Sowing seeds

The Baltimore Sun

We approach one of the two most important events on the sports wagering calendar. College basketball's March Madness runs neck and neck with the Super Bowl for attracting wagers large and small from devotees to dilettantes.

But the NCAA men's basketball tournament, aka The Big Dance, has this particular special appeal as a predictive exercise (for fun or profit): the bracket.

The bracket works like a deck of cards in poker and bridge, or the dice and checkers in backgammon. It's something that approximates a tactile piece of apparatus over which one can puzzle and noodle with.

We have this graphic representation of a colorful, frenzied sports event that is happening all over the country in a concentrated span of days and weeks.

End result: millions of people wagering billions of dollars.

The great thing about the NCAA tournament bracket is that you can approach it any way you want -- as art or science, frivolous or serious.

Over the past few days, I've been speaking with a couple of college professors -- Allen Lynch, an economics guy from Mercer University in Georgia, and Jay Coleman, a business management teacher at the University of North Florida -- who have approached the tournament bracket with a forecasting model rooted in their respective academic disciplines.

It's called the Score Card, and it follows in the steps of another formula they concocted, the Dance Card, for predicting the at-large teams chosen on Selection Sunday.

For the Score Card, Lynch and Coleman toss four ingredients into the stew pot: a team's Rating Percentage Index; the ranking of a team's conference; whether a team won its regular-season conference; and the number of games a team won in its past 10. That yields a Score Card number.

The team with the higher number is the predicted winner of a game.

As Coleman explains it, the technique is similar to using known data to forecast outcomes in managing a business.

The professors have presented the Dance Card in an academic journal and hope to do the same with Score Card. Both models have been presented at a large annual data mining conference held in, of course, Las Vegas.

Last year, the Score Card system produced winners 79.7 percent of the time. (It was a more modest 67.2 percent the previous year.) There's a caveat, though: Lynch and Coleman picked their winners before each round.

Naturally, in the real NCAA bracket world, folks have to make all their picks before the tournament begins. So people who want to use the system are advised to simply fill out the bracket going with the team with the higher power rating each step along the way.

Is this an endorsement of Score Card? Absolutely not. When it comes to NCAA tournament brackets, it's every person for himself.

But if you're interested, go to and look for the link to the Score Card. It's updated regularly.

And noodle away.

Copyright © 2020, The Baltimore Sun, a Baltimore Sun Media Group publication | Place an Ad