Basketball is a spectator sport. Mathematics is not.


Or is it? This month, hoops-loving statisticians get to ply their craft on the dunking, dribbling, alley-ooping data set which is the NCAA men's basketball tournament. Yes, mathematicians have the March Madness, too.

Scientific American obliges with a quick rundown of scientists who are getting in on the
game, including University of Illinois computer scientist Sheldon Jacobson, who argues that seeding is irrelevant when it comes to picking winners in the Elite Eight and beyond. His study was published in the Journal of Gambling Business and Economics--the very existence of which merits a study of its own.

But if you really want to get your brain in a bunch, try figuring out the probability of picking an exactly correct bracket. The American Institute of Physics puts the odds at one in 9,223,372,036,854,775,808. That's one in 2 to the 63rd power. Thanks, combinatorics, for busting my bracket!

But Georgia Tech professor Joel Sokol thinks he has the secret to beating those odds, and it's called a Logistic Regression/Markov Chain. Didn't hear about it on PTI? The system uses scoreboard data from the regular season to pick likely winners on a neutral court, giving less weight to photo-finishes than to routs. Last year, it led him to a perfect Final Four, championship game, and tournament winner pick. Stay tuned to see how he'll fare this year.

You may now return to checking your bracket, which thankfully doesn't use up as many productive hours as you thought it did.
blog comments powered by Disqus