Statsgasm: Final Predictions for the 89th Academy Awards

Welcome back, friendos. Although I had hoped to get these done earlier in the week, I’m very happy to present the results of Awards Daily’s Statsgasm prediction models for the 89th Academy Awards. Just in the nick of time for many of you as you start to finalize your Oscar ballots!

This is the fourth year of Statsgasm’s existence, and although many might consider this year’s Oscars to be a cut-and-dried (i.e. BORING) affair in terms of predicting winners, frankly I think this year is the most exciting I’ve seen it recently from a stats perspective. As you will soon discover, Statsgasm’s models are anticipating quite a few closely contested races in a number of different categories, both above-the-line and in the techs.

To do a very brief review, Statsgasm utilizes small sample (penalized logistic) regression models to predict winners in 21 categories. Regression analysis is the underlying methodology for the prediction models used by Ben Zauzmer at The Hollywood Reporter and the venerable FiveThirtyEight.com, although there will still be differences in our results based on variable selection, ad-hoc adjustments, and other factors. If you’d like to have a more wonky but still guided tour on regression analysis, please review my earlier writings on the subject here.

As always, I must deliver the following comment before I reveal all 21 of Statsgasm’s predictions: probability by its nature deals with possibilities and not absolutes, so I don’t expect Statsgasm to bat 21/21, and neither should you. Statistical prediction models are dependent on the historical data set used to generate its predictions as well as the skill and intuition of the model builder. Often times, history is a good guide for the future. But we must also remember that sometimes the future may not decide to follow the precedent of the past, and will instead choose to defy and confound our expectations.

Warning aside, let’s see what Statsgasm has for us this year!

(NOTE: The following Tableau dashboard detailing Statsgasm’s prediction results is best viewed on a desktop computer)

In summary, Statsgasm has La La Land in the lead in 10 different categories. But there is a lot of uncertainty in a number of categories this year, so accordingly there is a wide variance in La La Land’s potential Oscar win count. I personally still feel the presumptive BP winner will take home between 7-9 Oscars.

Please feel free to ask any questions you may have in comments, or reach out to me directly! And happy predicting!