Big Data: March Madness Math in the Real World

Shane Larkin #0 of the Miami Hurricanes concentrates at the free throw line
Lance King | Getty Images Sport | Getty Images
Shane Larkin #0 of the Miami Hurricanes concentrates at the free throw line

They did it for the love of The Big Dance—and Big Data.

A couple of self-described frustrated jocks teamed up more than a decade ago to produce a model for predicting which college basketball teams would make it to the NCAA Tournament.

Jay Coleman and Allen Lynch, now both business professors, also bring those same analytic tools to practical business problems—be it a construction outfit weighing bids or an NBA team trying to maximize ticket sales.

They bumped into each other in the halls of University of North Florida one day, "And Jay told me he had this great idea," said Lynch. Soon they'd produced what they called The Dance Card, which crunches such data as team RPI, wins vs. top teams, conference strength and more, to predict what 68 teams of the eligible 347 the selection committee would invite.

(Read More: 'Moneyball's' Beane: Business of Baseball Has Changed Forever)

The first 31 are easy—they're the conference winners who get automatic berths. Using statistical analysis software called SAS, in the first decade they predicted the remaining 37 teams with about 93-94 percent accuracy. They've since tweaked their formula and brought in a third member, economist Mike DuMond of Charles River Associates, and last year got 36 of 37 right. This year—perfection.

"We introduce this in the classroom because it's so useful in the real world," said Lynch, who is now a professor of economics at Mercer University in Macon, Ga. While quantitative research might usually be a hard sell to undergrads, "It's a great way to get their attention when you walk into the classroom on the eve of selection with The Dance Card."

The key usefulness of this kind of analysis is to predict the likelihood of a given event based on an institution's history, which is captured in data, Lynch said.

(Slideshow: Is March Madness So Bad? Comparing Productivity Killers)

The former research analyst with Blockbuster offers the example of an acquaintance whose construction company was trying to determine for which public works projects it should submit bids—a costly process. The company input data on the kind of bids it has won in the past: the region where the work will be performed, the size of the project, the type of construction, even known contacts with public officials. The software offered probabilities on winning bids, Lynch said, so the company could minimize time and money wasted on losing ones.

Sports is a natural fit for this kind of analysis, because it tends to produce a lot of quantifiable information so performance can be measured very efficiently, Coleman said. Indeed, data-driven team management is the basis of the hit movie "Moneyball," which was based on the book by Michael Lewis.

Since 2007, MIT's Sloan School of Management has hosted a conference focused on the increasing role of analytics in the sports industry. Attendance has grown from a couple hundred in its first year to 2,200 at the event earlier this month, when panelists included Lewis, New York Times stats guru Nate Silver and Dallas Mavericks owner Mark Cuban.

Now Butler University Coach Brad Stevens has a statistician on his staff who helps determine his lineup by figuring out not just which players are best—that's easy—but what combinations of players perform best. The Miami Hurricanes use U-BALL, a software built on SAS technology, for this kind of "box score analytics," according to Mike Nemecek of SAS.

In the NBA, the Orlando Magic is in the 20th largest market but manages to rake in the seventh highest revenue. Since 2011, they've used SAS analytics to get there, studying the ticket resale market to adjust prices for season ticket holders at risk of defection. The software also generates a continuously updated heat map showing unpurchased seats so marketers can make quick decisions on offering specials to fill them.

Coleman said a team might also target a yes-or-no problem such as "Will a season ticket holder renew?" by looking at demographics and whether he's a court-side customer who's likely well-off or a nosebleed-seater who may be on a tighter budget.

(Read More: Maybe March Madness Is GOOD for Productivity)

They also might look at season ticket renewals vs. past team performance, though that relationship is not always predictable.

Lynch, who is from Buffalo, N.Y., notes that Bills fans keep filling Ralph Wilson Stadium even though the team hasn't had a winning season in nine years. Bills supporters seem to be so impervious to losing, he said, that it could prompt a bottom-line-focused front office to fail to invest enough in the payroll to turn the tide.

"Sometimes what's not significant—in this case, winning—is more surprising than what is," Coleman said.

Follow @Matt_Twomey on Twitter.