Okay, friends, we’re definitely in crunch time for this year’s Oscars, and I am pleased to finally (belatedly) reveal the findings of Awards Daily’s Statsgasm prediction models for the 88th Academy Awards. If you’re new to the site, I developed these math-based models back in 2013, using a statistical technique called small sample (penalized likelihood) logistic regression analysis to analyze historical data. This is a similar methodology used by the Oscar prediction models from FiveThirtyEight.com and Ben Zauzmer (among others), but I do believe that the years of intuition and insight I have acquired about the Oscars from being a long-time fan (and friend) of OscarWatch/Awards Daily will give these models a leg up over their brethren. I won’t go into further detail in this post about regression analysis, but if you are interested, I strongly encourage you to revisit Statsgasm’s archives here on the subject.
My one reminder/comment before I reveal the results of all 21 Statsgasm prediction models: statistics and probability are *not* absolute! Statsgasm only indicates what may be likely to happen, what the expected (average) outcome might be given the historical sample that was used to generate its predictions. The process of statistical model building is as much art as it is science: much depends on the creativity and intuition of the builder in addition to knowledge of statistical methods. The underlying methods may be similar among different models, but modelers do have some latitude in variable selection and adjusting the estimated weights of the predictors in accordance with their views. And sometimes, there are outcomes that simply defy all historical precedent. Such results may very well happen on February 28th.
TL;DR – I don’t expect Statsgasm’s models to bat 21/21, and neither should you! For the record, Statsgasm was 16/21 in 2013 and 15/21 in 2014. Statsgasm’s Best Picture model produced Gravity as the predicted winner in 2013 (incorrectly) and predicted Birdman to win last year (correctly).
That said, let’s see what Statsgasm has in store for us this year. “Ready when you are, Sgt. Pembry.”
“Good evening, ladies and gentlemen. We are tonight’s entertainment!”
I will be adding more detailed commentary and analysis to each of the model’s findings between now and Oscars Sunday, so please revisit this post for updates. And as always, feel free to leave comments or contact me on Twitter or email if you have any questions!