DVO Podcast: Predicting Blockbusters?

DVO Podcast: Predicting Blockbusters?

Recently published studies claim to be able to predict a movie’s success.  We get out our clipboards and run the numbers.


DVO Podcast: Predicting Blockbusters?
Reading scientific studies 101: Shit versus Shinola.

The Guardian recently publicized a new study that claims to know which kind of plot structure makes the most reliable amounts of money.  Digging deeper, you can find dozens of similar studies that claim much the same.  One says it can predict opening weekends by tracking Wikipedia entries.  Another claims watching Twitter trends is the key.  Even Google is being fielded as a potential box office prognosticator.  Erik and I sift the studies, looking for Shinola and asking our own questions:

  • Are these models any better than going by historical trends?
  • How much predictive power is there in such overly broad categories they use?
  • If you can predict even modest, but reliable returns, will it change which movies get made?
  • What about the dark horses, flukes and trend setters that tend to shape any given decade?

Follow us as we try to separate the signal from the noise, and discuss how the growing reliance on algorithms may shape the movie making landscape.

DVO Podcast: Predicting Blockbusters?


Time Stamps:

1:00 – overview of the main studies.

6:00 – laying out the questions these studies bring up.

11:00 – picking apart the methodology of the major studies.

17:00 – exploring algorithms and their growing consequences.

19:00 – accounting for statistical noise and unexplained aberrations.

20:40 – dealing with outside factors like franchise fatigue, over-saturation.

29:00 – comparing these studies against good old historical trend spotting.

33:10 – if the algorithms are right. what changes for Hollywood?

Be the first to comment

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.