Netflix, for several years, has been running a competition to try and improve the accuracy of its prediction algorithm. Teams have been competing to improve its movie recommendations by 10%. If reports are accurate, then the prize appears to have been won.
But not just by one team. As per the rules, once a team breaks the 10% barrier, other teams get 30 days to surpass it before the competition closes. A team called BellKor’s Pragmatic Chaos, made up of engineers from AT&T, Commendo, Pragmatic Theory, and Yahoo Research, has hit the 10% mark.
But, in a nail biter, it looks as if another team - The Ensemble might have pipped them to the prize. For the dataset the teams have been using, Ensemble wins out, but now they'll both be compared to a completely new set of data, and the scores averaged.
Whichever team wins will be announced in a couple of weeks, and they'll get $1 million for their trouble. The loser gets nothing, so you can bet that members of both teams will be lying awake at night, wondering what they'll spend the cash on.
A subscription to Netflix, perhaps?