Researchers Solve Netflix Challenge, Win $1 Million Prize

A team of researchers Monday won a $1 million prize for developing a formula that improves the accuracy of the Netflix movie recommendation algorithm " beating another team by just 20 minutes.

In October 2006 Netflix offered the $1 million prize to anyone who could improve its recommendation algorithm by 10 percent. The challenge proved to be a complex problem in statistical analysis and predictive modeling applied to a data set of 100 million movie ratings.

Thousands of teams from around the world were in on the competition. In an interview with The New York Times, Netflix CEO Reed Hastings said the contest, despite the $1 million prize, proved to be a bargain for the company. "You look at the cumulative hours and you're getting Ph.D.'s for a dollar an hour," he is quoted as saying.

The winning algorithm was developed by a team known as "BellKor's Pragmatic Chaos" that was made up of seven researchers, engineers and scientists from around the world. Their algorithm, which improved the predictability of movie preferences by 10.09 percent, was submitted at 6:18 p.m. on July 26, according to the contest leaderboard.

That was just 20 minutes ahead of a submission by a team called "The Ensemble" whose algorithm offered accuracy improvement of 10.1%

The algorithm Netflix was seeking to improve uses data from subscribers who regularly provide ratings on movies they've watched, according to an Associated Press story. Netflix is planning on running a second, even tougher contest, that will seek better ways to predict the preferences of subscribers who rarely " or never " provide movie ratings, according to the Associated Press.