The Ratings Percentage Index has existed since 1981 as one of men’s college basketball’s main influences and organizing standards.
The sport’s prominent question, often asked instead of “What’s a team’s record?” is “What’s the RPI?”
Perhaps it’s time for that to change. Former Wichita State athletic director Jim Schaus is chairing an NCAA sub-committee working on several NCAA Tournament-related subjects, including how the RPI is used to select at-large teams.
“We’re going to really reevaluate the RPI as such a prominent metric and take a look at something that would be more of a composite metric,” said Schaus, athletic director at Ohio University.
Never miss a local story.
On Friday, college basketball analytics gurus will take part in an NCAA-hosted meeting in Indianapolis that aims to improve the tournament selection process by looking at composite measurements. The RPI is under increased scrutiny for its flaws in recent seasons. While committee members say the RPI is one of many factors considered when evaluating at-large teams, the ranking’s weight is likely to further decrease.
In its place could rise a blend of measurements, a composite that will take into consideration results-oriented stats such as the RPI with ones with more predictive elements, such as ones developed by Ken Pomeroy and Jeff Sagarin. Pomeroy’s rankings factor in margin of victory and measures a team’s success offensively and defensively as a way to rank teams, who can play schedules that vary significantly, and predict success. ESPN’s College Basketball Power Index measures team strength to predict performance. Its formula accounts for opponent strength, pace of play, site, travel distance, rest and altitude to simulate results.
The RPI uses winning percentages of teams (25 percent), their opponents 50 percent) and their opponents’ opponents (25 percent) to produce a ranking. It is now seen as too simple when compared to other metrics. Pomeroy, Jeff Sagarin, Kevin Pauga of KPI Sports, and others, will attend Friday’s meeting. So will Schaus, representing the men’s basketball committee, and NCAA senior vice president of basketball Dan Gavitt.
“I think it makes a lot of sense to take a look at it,” Schaus said. “This is just one of many things we’re looking at. The final answer is yet to be determined as to what we’ll learn from this.”
Schaus said the sub-committee ran test cases and found its experimental composite metrics predicted results accurately.
“We went back and looked at some of the potential combinations that may be considered and then we looked at how teams were selected and seeded,” he said. “Then we applied where they would have been with the metrics that were available at the time and we saw how they did in the tournament. It was a prediction that, ‘Yes, they should have been in the tournament’ and, ‘Yes, they did well in the tournament.’ ”
Changes will not affect how teams are selected and seeded for the 2017 NCAA Tournament. The change would likely not be instituted until 2018.
“Our cursory research has shown that a composite metric may be better than just the RPI in its overall predictions,” Schaus said. “If that’s true, then it’s probably a better standard to use to identify where teams are as a top-50, top-100 win.”
Even when other metrics are used — and a strong Pomeroy ranking helped Wichita State’s case in 2016 — the RPI remains important as a way of organizing teams by wins against top-50 and top-100 opponents. Using a composite to produce the “team sheets” that now rely on RPI may provide a more accurate starting point to evaluate teams.
“The impetus for the stats summit comes after years and years of haranguing the NCAA due to its stubborn default reliance on the RPI, an uncomplicated and easily manipulable metric,” Matt Norlander of CBSSports.com wrote. “The team sheets, which are available for public consumption, are built through RPI numbers. The point is, if it was a composite ranking that constructed the team sheets, the data therein would be more objectively accurate.”
Missouri Valley Conference commissioner Doug Elgin, along with coaches such as Bob Huggins, Mark Few and Mark Turgeon, worked on a National Association of Basketball Coaches committee on tournament selection, seeding and bracketing. Coaches asked for more transparancy and more guidelines as to how teams are selected and seeded, Elgin said.
“This is another step in which the basketball committee is embracing outside opinions and outside information in an effort to get it right,” Elgin said. “It’s going to produce a better way of understanding what the (selection) committee actually evaluates.”
Elgin, who previously served on the tournament selection committee, said it has long used other metrics.
“The RPI has always been viewed as just another tool, another source of information,” he said. “This may structure the decisions a bit more. There’s not going to be an absolute mathematical formula that dictates who gets in. This will never take the place of the subjective decisions that have to be made.”