How a Harvard maths graduate correctly predicted 20 Oscars winners
A Harvard maths graduate has revealed how a statistical algorithm can be applied to successfully predict the winners of the biggest Oscars categories.
Ben Zauzmer (@BensOscarMath), who works as a sabermetrician - a baseball analyst - for the Los Angeles Dodgers, successfully predicted 20 out of 21 Oscars winners.
He made predictions for 21 of the 24 Oscars categories in total, clarifying there was not enough data available to predict the three short film categories mathematically.
The 25-year-old, who graduated from Harvard with a degree in applied math in 2015, has been predicting Oscars winners for the past seven years, using his complex mathematical model to calculate a nominee’s percentage chance of winning.
His predictions are based on a number of different factors; including critics' scores, betting odds and the film’s performance at other awards shows leading up to the ceremony.
Last chance to fill out those #Oscars pools! Here are the top two most likely winners in each category, according to my mathematical model published in @THR. pic.twitter.com/CuF9oKun6I
— Ben Zauzmer (@BensOscarMath) March 5, 2018
The reason the denominator is 21 and not 24 is that there's not enough data to predict the three short film categories mathematically.
— Ben Zauzmer (@BensOscarMath) March 5, 2018
“In every Oscar category, I gather data on every nominee: which other categories the film is nominated in, how the film did at other awards shows - such as the BAFTAs - and so on,” he told The Telegraph.
“Then, I use statistics to weight these predictors based on how well they've predicted that Oscar category in prior years.
“Finally, I apply those weights to this year's nominees to get the chance each contender wins in every race.”
He was supremely confident in some categories, giving Gary Oldman a 77.9% chance of winning Best Actor for his portrayal of Winston Churchill in Darkest Hour.
The data analyst also convincingly predicted Frances McDormand’s Best Actress win (75%) and Sam Rockwell’s supporting actor victory (74.8%) for Three Billboards Outside Ebbing, Missouri.
The Best Motion Picture of 2017 is @ShapeOfWater!! What an incredible night for math at the #Oscars!!! My model ends the night at 20/21. I'd like to thank those I worked with at @THR and @nytimes this year, and of course, I'd like to thank @TheAcademy. pic.twitter.com/wRppMtDJ4M
— Ben Zauzmer (@BensOscarMath) March 5, 2018
Zauzmer called the Best Picture category “the closest race in at least two decades”, correctly predicting that Guillermo del Toro’s The Shape of Water (36.1%) would beat Three Billboards Outside Ebbing, Missouri (31%) to the award.
“My confidence in these predictions varied greatly by category, and that was reflected in the final percentages,” he explained.
“For Best Animated Feature, Coco had an 85% chance to win, so I felt very sure about that one. For Best Picture, on the other hand, The Shape of Water held only a 5% lead over Three Billboards, so I was nervous right until they opened the envelope.”
A flawless evening of predictions for the maths expert was spoilt with Netflix’s anti-doping exposé Icarus beating his French documentary pick Faces Places, which he had given a 50% chance of winning.
Following last year’s infamous blunder, in which La La Land was mistakenly announced as Best Picture winner, the 90th Academy Awards was plainly a more predictable affair.