OscarLytics: Which Nominations Are a Lock — And Which Might Surprise

Editor’s Note: Ben Zauzmer (@BensOscarMath) is a big fan of the Oscars, as well an applied math major at Harvard. For the past two years, he has predicted the Oscars using nothing but math, calling 75 percent in 2012 and 81 percent in 2013. This year, he’s teaming up with The Hollywood Reporter to bring you his Oscar predictions, as well as weekly updates, combining numbers and movies in the weeks leading up to the Academy Awards. Ben’s predictions are purely math-based; see THR‘s Feinberg Forecast for a look at the Oscar race that takes into account other factors).

Statistically speaking, the filmmakers behind 12 Years a Slave are assured an early morning call when the Academy of Motion Picture Arts and Sciences announces its best picture nominees Jan. 16. Based on all the available evidence to date, the probability of Slave getting a best pic nom is 99.9 percent.

But because no one knows how many movies the Academy will nominate — there will be at least five and as many as 10 — many other filmmakers can expect to to suffer through a sleepless night waiting for the nominations to be revealed.

MOST LIKELY BEST PICTURE NOMINEES

Here’s a behind-the-scenes look at the formula I used to create these probabilities. My method relies only on math; no personal opinions or hunches are involved.

For each Oscar category, I compiled all of the significant indicators from past years: awards shows, guild awards, various film critics groups, and Metacritic and Rotten Tomatoes scores. With all of this data in a chart for each potential nominee in a category, I next used a statistical formula to derive the best approximation of the relative weights of each award and critic score. Indicators that have performed well during past awards seasons are weighted more heavily. These weights were then applied to this year’s potential nominees — one formula for each category — to calculate the final percentages for each movie or star within each category.

For the more mathematically inclined, here’s the best picture formula when written like something out of a high school math class:

y = a₀ + a₁x₁ + a₂x_{2 …} + a_Nx_N

The a‘s are the weights for each predictor, the x‘s are the different Oscar nomination predictors, and the y is the final percentage for each movie.

Consider the best actor race, for example: Forest Whitaker and Idris Elba each have one major nomination so far — Whitaker from the Screen Actors Guild and Elba from the Golden Globes. But the Screen Actors Guild Awards are historically better than the Golden Globes at predicting Oscar nominees for best actor, so Whitaker (39 percent) has a higher nomination chance than Elba (9 percent).

MOST LIKELY BEST ACTOR NOMINEES

Of course, the Oscars are a human endeavor, so they cannot be modeled perfectly by mathematics. The Academy of Motion Picture Arts and Sciences is just that — a combination of art and science. These predictions may prove to be accurate, but indicators alone cannot guarantee any results.

Once we learn who the nominees are on Jan. 16, I will use a similar method to regularly track the likeliest Oscar winners in each race right up to the Academy Awards on March 2.

Till then, here are how the probable nominees stand in the other three acting categories.

MOST LIKELY BEST ACTRESS NOMINEES

MOST LIKELY BEST SUPPORTING ACTOR NOMINEES

MOST LIKELY BEST SUPPORTING ACTRESS NOMINEES

Email: bzauzmer@college.harvard.edu
Twitter: @BensOscarMath