Netflix Prize is Taken

I have stated on this blog that I am fascinated with the use of competition and prizes to find solutions to technical problems. The one real example that i have observed is the Netflix prize on which I last wrote about here. To recap, the prize of US$ 1 million was offered by a Netflix corporation in 2006 to any team or individuals that would provide a model to improve the predictions of Cinematch by at least 10%. I entered the contest sometime last year not because there was a chance that I could win but merely to get access to that data.

It is now vertain that the prize will be claimed within 30 days because the Netflix Prize Leader Board shows that the Bellkor's Pragmatic Chaos team presented an algorithm that has managed to improve the predictive ability of Cinematch by 10.05%. While I understand that there's a probability that another team could present an alternative that is better, I think that the probability is too low that Netflix should just write the cheque.

That aside, the lesson from this to me is that innovations cares not for recessions but through competition, the learning takes place and real solutions are developed. Secondly, it is important to note that the prize was announced in October 2006 and the teams have worked hard to deliver the prize in less than three years. The third conclusion that I draw is that the similarity of names for a good number of the leading teams suggests that they may be the same teams using different techniques. Since the prize requires that the winning team will not only claim the prize but also explain how this was done, this last point will be proved one way or another.

From Bookish Posts, congratulations to the Bellkor's Pragmatic Chaos team for delivering a lesson in the use of prizes. If only the public sector would follow suit and pay for results and not effort.