Wednesday, March 18, 2009

March Madness

It's March, and that means time to fill out the college basketball bracket! For some reason, this year there are a number of articles concerning the math of making predictions. Here's a computer model from a professor at Georgia Tech, and here's a rundown of efficiency measures in light of this year's bracket. Note that accuracy based on last year's tourney should be taken with a grain of salt-- models like these generally pick higher seeds over lower seeds (as they should), and last year was an aberration as all four top seeds advanced to the Final Four.

Feels like a Louisville year.

6 comments:

rolub said...

No interest in guessing the over/under of wins by the #1 seeds? 13.5 by the way.

Justin M Ross said...

I have WVU falling to NCU in the final.

No bias there ;)

Matt E. Ryan said...

13.5? 12 gets them all to the Elite 8...I'll take the over, but enjoy tourneys more when higher seeds lose.

rolub said...
This comment has been removed by the author.
rolub said...

Bad math in previous post. Should have read:

Matt, last year was the exception, not the rule. The under is plus money and I've happily thrown some change at it. If one of them goes down in round two, you need chalk all the way through to win. Any other loss that does not optimize the #1 seed's path without getting in the way of another #1 will pay the under.

Matt E. Ryan said...

An early loss certainly would put a wrench in it. And per the post, it's certainly not worth putting a lot of stock on a theory based on being accurate last year. :)