September 21, 2009

Netflix Awards $1 Million, Announces New Challenge

Netflix has announced the winner of its $1 million challenge to researchers who came up with the best way to improve its movie recommendation system.

On Monday, the online movie rental service announced that team "BellKor's Pragmatic Chaos," had been awarded the $1 million prize for reaching the company's goal of a 10 percent improvement over the accuracy of its movie recommendation system at the start of the challenge in 2006.

BellKor's was actually a combination of three different groups that had previously competed against each other in a different competition. The team narrowly beat out a group called Ensemble for the $1 million prize.

BellKor's algorithm accounted for a 10.06 percent improvement over Cinematch's score on the test subset at the start of the contest.

"We had a bona fide race right to the very end," said Netflix Co-Founder and CEO Reed Hastings.

"Teams that had previously battled it out independently joined forces to surpass the 10 percent barrier.  New submissions arrived fast and furious in the closing hours and the competition had more twists and turns than 'The Crying Game,' 'The Usual Suspects' and all the 'Bourne' movies wrapped into one."

Team members Bob Bell, Martin Chabbert, Michael Jahrer, Yehuda Koren, Martin Piotte, Andreas Töscher and Chris Volinsky were awarded the grand prize at a ceremony in New York City on Monday.

When Netflix issued the challenge in 2006, it provided engineers with 100 million anonymous movie ratings ranging from one to five stars.

"All personal information identifying individual Netflix members was removed from the prize data, which contained only movie titles, star ratings and dates but no text reviews," Netflix said in a statement.

According to the Associated Press, the most recent count showed there were more than 51,000 contestants from 186 countries.

Also, Netflix on Monday announced the "sequel" challenge. The next $1 million challenge will involve Netflix users who rarely rate the movies they watch.

"Accurately predicting the movies Netflix members will love is a key component of our service," said Netflix Chief Product Officer Neil Hunt

"This extreme level of personalization is like entering a video store with 100,000 titles and having those that are most interesting to you fly off the shelves and line up in front of you.  We take the guess work out of renting by presenting the movies and TV episodes we believe each Netflix member will most enjoy."


On the Net: