Education

They picked every tournament team correctly. Next goal: all the winners

Allen Lynch, associate professor of economics and quantitative methods, was part of a three-person team that correctly predicted all 36 at-large bids for this year’s NCAA men’s basketball tournament using their Dance Card statistical model.
Allen Lynch, associate professor of economics and quantitative methods, was part of a three-person team that correctly predicted all 36 at-large bids for this year’s NCAA men’s basketball tournament using their Dance Card statistical model.

College basketball fans are fine-tuning their bracket picks for the men’s NCAA basketball tournament, hoping for a good showing. In recent days, though, a Mercer University professor helped make some tournament predictions, and they ended up being exactly right.

Allen Lynch, an associate professor of economics and quantitative methods, was part of a three-man team that correctly picked all 36 at-large bids for this year’s tourney.

Thirty-two teams get automatic bids by winning their conference tournament, and a selection committee chooses the other 36 teams based on data comparisons, observations and discussions.

Lynch, University of North Florida professor Jay Coleman and economist Mike DuMond, of Charles River Associates, have been guessing the at-large bids since 1994 using the Dance Card statistical model they created. The mathematical equation looks at key variables such as team performance data and past committee decisions, Lynch said.

“We were obviously very excited. The most gratifying part of it was that our predictions were better than those of ESPN bracketologist Joe Lunardi,” Lynch said. “Joe Lunardi is very knowledgeable in the field, so it was nice to say that our model bested his predictions this year.”

Lynch and Coleman worked together at the University of North Florida, and DuMond and Lynch became friends while in graduate school at Florida State University. The trio are “sports junkies” who share a love of data analysis, and they have collaborated on a number of sports-related research pieces.

“There are so much data associated with athletic performances and outcomes that there are virtually an unlimited number of potential data analysis projects that can be conducted within the world of sports,” Lynch said.

The Dance Card model has had a 96-percent accuracy rate over the past six years, and 2013 was the only other year with perfect picks.

Lynch, Coleman and DuMond developed a model several years ago to predict individual game winners in the tournament, but as expected, they weren’t nearly as successful with that, Lynch said.

Nonetheless, Lynch has some personal predictions as the NCAA tournament begins. He thinks Wichita State and Cincinnati could have good runs, and he wouldn’t be surprised if Michigan ended up in the Elite Eight.

While he’d love to see Florida State take it all, he has picked Villanova to repeat as the national champ.

Andrea Honaker: 478-744-4382, @TelegraphAndrea

How does it work?

To learn more about the Dance Card statistical model, visit dancecard.unf.edu.

This story was originally published March 15, 2017 at 6:00 PM with the headline "They picked every tournament team correctly. Next goal: all the winners."

Get unlimited digital access
#ReadLocal

Try 1 month for $1

CLAIM OFFER