place ahead of the Oscars and then trying to pinpoint how much those awards matter.
We’re tracking which nominees are favored in the race for eight major Academy Awards. Read now »
If you’re a distinguished professional in the film industry who’s been invited to become a member of the Academy of Motion Picture Arts and Sciences, which runs the Oscars, there’s a good chance you’re in another club too. Maybe it’s the Directors Guild of America, or the British Academy of Film and Television Arts, or the American Cinema Editors. Those groups send ballots to their members just like the academy does; some prizes they give out have proved to be good predictors of how Oscar voters cast their ballots. Although it’d be lovely to have specific numbers for every vote behind these pre-Oscar awards, knowing the nominees and winners for each is a start.
Our model also takes into account the nominees and winners for some awards given out by critics and other outsider groups, for example, the New York Film Critics Circle. Even though members of the press don’t vote on the Academy Awards, there’s a good chance that movies they like are in the mix for an Oscar. These awards are a chance to get free air time in front of voters: Even though the Hollywood Foreign Press Association — which runs the Golden Globes — is a fundamentally ridiculous outfit, a few minutes looking good on national television accepting a prize for acting can make you a contender, at the very least.
So how do we come up with the scores that each Oscar nominee has in the categories we track?
First, we assemble the historical data, by pulling the winners of more than a dozen film awards going back as far as possible, up to 25 years. Some of the awards are from the entertainment industry guilds.1 We hypothesize that these are representative of the academy’s thinking. Some of the awards are given by non-industry groups.2 We hypothesize that these may be representative of the academy’s thinking.
Next, the math:
- For a given award, we look at how many times that award — say, the Golden Globe for best supporting actor — went to whoever won the analogous prize at the Oscars — in this case, the Academy Award for best supporting actor. In that example, 17 of the past 25 Globes winners won the Oscar; that’s 68 percent of the time.
- We square that fraction (weak scores get weaker and strong scores stay pretty strong) and multiply by 100 — e.g., 0.68 squared is 0.46, which turns into a score of 46.
- If it’s an insider award — i.e., an award voted on by people who may also be members of the Academy Award electorate — we then double the score — again, based on the hypothesis that awards from industry insiders are more likely to be in line with the academy’s decisions.
- If an Oscar contender was nominated for the award and loses, the person or movie gets a fifth of those points. The winner gets all of them.
So what’s new this year?
- We’ve implemented the algorithm for the documentary and animated feature Oscar categories.
- We’re including the top Annie Award in the model for animated feature.
- We’re multiplying scores by 100 because it looks way nicer.
So what should you be paying attention to if you’re looking for hints at what Oscar night could bring? If we group award shows based on how many points they could spread around, the biggest nights are the BAFTAs (on Feb. 12, with 513 points across all categories) and the Screen Actors Guild Awards (Jan. 29, 452 points). We’ve already had the biggest night from the outsider groups: the Critics’ Choice Awards (Dec. 11, 290 points). And while they’re more narrowly focused, the Directors Guild awards (Feb. 4, 262 points) and Producers Guild Awards (Jan. 28, 208 points) are some of the heaviest hitters in some categories.