Netflix’s lifeblood comes from the movies it recommends, a gesture of exclusivity that supplies less-known titles based on user ratings. Through Cinematch, which analyzes viewers’ individual ratings and trends, Netflix generates new suggestions and uses them to keep queues full and customers committed.
But the system has remained scattershot, prone to
ham-fisted recommendations from computers that can’t consider human
factors. In response, Netflix launched a $1 million contest to any
person or team that can improve the system by at least 10 percent, and
there are now several hopefuls that have stalled with less than 1
percent to go—largely, some say, because of Napoleon Dynamite.
According to a new piece in The New York Times Magazine, the 2004
comedy, a phenomenon for some and an inescapable plague for others, has
proven such a divisive movie that competitors for the Netflix prize
have traced a large part of the error in their calculations back to
it. Movies like I Heart Huckabees, Lost in Translation and The Life
Aquatic With Steve Zissou have demonstrated similar effects, but none
like Napoelon Dynamite, which one user said accounted for as much as 15 percent of
his error rate.
The closest team vying for the Netflix money is
stuck at 9.44 percent. Napoleon, meanwhile, just polished off a
whole new set of haters.
Related links:
News: Netflix brings 12,000 movies and shows to TiVo users
News: Netflix deal with Starz adds 1,000 movies to Watch Instantly
NetflixPrize.com
Got news tips for Paste? E-mail news@pastemagazine.com.


Be the first to comment
Click to leave a comment.