Can Math Predict the Oscars?

The Academy Awards ceremony - aka 'The Oscars' - are tonight. Herewith, a debate about whether mathematical models can predict the winners - and one person's prediction (he's been right at least 75% of the time since 2012, got 85% last year - and he's a Harvard-trained mathematician now doing statistics for the Los Angeles Dodgers baseball team).

Not that that makes him any more of an expert than any of the rest of us.JL 

Jody Avirgan reports in The 538 blog and Adam Zausmer predicts in The Hollywood Reporter:

The Oscars present the perfect forecasting challenge. It’s a small ecosystem (only some 6,000 members of the Academy get to vote), there are all sorts of inscrutable variables (Hollywood relationships, behind-the-scenes politicking, “buzz”) and there’s a limited sample of previous data to work with.

The 538 blog: The Oscars present the perfect forecasting challenge. It’s a small ecosystem (only some 6,000 members of the Academy get to vote), there are all sorts of inscrutable variables (Hollywood relationships, behind-the-scenes politicking, “buzz”) and there’s a limited sample of previous data to work with.
FiveThirtyEight’s Walt Hickey has been immersed — and we mean immersed — in the world of Oscars forecasting over the past few weeks. He discusses what he’s learned about this particular challenge, and offers his picks for this year’s winners. Plus, he guest-hosts the first ever What’s The Point mini-debate between Oscar forecaster Ben Zauzmer and The Wrap’s Jeff Sneider over whether you can predict the Oscars using math.
Stream or download the full episode above, or subscribe using your favorite podcast app. Below, a transcript of some highlights from the conversation.

Data and diversity

Jody Avirgan: This year in particular, the conversation around the Oscars is hitting upon the insularity of it, but in a lot of bigger ways. So obviously with regards to gender, and especially this year, race, we’re seeing that the Oscars are really white. Does data provide the promise of being able to crack that problem or understand the roots of that problem?
Walt Hickey: Data has actually been a little bit ahead of the curve on this. There’s a lot of research being done, in particular with the Media, Diversity & Social Change Initiative at the University of Southern California Annenberg. They actually categorize every single speaking character in the top 100 movies each year. They’ve been doing that for about six years now. And they’ve found pretty decisively that there are insufficient women on screen and insufficient minorities on screen, and that’s a problem. And it’s a pipeline problem. If you’re not getting the “Guy in Restaurant” role, then you’re not going to be able to get the role as a supporting character, you’re not going to be able to get the acting role, and then you’re not going to be able to get nominated for lead actor.

A WTP Debate: Opening statements

Ben Zauzmer: You can predict the Oscars with math. We’ve got all of this data from previous years — Guild Awards, other award shows, critic scores, which categories you’re nominated in — and we can use all of these things to figure out how important each of them are to predict each Oscar category. Then plug in this year’s data, and we’ve got ourselves a formula. It’s not perfect. Math is probably never going to go perfect, but, frankly, neither are qualitative predictors. I think at least math can definitely add something to the conversation each year.
Jeff Sneider: Obviously, there’s a lot of credibility behind math and I think that math is an important element to how I formulate my own predictions, but it’s not everything. I wrote Walt in the middle of the night — one of those ill-advised emails you immediately regret — and I was like, “There’s no way that math can predict the Oscars — there’s so many different elements and variables — sometimes it’s just instinctual. You can feel the current almost in the town.” That’s where I’m coming from. I’ll put my gut up against any formula, so to speak.

The winners that a mathematical model predicts in the top eight categories favor Leonardo DiCaprio and Brie Larson, but Sylvester Stallone may have to sweat it out.

The Hollywood Reporter

The Revenant is twice as likely as Spotlight to take the big prize of best picture at Sunday night’s 88th annual Oscars. Leonardo DiCaprio and Brie Larson are, statistically speaking, assured the top acting awards. But while many are predicting a Sylvester Stallone victory in the best supporting actor category, that race is still looking like a close call.
That, at least, is how my mathematical predictions are looking for the big night.
For the fifth consecutive year, here are my Oscar predictions using only math. In these standings, I don’t factor in my own opinions of the movies, I don’t consider what my gut tells me will win, and I don’t allow myself to be biased by industry buzz. My statistical model uses only cold, hard facts, such as who won the BAFTA in each category and how many nominations a film received. And I then use historical Oscar data to determine the relative importance of each predictor.
I understand that the Oscars are a field of human endeavor subject to bias, personal opinion and random chance. But, all of the predictors are subject to those exact same whims, which is why my model has correctly predicted at least 75 percent of categories every year I’ve done this. Last year, it went 18 for 21.
If you use this to fill out your Oscar pool, keep in mind that these are percentages, not guarantees. Upsets do happen. Even if I list a nominee with a 90 percent chance, that means one time out of 10, a nominee with that Oscar résumé will fail to emerge with the little golden man. But what this exercise does offer is a guide to which nominees are favored going into Sunday night.
In this article, I’ll take a look at the eight top categories. The picks in remaining categories will be appear tomorrow and Friday, both here at The Hollywood Reporter and at @BensOscarMath.
Best Picture

Talk about momentum. The Revenant failed to get a screenplay nomination. It also failed to get a Screen Actors Guild nomination. It lost the Producers Guild of America’s award. But then everything changed. A Directors Guild win. A BAFTA win. Add in the fact that it received the most nominations (12) and won the Golden Globe for best drama, and we have a relatively weak best picture favorite.
But SAG champion Spotlight and PGA victor The Big Short, both of which also earned Writers Guild of America award wins, are still very much in the running. No film has ever won the best picture Oscar without either a screenplay nomination or a WGA nomination since the WGA’s 1949 founding, and The Revenant is trying to be the first.
Best Director

Alejandro G. Inarritu would have been the favorite even without his film’s 12 nominations. He would have been the favorite even without his BAFTA win. This is one of the few categories where one predictor is so good that it dominates all the others, and that predictor is the DGA. There have only been four times in history when a DGA champion who was also nominated for an Oscar failed to win it, and none since 2002, when Roman Polanski (The Pianist) lost to Rob Marshall (Chicago). Don’t look for this to be the fifth time.
Best Actor

The time has come. Leonardo DiCaprio’s year has finally arrived, and he’s not leaving anything to chance: Golden Globe win? Check. SAG win? Check. BAFTA win? Check. And, just for good measure, many other groups also have awarded his work in The Revenant. Don’t overthink this one. DiCaprio wins in a landslide.
Best Actress

Room’s Brie Larson has similarly dominated the other lead acting category. She claimed the Golden Globe, SAG, BAFTA and many other honors, and has emerged as the clear favorite. She actually has a higher percentage than DiCaprio — even if it less than 1 percent higher — mainly because her competition is slightly weaker. Neither Oscar nominees Jennifer Lawrence nor Charlotte Rampling received SAG or BAFTA nominations. If you think of each race as a pie with 100 percent of total slices to go around, when Lawrence and Rampling take less of the pie, that’s more pie available for Larson.
Best Supporting Actor

This is the single closest race of the night. Normally, the SAG winner helps one nominee emerge from the pack. This year, the SAG award went to Idris Elba (Beasts of No Nation), who was controversially not nominated by the Academy. In that case, the math essentially says that there is no clear favorite among any of the Oscar nominees. That’s why everyone is so bunched together. Mark Rylance of Bridge of Spies built a last-minute case with his BAFTA victory, but Creed’s Sylvester Stallone picked up Golden Globe and Critics' Choice wins and those just barely give him the edge. Not to turn to a cliché, but this is still anyone’s Oscar.
Best Supporting Actress

The other supporting acting race is almost as close. The Danish Girl’s Alicia Vikander took the SAG and Critics' Choice Awards, making her a not-so-comfortable favorite. Steve Jobs’ Kate Winslet took the Golden Globe and BAFTA, though in both those cases Vikander’s Danish Girl role was categorized as leading, so they weren’t direct competitors in the supporting category. To confuse things even more, Vikander was nominated for supporting actress by both of those latter two bodies, but it was for her role in Ex Machina. My model’s conclusion to this mess is basically that no clear frontrunner has emerged, so everyone has a shot. But the shots belonging to Vikander and Winslet are a bit better than everyone else’s.
Best Original Screenplay

Will Oscar voters give this to Spotlight as a consolation prize if it doesn’t win best picture? The math doesn’t account for such speculation, since it’s impossible to get inside the voters’ heads. But the math does account for the fact that Spotlight won the WGA, the BAFTA, the Critics' Choice Award and many other honors. It was also the only one of these five films nominated for the Golden Globe, though that honor ultimately went to un-nominated Steve Jobs. In total, due to its dominating performance this Oscar season, and a lack of anything going for its competitors, Spotlight is the only choice here.
Best Adapted Screenplay


The top eight categories come in pairs this year. The top two have The Revenant in first. The lead actor categories are blowouts. The supporting actor races are thrilling. And the screenplay competitions are lopsided. Just like Spotlight, The Big Short won screenplay honors from the WGA, the BAFTAs and the Critics' Choice Awards. WGA-ineligible Room has an outside shot, but The Big Short is the clear frontrunner.
Coming tomorrow:  Animated Feature, Documentary, Foreign Film, Original Score, Original Song, Production Design, Cinematography
Ben Zauzmer (@BensOscarMath) uses math to predict and write about the Oscars for The Hollywood Reporter. He recently graduated from Harvard with a degree in applied math, and he now works as a baseball analyst for the Los Angeles Dodgers.