OscarLytics: How Second-Place Picks Could Decide the Winner
With "Gravity," "12 Years a Slave" and "American Hustle" locked in a fierce three-way battle, voters' second choices loom very large.
This story first appeared in the Feb. 21 issue of The Hollywood Reporter magazine.
When the final Oscar voting begins Feb. 14, Academy members will be asked to choose their favorite movie of the year. But in this tight contest, the more important choice they make might be the film they mark down as their second-favorite film.
It's a best picture race that has gone back and forth among Gravity (winner of the DGA and a PGA tie), 12 Years a Slave (recipient of a Golden Globe for drama and that PGA tie) and American Hustle (the Globe for comedy and the SAG ensemble prize). So it's very possible that no movie will have the requisite majority to win on the first count. In that case, the movie ranked No. 2 by a voter who quixotically ranked a long shot like Her as his No. 1 choice could prove crucial.
Why? It all has to do with the Academy's preferential voting system, which has been used since 2010 to determine the best picture winner. Each of the 6,028 voters will be asked to rank all nine nominees from one to nine. Initially, the ballots will be sorted into piles according to which movie each voter ranked No. 1. Then, the movie with the smallest pile (that is, the movie with the fewest first-place votes) will be eliminated. The ballots from that smallest pile then are redistributed based on the films they have ranked in second. Then, with eight movies remaining, if no one movie has captured a majority of votes, the PricewaterhouseCoopers accountants repeat the process and the smallest remaining pile is redistributed.
The Academy never releases vote totals, so it's impossible to know exactly how past scenarios have played out. But 2010's The King's Speech was probably a landslide winner, receiving first-place votes from more than 50 percent of voters. In that case, that movie's pile was larger than all of the other piles combined, and no redistributions were necessary. Second-place votes never came into consideration.
But in a non-landslide year -- which this one, with three front- runners, appears to be -- it is unlikely that any movie starts off with more than 50 percent of first-place votes. So redistributions will be necessary to decide the outcome.
Say, hypothetically then, that the three frontrunners all receive a first-place vote on about 30 percent of ballots. The other 10 percent of voters make their first-place choices from among the other six nominees. In that scenario, at least two redistributions -- and probably more -- are necessary. If every one of the nominees earns at least one first-place vote, then at least seven redistributions are necessary.
After the first six redistributions, if the people who ranked those six lower-tier nominees at No. 1 all preferred Gravity to its two main competitors, Gravity would be leading 40 percent to 30 percent to 30 percent. At that point, Gravity has much better odds of winning after the next redistribution. Say 12 Years a Slave then barely edges out American Hustle for second, then Hustle's ballots are redistributed according to whether its fans prefer Gravity or 12 Years as their next choice. Now, Gravity only needs one-third of Hustle's pile to get over the crucial 50 percent mark and win best picture. 12 Years, on the other hand, needs two-thirds of Hustle's pile to win.
What does it all boil down to? Keep in mind that piles are redistributed initially based on which film on a given ballot is ranked No. 2. If that movie has been eliminated, then that pile is redistributed based on its No. 3 choice, and so on. So, those No. 2 votes are especially vital. If a voter puts Nebraska as No. 1, it might not mean much in the end, but that voter's pick for second place could wind up deciding the entire race.
Ben Zauzmer, an applied math major at Harvard, has developed a math-based model using previous awards and other factors to predict the outcome of the Oscar race. His updated projections can be found at thr.com/Oscarlytics.
(Click to enlarge chart)