March 22, 2013
Computer Makes March Madness Picks
It is that time of the year that employers hate; winter is about to give way to spring and employees will soon take longer lunch breaks and summer vacations are around the corner. But there is something else happening as well. We’re now in the middle of March Madness, which will continue for the next few weeks until the final four teams head to Atlanta and take court at the Georgia Dome.
Even the President, who has faced a budget crisis and the sequester, had time to make his picks for the yearly NCAA Basketball Tournament, and the media found time to cover it as well.
“The commander-in-chief likes Louisville, Indiana, Ohio State and Florida to make it to Atlanta, with Baylor, Connecticut, Notre Dame and California advancing in the women’s bracket,” USA Today reported.
ESPN’s Sports Center unveiled all his picks.
President Obama isn’t alone in liking Florida either. Georgia Tech’s Logistic Regression/Markov Chain (LRMC), the college basketball ranking system designed to use only basic scoreboard data also predicted Florida to make the Final Four and actually go all the way.
This system utilizes Logistic Regression” and “Markov Chain,” which are the two primary mathematical techniques and these take in account just the basics, including the two teams, whose court they played on and the margin of victory.
Perhaps Obama had some computer-like insight as LRMC further called Louisville and Indiana. Of course, that’s pretty easy as both schools are seeded number one. Obama chose Ohio State while the computer program picked Gonzaga, which is seeded number one, compared to OSU’s second place seeding. For the record, Florida was also a second place seeding.
In other words both Obama and the program are betting on the likely winners. In this tournament that’s essentially hedging one’s bet. It is unlikely all four teams will actually make it to Atlanta; after all part of the Madness of this is that there are going to be some upsets. A reported 25 percent of all NCAA tournament games end up as upsets.
However, it is unlikely that one (or more) of the teams will go all the way to the Final Four.
What makes this prediction by LRMC is that it is the first time in its 10 year history that a team that isn’t a number one seed has been called out to win the whole thing. However, given what Georgia Tech’s School of Industrial & Systems Engineering (ISyE) have put into the programming perhaps Florida simply deserved a higher spot in the seeding. Given that the LRMC remains one of the most accurate predictive ranking systems over the last 10 years, and has outperformed more than 80 others, including the NCAA’s Ratings Performance Index, there could be something to this.
It is the NCAA’s RPI that most experts now use to justify which teams should and should not get into the tournament and LRMC seems to outperform it.
“Our system combines the aspects of performance and strength of schedule by rewarding game performance differently according to the quality of each opponent,” said Professor Joel Sokol of Georgia Tech, whose research specialties include sports analytics and applied operations research, oversees the annual project. “Compared to something like RPI, LRMC is able to predict which team is better by taking the margins of victories and losses into account.”
The program goes a little further, dispelling some of the most popular myths about the tournament.
“The reason that you hear people say things like ‘Duke is one of the toughest home courts – it’s so hard to win there’ isn’t because of the court or the fans,” added Sokol. “It’s that Duke is usually such a good team. When you give them even a three- or four-point home court advantage on top of the skill advantage they usually have, it’s hard to overcome.”
Of course, as noted, the computer program is only predicting and not actually competing; it certainly can’t get past the human element, namely that humans can make mistakes and lose games they should otherwise win. This certainly explains IBM’s Watson super computer, which actually beat the top two Jeopardy! winners.
Image Credit: Photos.com