Using historical data to predict the Oscars – an exercise I’ve undergone each of the past ten years – is ultimately a statement that the past is a decent predictor of the future.
My model encompasses data from a number of critic circle awards, guild awards, and other sources, weighting each of them based on how well they’ve predicted each category in past years.All things considered, the model has done fairly well.
Last year’s nominee predictions went 40/44, and last year’s award predictions went 17/21. But what happens when the present is so strange as to look nothing like the past?Welcome to the weirdest year of Oscar prediction.
Read more on hollywoodreporter.com