Why Microsoft Beat Vegas on Oscar Predictions

The Bing search engine has gotten a lot of flack over the years, but it’s a crystal ball if you use it right.

Unless you’re a genius or a ballot counter at the Academy of Motion Picture Arts and Sciences, Microsoft was better than you at picking Oscar winners. Researchers working with the Seattle monolith’s much-derided Bing search engine correctly predicted 20 of 24 categories on Sunday, including all six “major categories,” a subset including the acting, directing and best picture awards. Last year, David Rothschild, who works in Microsoft’s research lab, went 21 for 24, so this year marked a minor deviation toward the mean. What made his predictions more impressive than last years was the atmospheric turbulence in which the awards took place.

To understand how tricky it is to predict the awards, you have to understand that, after the nomination process, voting takes place from February 15 right up until the ceremonies are almost ready to start. In other words, voters’ behavior is affected by media story lines. For instance, much was made of the lack of black nominees in major categories this year. Was that critique responsible for Richard “White Dude” Linklater’s unexpected loss in the Best Director category? It is statically probable. By using a search engine, Microsoft can track changing attitudes as reflected by changing search terms.

This is not the first time that the Bing researchers have proved prescient. In fact, the search engine’s mechanics have been accurately calling athletic contests for some time. Microsoft’s Oscar success illustrates the unexpected virtue of searching for answers in unlikely places. In order to win next year’s pool, you just have to be the one guy in the room checking Bing. Consider your winnings another charitable donation from Bill Gates.