An IT Angle to March Madness

As the NCAA basketball tournament and the inevitable office pools have become more popular over the years, so have the methods and models used to predict winners. Some have become quite sophisticated, with many relying on information technology to create evermore elaborate analyses. A site that has taken college basketball forecasting (and by extension predictions on NCAA tournament matchups) to a new height could be Ken Pomeroy’s Web site kenpom.com, which was the subject of a New York Times article this week.

Pomeroy has really gone deep into statistical analysis, relying on such concepts as the Pythagorean calculation for expected winning percentage and the log5 formula. (Those are his links from his Web site.) Pomeroy calculates offensive and defensive efficiencies for each team. If you want to find out more, go to his methodology page here. For you technical IT folks, the read shouldn’t be a problem.

If you spend enough time on Pomeroy’s site, you may start to figure out how the analysis may give you a bit more information than simply flipping a coin to pick winners. But hurry, you only have less than two days to fill out your brackets, and the site is packed full of numbers.

For more on how prevalent NCAA tournament pools are in the federal workplace, go to our FedBlog for a discussion.