http://archive.ics.uci.edu/ml/datasets/Netflix+Prize
http://www.netflixprize.com/faq
simple corrupt file or corrupt contest? i throw around the word 'corruption' not cavalierly, but because when there is a million dollars at stake, and the future legitimacy of large-scale programming contests, it's a little odd that netflix hasn't released the actual scores of the top two teams. nytimes, bbc and others report that the 'percentage improvements' were equal, but that's not what the rules say. the rules say it should be the best RMSE down to 0.0001 (if i recall). so my point is -- with so much on the line, why hasn't netflix released the final scores? and why don't the checksums match?
here's the other hacker news page where i got downvoted like a m-f-:
http://news.ycombinator.com/item?id=834632
again, i have no association with any team or any part of this programming contest. i just read 'the ensemble's blog post a month ago and thought it was cool how they raced to the deadline and appeared to win. and also, i think it's f-ing ridiculous if a corporation backed down on its rules.
hopefully they just uploaded the wrong file. but with that one checksum basically legitimizing the entire contest, you'd think they'd take it a little more seriously. just doesn't add up.
p.s. i'd contact netflix directly if they had an e-mail address available or something -- but i have no interest in registering for the forum. i'll delete this post as soon as someone explains why the checksums don't match. don't want to make a false accusation.