1 Comment

Summary:

Earlier this week, I weighed in on the NCAA tournament with my data-influenced picks, and now the experts have crunched the numbers and are weighing in with their purely objective picks. If you haven’t filled out your bracket, you might want to read this.

generic_mbasketball_stan

Earlier this week, I weighed in on the NCAA tournament with my data-influenced picks, and now the experts have crunched the numbers and are weighing in with their purely objective picks. If you haven’t filled out your bracket, you might want to read this.

Rothschild's picks.

I’m focusing on info from Yahoo, whose David Rothschild based his predictions on info from Betfair and PredictWise, and the New York Times, whose Nate Silver has his own round-by-round model for picking favorites. Yahoo says it can predict the outcomes of events with a 3 percent margin of error, while the Times model performed very well in the early rounds last year but was thrown for a loop in the upset-heavy later rounds.

Here are some key takeaways:

  • Kentucky is the odds-on favorite no matter whose model you choose. Yahoo gives it a 26.9 percent chance of winning it all, while the Times gives it a 26.4 percent chance.
  • Both models have Kansas as the fourth most-likely winner, although they swap Ohio State and North Carolina at No. 2 and No. 3 spots.
  • Silver says the No. 16 teams are better than they have been in years, but that still doesn’t make them likely to win a game.
  • If you want to pick a first-round upset, Silver’s model says it’s 11-seed North Carolina State in the Midwest region, with a 59.8 percent chance of winning.
  • Of the teams seeded fifth or worse in their regions, both models have Vanderbilt in the East region as a good Cinderella pick to make a deep run, but Silver’s model also likes 5-seed New Mexico and 8-seed Memphis in the West region.
  • The models are actually pretty close through the top 10 favorites to win the tournament, with one glaring exception — Florida State. It doesn’t make Silver’s top 10 (in fact, his model gives it a 0.9 percent chance), but Rothschild has it in the No. 7 spot with a 5 percent chance.

Of course, picking basketball games is just fun. If you want to hear about some example of analytics actually making people money in business rather than with bookies, come to or watch our Structure:Data conference next week.

And if you want to know where to watch the March Madness action online, here’s your guide.

Image courtesy of the University of Kentucky.

You’re subscribed! If you like, you can update your settings

  1. Interesting post. I did my own analysis of March Madness here:

    Predicting March Madness through Social Media Analytics http://bit.ly/yKDewq #marchmadness #socialanalytics #ncaa #sports

Comments have been disabled for this post