(WIRED) -- For the past three years some of the world's most resourceful software engineers and mathematicians have worked feverishly in the spirit of friendly, Darwinian competition -- and in pursuit of a $1 million prize -- to improve the Netflix movie recommendation system by 10 percent.
Netflix says it will announce the winner of its $1 million Netflix Prize at an event September 21.
It's been a long journey, but it's almost over.
Netflix said Tuesday it would announce the winner at an event in New York on Sept. 21 -- where details about the Netflix Prize 2 sequel would also be unveiled.
As things stand now, nobody outside of Netflix knows which team will win the million-dollar prize, but the winning entry will power Netflix's upcoming movie recommendation engine. Because nothing about the contest barred entrants from licensing their technology to companies other than Netflix, both qualifying teams could find plenty of licensees -- even if they come in second.
And not much separates the two top teams. Teams Bellkor (AT&T Research), Big Chaos and Pragmatic Theory combined to form Bellkor's Pragmatic Chaos, the first team to qualify for the prize on June 26 with a 10.05 percent improvement over Netflix's existing algorithm.
This triggered a 30-day window in which other teams were allowed to try to catch up -- and indeed, a team called The Ensemble, made up of lower-ranked contestants, submitted a higher score of 10.10 percent as time ran out -- a hair better than Bellkor's Pragmatic Chaos' final score of 10.09 percent.
Before Sept. 21, Netflix must decide (assuming they haven't done so already) which of the two qualifying teams has the best algorithm based on how they score on various undisclosed tests. The company must also determine whether the winning team's results are reproducible, meaning that its algorithm handles new data as well as it did the test data.
When we wrote about this last month, after the window for qualifying closed, Netflix declined to say whether it will reveal details about its methodology for picking the winner (whichever team loses, they're going to want to know how, exactly, that happened -- as will curious onlookers around the world).
A Netflix spokeswoman told us, "We are definitely going to touch on the methodology in the press conference," and offered to let us interview Netflix executives to find out more about how they picked the winner.
Netflix vice president of communications Steve Swasey told us last month that the company is pleased with the results of the contest, and apparently, it hopes there's more where that came from. At the same New York event where it will announce the Netflix Prize winner, the company plans to unveil details about Netflix Prize 2.
Writing on Netflix message boards, chief product officer Neil Hunt already offered some hints: "The next contest will be a shorter, time-limited race, with grand prizes for the best results at 6 and 18 months. While the first contest has been remarkable, we think Netflix Prize 2 will be more challenging, more fun, and even more useful to the field."
More than 40,000 teams from 186 countries competed for the first Netflix Prize. Some contestants doubted that any team would be capable of surmounting the 10-percent-improvement barrier, but two teams managed to do so.
Clearly, Netflix has stumbled onto a winning formula here, and will walk away from the contest with a measurably improved movie recommendation engine -- and recommendations are a linchpin of its business.
As the field of "prize economics" evolves, offering organizations increasingly efficient ways to solve their problems, we expect results like this to become more commonplace.
Subscribe to WIRED magazine for less than $1 an issue and get a FREE GIFT! Click here!
Copyright 2009 Wired.com.
|Most Viewed||Most Emailed||Top Searches|