Netflix prize awarded to BellKor's Pragmatic Chaos

New contest starting up

View more images

22 September 2009 9:27 GMT / By Duncan Geere

Netflix's million-dollar prize for the team which could most improve its movie recommendations service has been awarded to a team of statisticians, machine-learning experts and computer engineers from across the world who call themselves "BellKor's Pragmatic Chaos".

It had previously been reported that the prize had been won, but that it wasn't clear by whom - the contest was to make the recommendations 10% better but two teams crossed that threshold before the timer expired - BellKor and another group called the Ensemble.

The race was close - BellKor presented its final submission 20 minutes before the deadline, and then the Ensemble got theirs in a few minutes later - just before time ran out. The teams were mathematically a dead tie, so the contest rules dictated that the team who submitted first took the prize.

Netflix is setting up a second prize - an attempt to model the "taste profiles" of individuals. Contestants will be given demographic and behavioural data of 100 million individuals and asked to recommend them movies. The results will then be compared with what the individuals actually rented.

There's no target this time round, instead it's a straight race - with $500,000 given to the leading team after 6 months, and another $500,000 given to the team doing the best after 18 months. "Three years was a long time to compete in Prize 1", said Netflix's Neil Hunt, "So the next contest will be a shorter time limited race, with grand prizes for the best results at 6 and 18 months".

If you feel like entering then full details will be on netflixprize.com shortly.