Shannon Ryan: Applaud NCAA, which is open to more metrics in tournament selection process

The NCAA basketball tournament used a seeding process for the first time in 1978.

Four teams in each region received automatic bids as conference winners that season, and they were seeded based on their conference’s win-loss record in the five previous tournaments.

The tournament has been evolving since, continually searching for methods to create the best possible bracket.

And, thankfully, it’s not done yet.

On Friday, the NCAA will host basketball analytics experts to discuss including more metrics in the tournament selection process – something the National Association of Basketball Coaches had wisely requested.

Major League Baseball increasingly has relied on analytics in recent years. College basketball coaches can’t get enough of number-crunching when scouting opponents. The NCAA selection committee should also use the available data – and there’s more of it than ever before – in the most intelligent way.

The goal Friday at the exploratory meeting is simple, said Kevin Pauga, who runs the increasingly popular metrics system on KPI.com.

“How can we make things better?” said Pauga, who also serves as assistant athletic director at Michigan State. “Let’s help get more information that the committee can look at but look at it in an efficient way. Really, what can come out of this meeting is there are a lot of people that have all sorts of levels of understanding from all sorts of backgrounds and can provide knowledge (to decide) is a composite metric the way to go.”

An aggregate metric could be a part of the selection process as early as next season, according to an article on NCAA.com. The NCAA will tap into the minds of metrics gurus like Pauga, Ken Pomeroy and Jeff Sagarin along with ESPN’s Ben Alamar and CBSSports.com bracket predictor Jerry Palm.

The harshest criticism of the current model is typically aimed at the reliance on the Ratings Percentage Index. The RPI measures aspects like winning percentage, opponents’ winning percentage and the winning percentage of opponents’ opponents to assess teams and rank them into weighted categories, such as top 50 and top 100.

This system isn’t nuanced enough given the availability of formulas that go deeper.

Many of the more complex systems weigh aspects like opponents’ records and place more value on road victories than the RPI does, for example. Combining them for a composite to help guide the 10-person selection committee will help make the process more clear-cut.

“There’s a pretty significant difference between beating 49 at home and beating 51 on the road,” Pauga said. “The win over 49 at home is in that top 50. And that win over 51 on the road is not.”

Palm said he hopes the NCAA does not pick just one system but uses a composite with transparent formulas. Each system can produce an outlier, which is another reason why it’s important for the NCAA to go with a composite model.

There should be caution to avoid a lifeless computer model deciding the field of 68. The hope is to keep the fun and drama of Selection Sunday alive and not turn this into the Bowl Championship Series. A composite should be a guide, not a directive. And it sounds like the NCAA and the metrics experts are in agreement on that.

“There are things computers can’t measure and every computer system has anomalies,” Palm said. “I think it would be helpful to have something a little more sophisticated than the RPI. I don’t think anyone wants the subjective element to go away. I think what they’re talking about is a more subtle change than people might be led to believe.”

While it won’t be a major shake-up, it will make the selection process stronger. The system could go into place as soon next season, according to the NCAA.com article.

“There’s not a composite metric that’s going to say a committee missed on 15 teams,” Pauga said. “We’re talking about giving more information to the committee and giving an explanation with that information to tell a story.”

Credit the NCAA for adapting and letting the selection process evolve.