The Netflix movie rental business has given a $1 million cash prize to seven programmers and statisticians who developed a new algorithm to provide more accurate movie and TV episode recommendations to subscribers.
Netflix, which offers more than 30 billion recommendations each day, will be able to reportedly double the accuracy of predictions in the future. The contest was first launched in Oct. 2006, and requested participants only increase movie recommendations by just 10%. More than 40,000 teams from 186 countries participated, with two teams eventually tying for the original prize.
"Right now, we're driving the Model T version of what is possible," Netflix CEO Reed Hastings said in 2006 when speaking of its recommendations. "We want to build a Ferrari and establishing the Netflix Prize is a first step."
The winning team, BellKor Pragmatic Chaos, submitted the new winning algorithm that had a 10.06% accuracy increase, according to Netflix.
The team that finished in second place -- which lost the competition due to time restrictions agreed upon at the start of the challenge -- isn't worried about the $1 million cash prize, as the statistical analysis and predictive modeling design created for the project is now being used by several of the company's clients.
"We've already had a $10 million payoff internally from what we've learned," Opera Solutions CEO Arnab Gupta told the New York Times. "So for us, the $1 million prize was secondary, almost trivial."
I've heard complaints from several Netflix subscribers who have had difficulty with the company's recommendations. I used to find it sometimes amusing to be recommended movie titles not even remotely related to whatever I was watching -- but I did understand their frustration.
