While the Ospreys may not have made it to the NCAA Division 1 Men's Basketball Tournament this year, the spirit of the Big Dance lives on through the University of North Florida's very own Dance Doctor.

Associate Provost Jay Coleman’s analytical formula, known as the Dance Card, has once again accurately predicted all 36 teams receiving at-large bids to this year’s highly esteemed and sought-after tournament.

Coleman received his doctorate in Industrial Management from Clemson University in 1988. Shortly thereafter, he joined the faculty at the Coggin College of Business at UNF. Nearly 30 years later, the basketball-ranking formula he co-authored has for the second time done what some might consider close to impossible.

To be clear, Coleman’s formula for predicting which teams will get the blessing from the NCAA tournament selection committee is not the only one out there. However, the Dance Card is one of the most detailed and surprisingly accurate statistical models to date, correctly predicting 209 of 218 at-large bids in the past six years.

Just how exactly do they do it, you might be wondering. The Dance Card, per Coleman, does use one of the most commonly referenced metrics, the Ratings Percentage Index (RPI), but they also implement their own unique set of quantities in the formula.

“It’s how you are doing against various groups of teams in the RPI,” Coleman said. “So, how man wins do you have against the top-25, what’s your record against the second-25, if you will – teams ranked 26 to 50 – your record against teams ranked 51 to 100, how do you do on the road, whether you have a losing record in the conference. That’s about it. It’s really not fancy.”

Fancy, maybe not, but the Dance Card was once used to prove the existence of an intrinsic historical bias among the ten-member selection committee. Coleman says the Dance Card, during various time periods, found statistical evidence of the committee showing favoritism towards the selection and seeding of major D-1 teams, as well as bias towards teams with some sort of committee representation.

“We know they are familiar with our work,” Coleman says. “We also know through indirect channels that in 2012 the committee itself, during their deliberations, was made aware of our line of research and that there was some evidence, statistically anyway, of bias. And that year they were very unbiased and, quite frankly, every year since they’ve been rather unbiased or at least there isn’t strong evidence of it.”

Coleman says he and his associates would like to think that their work with the Dance Card contributed to the purge of bias seen in committee selection, but he admits this is only speculation on their part.

One of the more interesting factors Coleman said they’ve found in the Dance Card metrics is that wins against the top-25 RPI ranked teams are very important, while losses against the top-25, not so much. This being something mid-major and minor D-1 conference teams, should take note of in seasons to come.

“So, the satisfying element of that is, go play the top-25,” Coleman proclaimed. “Take the risk of, you know, going out and trying to play against the best teams. If you win, you get brownie points. If you lose, no biggie.”

According to Coleman, more “advanced metrics” could soon be introduced by the selection committee. A “hue and cry,” Coleman says, to get away from the RPI has arisen. In response, the selection committee chairman Mark Hollis met with various experts on basketball metrics in January to discuss the implementation of a better way to determine team strengths. If changes are made, the current Dance Card formula might also have to adapt, Coleman says.

“The fact that we were perfect this year, not only not using those advanced metrics and using an old version of the RPI, suggests that if the advanced metrics are going to be used, they haven’t kicked in yet,” Coleman said. “And to be fair, there was no promise that they would kick in this year, but there is a promise, of sorts, that they are going to kick in next year.”

Unlike the results-oriented RPI, the Dance Card is not currently designed to predict how a team will be seeded once it has been selected by the committee. Per Coleman, the idea that the selection committee weighs similar factors in the selection process and the seeding of teams may actually be incorrect.

The research Coleman and his associates published in 2010 showed incongruent patterns in the two processes, as well as statistical evidence of bias in how teams were seeded.

“Now, for a variety of reasons, our model doesn’t do seeds and we’ve not continued that or updated that line of research on the seed side,” Coleman says. “So, I don’t know if that bias has waned.”

Next year looks to be a defining one for the Dance Card, as the potential for new, more advanced metrics could impact the accuracy of predictions. The Dance Card, after all, is directly correlated to the consistency of the committee, per Coleman.

Coleman believes one of the advanced quantities the committee may be looking to adopt is a tempo-based, predictive metric, made famous by Ken Pomeroy, who was one of the experts Hollis met with earlier this year.

Hollis hinted after that meeting that there will most likely be analytical updates to both the selection and seeding process, calling for a "fresh look." Whether one set of metrics is better at determining the seeding and one at the selection of teams remains to be determined.

There doesn’t seem to be much concern over the future of the Dance Card though. Coleman says, after looking at nearly 30 other more advanced metrics, there was no tangible proof that the Dance Card was any less accurate in predicting teams.

“So, even though those things have been around for a long time, there is no real evidence that they are superseding, if you will, the old-school ways of doing things,” said Coleman.