CompassPoint Media takes on the 91st Academy Awards

For the last four years, out Marketing Performance group has predicted the big four Oscar categories – Best Picture, Best Director, Best Actor, and Best Actress.  Using nothing but data and a few machine learning algorithms, the team has correctly predicted 14 of the past 16 winners in these categories.


Last year, they got all four right.  This year, we’re at it again with the following predictions for the 91st Academy Awards ceremony this Sunday, February 24, 2019.


The predictions below summarize the modeled probability of our predictions and compares them to current Las Vegas odds.  Las Vegas odds were collected on 2/20/19 from


Best Picture = Green Book

Model probability of winning: 10-49%

Las Vegas probability: 22%

Key Data Points:

·       Winning best picture at Producer’s Guild Awards

·       IMDB user rating

·       Total days between the film’s release in the US and the Oscar ceremony


The odds makers have Roma as the favorite in this category, with a 77% chance of winning, but Green Book got just a little more love from our models.  According to our data, this category is the most difficult to predict this year, so if we get one wrong, this is likely the one.


Best Director = Alfonso Cuaron (Roma)

Model probability of winning – 61-99%

Las Vegas probability – 95%

Key Data Points:

·       Winning best director at Director’s Guild Awards

·       Total Oscar nominations

·       Winning best director at Golden Globes


The models and the betting public are in violent agreement on this category.  We’ve never missed in this one, so we’d be VERY surprised if this award went to anyone else.


Best Actor = Rami Malek (Bohemian Rhapsody)

Model probability of winning – 66-99%

Las Vegas probability – 80%

Key Data Points:

·       Winning best Actor (drama) at Golden Globes

·       Winning best Actor Screen Actors’ Guild

·       Age of Actor in release year


Our models and the Vegas line agree on this category as well.  Rami Malek will likely win, but don’t count out Christian Bale.  We’ve never gotten this category wrong either, but there is a first time for everything.


Best Actress = Glenn Close (The Wife)

Model probability of winning – 62-63%

Las Vegas probability – 83%

Key Data Points:

·       Age of Actress in release year

·       Winning best Actress at Golden Globes

·       Total nominations for picture


Best Actress has historically been the toughest category for our models to predict.  This year, the models are split between Glenn Close (The Wife) and Olivia Colman (The Favourite).  We’re going with the two models that have the best historical performance and picking Glenn Close.  This agrees again with the Vegas line as well.


The Science

To generate our predictions, we used three machine-learning models commonly used to classify unknown quantities based on known data.  In this case, we are classifying the Oscar nominees as either winners or losers based upon a data set with 99 different factors.  What we know about the winners from previous years then informs the model’s picks for the current year.  We encounter similar models all the time.  They help us find movies we’d like to watch on Netflix, or books we’d like to read on Amazon.


Data Sources:

Academy Awards Database:

Internet Movie Database:

Roger Ebert Reviews:

Screen Actors Guild:

Producers Guild:

Directors Guild:



Pardoe, I. and Simonton D. K. (2007) Applying Discrete Choice Models to Predict Academy Award Winners, J. R. Statist. Soc. A(2008), 171, Part 2, pp. 375-394