Luke DeCock

NCAA basketball tournament seeding changes will have impact

Duke players celebrate their 68-63 victory over Wisconsin in the NCAA Final Four college basketball tournament championship game April 6, 2015, in Indianapolis.
Duke players celebrate their 68-63 victory over Wisconsin in the NCAA Final Four college basketball tournament championship game April 6, 2015, in Indianapolis. AP

Typically, there’s very little in the world of college basketball that escapes the notice of the ACC, but the NCAA’s announcement of some changes and tweaks to the way teams are seeded for the NCAA tournament came Monday, when the ACC was very much immersed in football at Pinehurst.

These changes could potentially have an impact on ACC teams, so it’s worth taking a step back for a slightly deeper look. And what may be the biggest tweak of all wasn’t even announced. It remains unofficial, the Division I men’s basketball committee’s use of advanced metrics in addition to the RPI to evaluate teams.

First, the official changes: Instead of sending the last four at-large teams voted into the field to the First Four, the four lowest-seeded at-large teams will go to Dayton instead. And the best of the No. 2 seeds will be bracketed with less concern for geography to avoid a top-heavy region with the No. 1 overall seed, although conference affiliation will still matter.

For teams near the back of the field, as N.C. State has been two of the past four seasons, this means the true “last four in” will end up in Dayton. (Texas and UCLA avoided Dayton based on when they were voted into the field but should have been there given their eventual seeding.) And for teams at the top of the field, as Duke has been in seven of the past eight seasons, the best team will not have to face the most powerful No. 2 seed in the regional final.

“When it was all over, (NCAA executive) Dan Gavigan and I were driving to Dayton, and during the car ride we had a sort of post-mortem on the process,” said outgoing committee chairman Scott Barnes, the new athletic director at Pittsburgh. “This was certainly an area where we said, ‘Why do we do it this way?’ It’s a good step in the right direction.”

These commonsense changes aren’t nearly as significant, though, as the committee’s increasing willingness to move away from the simple but inevitably flawed RPI to evaluate teams. Barnes acknowledged on Selection Sunday that the committee looked beyond RPI in deciding to leave Temple and Colorado State out of the field, but this week confirmed the use of such metrics as Ken Pomeroy’s and Jeff Sagarin’s efficiency ratings was even more extensive than Barnes described that day in March.

“More this year than any prior year, we looked at other systems when there were gaps and inconsistencies in the RPI,” Barnes said. “We talked more about it, how the RPI doesn’t tell the whole story.”

This is a change from past practice, when the committee hewed so closely to its procedures and principles that, combined with its slavish devotion to RPI, it was possible for the increasingly educated consumer to mimic the process from outside the committee room. The average score on Bracket Matrix, a website which tracks bracketologists, went up each year from 2010 through 2014 as predictions, collectively, improved.

That changed in 2015, when teams like Texas and UCLA benefited from strong KenPom ratings to get into the field and Oklahoma was seeded ahead of Maryland. In each case, use of the RPI-based resume to compare teams would have suggested different results.

With Michigan State athletic director Mark Hollis taking over as committee chairman in 2017, this trend does not figure to abate soon. (Oklahoma’s Joe Castiglione is the chairman this season.) Hollis has a history of embracing innovation – the Carrier Classic was his idea, among others – and one of the new metrics, KPI, was invented by a Michigan State basketball staffer, Kevin Pauga and introduced to the committee by Hollis.

The new numbers produced some unexpected results on Selection Sunday, but they’re clearly here to stay.

DeCock: ldecock@newsobserver.com, @LukeDeCock, 919-829-8947

Related stories from Charlotte Observer

  Comments