Predicting box office success

Lots of movies are coming out during the holidays. While they're fun for us, for Hollywood, it's white-knuckle time, as they wait to see if their huge investments pay off. Two scientists at Oklahoma

Friday, December 16th 2005, 9:50 am

By: News On 6

Lots of movies are coming out during the holidays. While they're fun for us, for Hollywood, it's white-knuckle time, as they wait to see if their huge investments pay off. Two scientists at Oklahoma State University think they've come up with a computer program that can take the guesswork out of movie-making.

News on 6 reporter Steve Berg has more on their movie magic.

Movies do have a magical quality, not exactly something you would associate with advanced mathematics. Dr. Dursun Delen: "Most of the people think that movies are unpredictable. We argue otherwise." Dr. Ramesh Sharda: "It's one of the really fun projects that we've worked on."

Oklahoma State's Dr. Ramesh Sharda and Dr. Dursun Delen came up with the Movie Forecast Guru. It's the product of 8-years of number-crunching and draws data from 834 movies. Predicting movies box-office results according to nine categories. Dr. Ramesh Sharda: "Class 1 is flop, less than 1-million. Class 2, just as an example, is 1 to 10-million. And Class 9 is over 200-million."

In other words a blockbuster. So how do they do it? They input 7 different variables, everything from special effects, to the movie's rating to star power. What makes their program different, they say, is that it uses what they call a neural network, which they say can measure the effect that all the different factors have on each other.

Dr. Ramesh Sharda: "A neural network is able to take these various parameters in combination, not just individually. So it's quite possible that one Rated "R" movie will do poorly, but another Rated "R" movie will do quite well, because it's able to take into account all the other factors."

The computer has correctly guessed 75 percent of movies within one class or less, but not always. Dr. Dursun Delen: â€œfor instance, we would miss Waterworld, we would classify Waterworld to turn out to be a blockbuster. But it did not turn out to be a blockbuster, not even close. Another flipside of it for instance would be My Big Fat Greek Wedding. We would never classify that movie as a blockbuster, but it was one of the major blockbusters, it's surprising."

The hardest thing to guess? Star power. Dr. Ramesh Sharda: "Because the same stars do not remain stars over time."

The professors say they are already in talks with one major studio about the rights to their program.