The Netflix Prize and other challenges.

On October 2, 2006 Netflix offered the following challenge. The first team to estimate ratings 10% better than Netflix’s system would earn $1 million. The choice of 10% was remarkably prescient. After two years of concerted effort, the best current algorithm is 9.44% better than Netflix’s algorithm, thus creating lots of excitement about the prize and lots of good publicity.

It has also generated folklore that is clearly wrong, but widely quoted. It is referred to as the “Napoleon Dynamite Problem.”

Len Bertoni, a computer scientist devoted to developing a winning algorithm, said that the reason he can’t reach the 10% improvement level is because of movies such as Napoleon Dynamite, which are notoriously hard to predict. But Bertoni has it entirely wrong. It is precisely the movies that are extremely hard to predict that provide opportunities for Bertoni and others to reach the 10% goal.  Napoleon Dynamite is hard for Netflix’s system too, and leaves substantial opportunities for improvement. On the other extreme, if a movie is easy to predict, then the Netflix system may already do so well that there is little room for further improvements.

I hope that many other companies copy Netflix’s idea, and offers their own challenges. This already occurs on a regular basis because of the company Innocentive Inc., .  This company acts as a broker for prizes, where the prize value ranges from $10,000 on up.  The prizes are for technical work in lots of different fields. Some of the prizes may be of interest to those in the OR community. (Thanks to Mary Ann Gorman for alerting me to this company.)


One Response to “The Netflix Prize and other challenges.”

  1. Laura Says:

    I had the same reaction to the “Napoleon Dynamite Problem” — if it weren’t for a few polarizing movies, the whole thing wouldn’t be as hard as it is. It has been exciting to see all the articles about statistics/math/CS/OR in popular outlets about the Netflix prize. Hopefully it will encourage kids to pursue scientific disciplines.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: