Huffington Magazine Issue 35-36 | Page 25

THE OSCAR ISSUE / HUFFINGTON / 02.10-17.13 v o i c e s david rothschild subjective ratings, Oscar nominations and a nominee, they are not predictive of biographical data. The reason is simple: the eventual winner. For example, movPrior to the other awards shows, there is ies released late in year are more likely a dearth of variables that properly idento get a nomination for an Oscar, relatify individual award categories, as most tive to movies released in the spring; but data is just movie specific. conditional on getting nomination, they But, there are two goals of fundamenare no more likely to win the Oscar. tal models: forecasting and determining which variables have predictive power. BOX OFFICE SUCCESS: This category includes While fundamental models do not make variables like: gross revenue, screens, avgreat forecasts for the Oscars relative to erage gross revenue per screen, these valother data including prediction markets, ues on the first week of wide release and they can still provide insight into which the first four weeks of wide release, and variables we should follow. many other combinations. Between gross All of the insights in this column are revenue and number of screens there are into the predictive power of variables, some really interesting variables to conconditional on a movie getting a nominasider here. This is further complicated tion for an Oscar, at the by the staggered opening time of the nomination. of many Oscar nominated How well a movie does in movies. After much investiIT IS NO the box office, especially gation, the predictive power SURPRISE after a few weeks, the popuin this category is highly THAT THE lar ratings, and h ow many correlated with the change OSCAR nominations the movie that happens over the first receives are all significant few weeks. A key inflection VOTERS predictive variables. point appears to be between VALUE weeks four and five. For THEIR OWN STUDIO INPUTS: This category Best Picture I follow this JUDGMENT , includes variables like variable closely: 2*Gross AND MOVIES budget, release date, genre, Week 5 - Gross Week 4. WITH MORE and when the movie goes From week four to week NOMINATIONS to wide release. Some of five, Argo went from $13.3 these variables correlated million to $9.0 million, TEND TO strongly with whether a while Lincoln went from DO WELL IN movie gets a nomination, $18.0 million to $12.4 milWINNING but conditional on being lion. Thus, from this ru- ‘‘ OSCARS. ‘‘