Oscar predictions get the FiveThirtyEight treatment for second straight year

Oscar prediction model on
The Oscar prediction model on hopes to exceed even last year’s forecasting success.

With only a few days until the 88th Oscars (ABC, Sunday, 7 p.m. ET), the wait is almost over. Will Best Actor nominee Leonardo DiCaprio break his “losing streak“? Will Best Director nominee Alejandro González Iñárritu take home the glory for a second year in a row? And, years after beating opponents in the fictional ring, can Sylvester Stallone beat the field in the Best Supporting Actor category?

FiveThirtyEight might have the answers. The site is featuring extensive coverage of the Academy Awards that will include live tweeting the ceremonies, anchored by lead lifestyle writer Walter Hickey. FiveThirtyEight also is forecasting the winners of the top six categories using an election-style Oscar model. Categories include Best Picture, Best Director, Best Actor, Best Actress, Best Supporting Actor and Best Supporting Actress.

Last year, the first year FiveThirtyEight released its picks, it correctly forecast five of the six main categories (missing only on Best Director).

Front Row spoke with Hickey about this year’s predictions, his quest to discover alternative Oscar-predicting methods and more:

Can you briefly go over the methodology and the process of creating your model?
This was informed by our political model. Since we can’t poll the folks who actually vote for this event, we look at other events preceding the Oscars that vote for Academy Award nominees, and try to figure out how much meaning to inscribe to each of those.

We do that by determining how accurate these events have been [at predicting the Oscar winners] over the past 25 years. In the end, we give a score to each nominee based on what awards they’ve won and on how good these awards are at predicting Oscar winners. [In depth methodology available here]

What have been some of the biggest challenges in analyzing this year’s contenders?
This year has been really interesting, especially if you look at categories like Best Supporting Actor. The person who won the single most awards [this season] was actor Idris Elba for “Beasts of No Nation,” but he wasn’t nominated for an Oscar. So we don’t have a lot of data for that particular contest.

You asked readers to submit their own predicting models for the Oscars – how surprised were you on the kinds of models you received?
We are always looking for new ways to reach out to different fields and look at different approaches and sets of data – whether its sports, culture or politics.

So we recruited eight contenders and essentially had a “bake-off”: They submitted their own methodologies on predicting the Oscars and we followed them along the process. What I liked the most was that we got one model from a guy working with the Massey Ratings for college football, [James England]. He is trying to factor in how they do college football rankings and apply that core concept into the Oscars.

As a movie-goer, what was the biggest surprise for you in terms of numbers?
The Best Supporting Actor category. Mark Ruffalo crushed on “Spotlight” and Christian Bale in “The Big Short.” Any other year, those guys would have been way ahead of the competition, but it seems like there’s a big campaign behind Stallone that I can’t account for. I personally don’t think he was the best part of “Creed,” so that was surprising for me.