NCAA Bracketology and Other Sports Analytics Winners

March 17, 2011
86 Views

Basketball Team2 150x150 photo (predictive analytics)My all-star pick in this year’s March Madness may never touch the hardwood.  He’s Wayne Winston. He is a math expert, analytics guru and modeling hero who wrote the book on math in sports. Really!

Basketball Team2 150x150 photo (predictive analytics)My all-star pick in this year’s March Madness may never touch the hardwood.  He’s Wayne Winston. He is a math expert, analytics guru and modeling hero who wrote the book on math in sports. Really! “Mathletics” explores how fans, teams, gamblers and owners use math to predict results and prepare strategies.  Is a two-pointer or a long-range three a better play?  Go for a tie then win in overtime? Or, take the higher-risk three and win outright?  Winston looks for answers as a professor of operations & decision technologies at Indiana University’s Kelley School of Business, and as the former statistical guru for the NBA’s Dallas Mavericks.

Maybe Winston cannot sink a fade away three pointer — but he sure can compute, analyze, model data and plan accordingly.  Winston explores how specific plays, players and outcomes can be predicted based on past performance.  The ability to weight stats, social input — other people’s choices — and other factors may give you an edge in picking this year’s winner.

For the NCAA basketball tournament, Winston’s blog posted Sagarin ratings, created in the mid 1980s by Jeff Sagarin who has analyzed sports teams and player performance for USA Today every year.   This rigorous analysis explores the chances that each team wins each possible number of games.  Ohio State, for instance, has a 24.3 chance of winning the championship, a 12.2% chance of losing in the finals.

Before the glow fades from the MIT Sloan Sports Conference on Sports Analytics,”Mathletics” gives us another chance to see how modeling and predictive analytics help teams win — in sports and business.  We watched Harvard’s hoopsters fall in a buzzer-beater to Princeton in a one-game conference playoff and Washington’s Huskies win with the same way against Arizona.

Who could have predicted those results? … Maybe John Ezekowitz (@JohnEzekowitz), a Harvard sophomore and Harvard College Sports Analysis Collective co-president. John recently completed a model that shows the likelihood of upsets in today’s and Friday’s NCAA Tournament games. And even though he only gives Gonzaga an 11.3% chance of upsetting St. John’s, I’m going with Gonzaga for sentimental, not mathematical, reasons.  But evidence-based details make it hard to ignore the Buckeyes of  Ohio State.  What are your picks for this year’s Final Four?

Editor’s Note: If analyzing data sets to predict sports or business events interests you, join us for a complimentary Webcast Friday, March 25th, 2011, 11:00 ET titled Predictive Analytics with Spotfire.

David Wallace
Spotfire Blogging Team