Either the RPI matters or it doesn’t. The NCAA needs to decide.

There were plenty of questionable factors in the 2016 bracket that made it look like Charles Barkley might have had a hand in creating the bracket not just analyzing it during the ill-conceived two-hour selection show. But how the committee did or didn’t use the RPI is certainly one of them.

For over two decades, the Ratings Percentage Index has been part of the NCAA’s selection process. The computer formula was developed as a way to compare schools in different conferences, who don’t play each other.

It’s not perfect. No system is. There are too many nonquantifiable factors that go into winning and losing. But for better or worse, it’s the system the NCAA uses, or at least claims to use, to aid in filling and seeding its men’s and women’s NCAA Tournament fields.

While there’s been plenty of years where a team with a better RPI gets an at-large bid over a team with a worse one, the numbers are usually close.

But St. Bonaventure was left out of the field despite being ranked 30, at least 15 spots higher than Wichita State (47), Michigan (56), Tulsa (58), Vanderbilt (63) and Syracuse (72).

Schools, especially outside of the Power 5, study the system and create nonconference schedules designed to maximize their chances for creating a strong RPI.

When St. Bonaventure athletic director Tim Kenney was the associate A.D. at UMass, he poured over the RPI when trying to create a schedule. RPI was part of the reason UMass was a No. 6 seed in 2014.

He’s certainly doing the same thing now with the Bonnies. When it works they have to get credit for it. If the NCAA says the RPI matters, it has to.

When Miss America candidates arrive at … wherever they do that silly event nowadays … they know they need to excel at the talent portion, the interview and swimsuit competition. The judge can’t suddenly tell them ability to speak Portuguese or solve a Rubik’s Cube is now part of the process. The goal line has to stay the goal line.

But selection committee chairman Joe Castiglione (the Oklahoma athletic director, not the Red Sox radio guy with the same name) cited advanced metrics as part of the reason Syracuse got in.

By advanced metrics, he likely means Ken Pomeroy’s computer rankings. The stats at KenPom.com are deservedly the gold standard among ranking system for college basketball coaches, fans and scouts who like analytics as a measuring tool. Most people believe they’re more accurate than the RPI.

Accuracy is good. The more accuracy the better to be sure. If the NCAA wants to use KenPom.com or ESPN’s BPI or Basketball State’s metrics, they should. But they should say so before the season starts. Schools will build their schedules accordingly. Students who take SAT prep courses can’t be expected to suddenly take the ACT.

The RPI was used tangentially.

Syracuse got credit for having more top 50 wins including one over the Bonnies. The top 50 in question is not the Billboard record chart. It’s the RPI. In other words St. Bonaventure’s status as a top 50 team helped Syracuse, but didn’t help the Bonnies.

Losing to Syracuse counts as a non-top 50 loss for SBU. Either having a good RPI matter or it doesn’t. It can’t go both ways.

The NCAA should be clearer how much computer rankings are utilized. College hockey uses a ranking system almost exclusively. There aren’t many surprises on selection day. Nobody complains of bias. You are what your Pairwise ranking says you are.

Nobody is advocating for that, but there needs to be a computer component. It’s the best way to keep the selection process from rewarding programs for their brand instead of their actual resume.

I’d argue for following:

Pick a computer ranking system. One of the ones that exist already, a new one designed for this or an average of several computer systems.

Take the top 35 teams based on those rankings and all of those teams get into the field.

Using KenPom.com as the system. This year eight of the top 35 would already have automatic bids and two (Louisville and SMU) are ineligible. The other 25 teams in the top 35 automatically get at-large bids.

Add that to the 21 teams ranked outside the top 35, who won their tournaments and got automatic bids, and you get 58 teams.

The committee would then choose the other 10 teams and then seed the entire 68. It would have more leeway to consider other circumstances when picking the last 10 and seeding the field.

Matt Vautour can be reached at mvautour@gazettenet.com. Get UMass coverage delivered in your Facebook news feed at www.facebook.com/GazetteUMassCoverage