Christopher Coulter asks about this article about a computer program that supposedly predicts box office
. Unfortunately, one of the variables used in predicting box office is how many theaters the movie's opening in. Typically, if a movie's a disaster, it doesn't open in as many theaters as a movie that's at least okay. The studios and exhibitors dump it. Another variable is strength of cast. You don't get a strong cast, on average
if the script is dreadful. So the program is meaningless, especially since it's only vaguely right three quarters of the time. If I could tell you within $100 million dollars how well a movie was going to do, but only 75% percent of the time, would you consider me a marketing whiz?
What's much more meaningful is that they can by and large tell by Sunday how well a movie's going to do box office, basically by looking at how word of mouth spreads from Friday to Saturday.