Algorithms Predicted Every Winner Of An Awards Show
An algorithm recently predicted nearly every awards show winner. It was scarily accurate, easily nailing category after category
This article is more than 2 years old
Ben Zauzmer strikes again. The Harvard statistics whiz who correctly predicted 20 out of 21 Oscar winners in 2018 has given this year’s Tony Awards his verdict and managed to get 18 out of 25 categories right using the strength of his analytical model alone. The young mathematician and author of Oscarmetrics: The Math Behind the Biggest Night in Hollywood uses a complex algorithm to deduce the results of awards shows every year. Though he originally received recognition for his accurate Academy Award predictions, Zauzmer also dabbles in sports, musicals, and every manner of award-giving bodies. And he’s got a whopping 75% success rate. Check out his Twitter thread for the Tonys right here.
Unlike most other data scientists, Zauzner’s propensity for predicting awards shows began with an earnest love for cinema. He told MLB in February he was an applied mathematics student at Harvard when he first caught the itch. He was interested in using a forecast model with which he could divine the results of almost every awards show imaginable — except the technology didn’t exist. Empowered by a need to inform audiences on their favorite movies and investors on where to focus their purchases, Ben Zauzner developed a revolutionary new model. He cobbled together data from an exact aggregate of past awards show results, betting markets, critics’ reviews, and which categories a film or actor is nominated in. And he has rarely been wrong since.
Ben Zauzner isn’t the only mathematician using science to predict awards shows ahead of time. There’s also Farsite, GoldDerby, Awards Watch, Galaxy Sciences, and almost every journalist writing for a major news outlet. But Zauzner is way ahead of the game when it comes to awards show prediction. While Farsite, for instance, could only figure through about 5 or 6 categories, Zauzner often has the whole show completely worked out. There’s no beating his razor accuracy. His prediction model even has betting websites beat. Bookmakers have nothing on Zauzner’s statistics, and it doesn’t even cost The Hollywood Reporter as much to collect information from him. Bettors pay big money to see their preferences take the cake, and yet their accuracy is dependent on a number of unreliable factors: popularity, social media trends, and money. Bookmakers certainly carry out exact aggregates of all bets made, and yet those aggregates are of the public’s general preferences. Zauzner wouldn’t rely on such an erratic source.
Given his scarily perfect track record, it is a wonder Ben Zauzner isn’t more involved in betting websites. His predictions on awards shows decimate bettors’ guesses by a mile and they are decidedly more scientific. His model is a collection of many relevant factors, including past precedence and a film or actor’s performance in other awards shows. The industry would benefit highly from his forecast model being revised to suit, say, who the next James Bond might be, or whether a fledgling cinematic universe is doomed to fail (or conversely, fated to last) depending on box office earnings, critical reviews, streaming numbers, and opening weekend performance.
His model has so much potential in other fields — not just awards shows! — that it could practically pave the road for an entirely new world of applied math and statistics that services only entertainment trends. This would help many businesses – studios, producers, game publishers, and other major conglomerates — develop the insight to make better corporate decisions that benefit both the company and consumer market as a whole. Amazon’s recent forays into space tourism, drone manufacturing, and game distribution, for instance, are gambles waiting to either implode or result in something phenomenal. Ben Zauzner’s model could predict the company trajectory and prevent messes from ruining investments. The same could be said of political maneuverings; does a particular candidate genuinely have a chance? Will a certain decision impact a president’s approval rating at all. The opportunities are endless.
Ben Zauzner has just accurately predicted most awards shows since the early 2010s. He doesn’t have to stop at just the Oscars and Tonys. The sky is the limit for this guy, and that’s no hyperbole. All math and science.
