FlowingData NCAA Tournament Bracket – Try and Beat Me

I just started the FlowingData NCAA tournament bracket. Join now. Try and beat me if you can.

The great thing about the tournament is that that you’re gonna hear tons and tons of statistics on what players have done, who’s favored to win, and who is without a doubt going to lose. Throw a huge dose of raw, human emotion and competitive spirit, and without a doubt, a lot of the data will mean absolutely nothing. I love it.

To make things interesting, to the winner goes a $20 Amazon gift certificate. If I win, nobody gets anything. Muauahahaha. Go on. I dare you. Join now and make your picks. Hurry though, because there’s only a couple of days left.

7 Comments

  • Andrew Ross March 17, 2009 at 4:09 pm

    Aw, I clicked through this post hoping you’d have a better design for the bracket sheet…

  • hegemonicon March 17, 2009 at 4:13 pm

    Interestingly enough, apparently using statistical data to predict NCAA winners is useless

  • I’m so not surprised. Thanks for the pointer.

  • Ditto on stats — there’s too much variability is in the data to be able to model it…even at the pro level (although a former professor of mine has had better luck http://tinyurl.com/dgh354, http://tinyurl.com/d28mhs). The only time I’ve won a pool was when I flipped a coin to pick the winners. I took that as a sign that pools are something I should avoid. Best wishes to all who enter.

  • Hmm, I tried this with two different browsers and it wasn’t working; I coudln’t add any picks. Oh, well. BTW, I tweeted yesterday about this article I saw in the Wall Street Journal about using statistics to pick games, though: http://bit.ly/N3r5

  • “Interestingly enough, apparently using statistical data to predict NCAA winners is useless”

    When is statistical data useful predicting human behavior?

  • Per Malcom Gladwell and Gerd Gigerenzer, I just went with who I recognized and had any feelings for. Seems to be working :)

    Gerd Gigerenzer has done research that seems to indicate that knowing about half of your data set and guessing the rest is the best way to win this type of thing. I’m hoping he’s right :)