ADVERTISEMENT

WAC to use KenPom to seed their tourney

glassmanJ

2500+
Jan 26, 2007
2,559
1,172
26
The Western Athletic Conference is changing how it seeds its postseason basketball tournaments by introducing an intrepid, evolutionary concept to the sport. The 2022-23 season will see the WAC's men's and women's basketball standings be determined not just by conference wins and losses, but also an algorithm that rewards and/or punishes based off performance against all teams they face in the regular season.

In what's believed to be a first in the history of college sports: a conference will seed its postseason tournament based on advanced analytics.

This means that, in the WAC, it's possible a team could finish with the second-best league record but end up seeded, say, fourth or fifth. Conversely, a team with the eighth-best record but with a more impressive résumé of wins could be rewarded for its degree of difficulty and earn a higher seed than its place in the traditional conference standings.

WAC commissioner Brian Thornton and associate commissioner Drew Speraw, both former basketball coaches, are the architects of the idea. They commissioned Ken Pomeroy -- proprietor of KenPom.com -- to concoct a reliable formula that would accurately reflect the achievements of WAC teams from early November until right before the WAC basketball tournaments begin in March. Thornton first went to Pomeroy with this concept when the conference was at risk of games being canceled due to COVID-19. That led to Thornton and Speraw mulling how such a method could be used for competitive balance and meritocratic representation beyond mere wins and losses in league play. College basketball has embraced advanced analytics for over a decade at this point. This is an extension of that philosophy.




"The goal (is) ultimately to protect the highest résumé with the highest seeds," Thornton told CBS Sports. "As we were coming up with this strategic plan from a basketball standpoint -- which was one of my big tasks when I came here -- nonconference scheduling always came up. … And ultimately, it becomes very hard to penalize people for what you're able to do from a nonconference scheduling standpoint."

The league is referring to this sorting metric as the WAC Résumé Seeding System. If a WAC team plays a tough opponent and wins, it will receive more statistical credit than if it plays a weaker opponent (and also wins). Conversely, if a WAC team plays a tough opponent and loses, it will not be dinged as harshly as if it plays a significantly weaker opponent and loses.

Wins and losses is a zero-sum result. But beating a top-50 team is a much bigger accomplishment than beating a team that rates somewhere in the 300s. The WAC's seeding initiative will reflect that reality. The algorithm will account for where games are played (at home, on the road, on a neutral court) and will be based on the NCAA's NET ranking, which is the selection committees' primary sorting tool for seeding and selecting the NCAA's basketball tournaments.





"Essentially, what we came up with was a way to utilize the NET in order to provide the reward/penalties for a particular game," Speraw told CBS Sports. "It's weighted depending on where the game is at, just as the committee would look at it, and so it's a system where, essentially, we're trying to promote Quad 1, Quad 2 games. And so when you look at it, not every game is worth the same inside Quad 1, Quad 2, just as the committee would look at that. Obviously, a top-10 win is not the same as a top-30 win. Similar, but not the same."

"The idea is to seed the conference tournament based on your full season play, not just your conference standings," Pomeroy told CBS Sports.

The WAC is coming off one of its strongest seasons in recent memory. The league finished 15th out of 32 leagues in KenPom's rankings, its highest finish since 2013. Thornton noted that New Mexico State, which was fairly awarded a 12-seed in the 2022 NCAA tourney, won the WAC Tournament and set the conference up for its best chance at Big Dance success. (New Mexico State finished 80th at KenPom last season, the highest in the conference.) The Aggies flew cross-country to Buffalo and upset fifth-seeded UConn 70-63 in the first round. Had Grand Canyon, Seattle or Sam Houston State won the WAC auto bid, a 12-seed would not have been on the table, and thus the WAC would have been much less likely to win a game and earn treasured NCAA Tournament "units" (which is industry lingo for monetary payouts).




"At the end of the day, one of them gets a 12-seed. If the other one is a representative, they probably get a 15-seed," Thornton said. "What are we trying to create in this situation?"

Another reason for introducing the concept is unbalanced league scheduling. There is no WAC round-robin play because the conference has 14 teams on an 18-game schedule. Thornton said there was pushback from some inside the league, due to the unorthodox nature of the experiment, but ultimately way more people than not backed it. And in a nod to traditionalism, the WAC will still rely purely on conference records for entrance to its men's and women's basketball tournaments. The 12-best league records will qualify for the postseason. Once the 12 are in each bracket, then they will be seeded according to the Résumé Seeding System.

This formula-based approach can also be an aid for the selection committees as they determine how to slot WAC teams for March Madness.




"A win is still a win, it's still positive, it's a plus in your win column, and a loss is still a loss, and it's a negative, just if you lose to a better team, it doesn't hurt you nearly as bad," Speraw said. "Whereas if you lose a Quad 3 or a Quad 4 game, it can be detrimental to your résumé."

As for how the algorithm works, Pomeroy said each game is calibrated based on the chances of the 150th-best team in the country winning in that scenario. For example, if the 150th-best team had to face the top-ranked team, Pomeroy's algorithm, which is based on mountain ranges of historical data, suggests that Team 150 would have approximately a 2% chance of winning that game on the road, a 4% chance of winning on a neutral court and a 7% chance of winning at home.

Why 150?

"I wanted it to be a little worse than their best teams would be so it would give some incentive to the best teams to schedule better," Pomeroy said.




Every game has a valuation of 1.0, meaning that every game in a season will carry weight in the Résumé Seeding System and each result will add (with a win) or deduct (with a loss) from a team's season-long total.

The concept is rudimentary; it's merely introducing win/loss merit to bring more nuance to the WAC's standings ledger. The concept rewards teams for scheduling well (and winning) and dings them for scheduling poorly (and losing). Since it's harder to beat a top-50 on the road than it is to beat a team ranked 200th at home, those achievements should be weighted more appropriately.

"If you play Gonzaga and you beat them, that's going to be a huge reward, low penalty. If you play against [a bottom-tier team], that's always going to be a very, very low reward. But if you lose that game, there's going to be a huge penalty," Thornton said.
 
The NCAA has infamously not revealed the coding or formula behind how the NET is built, but the WAC and Pomeroy are able to work around this by using Pomeroy's own algorithm, applying it to his credible rankings system and logging those inputs in accordance with the NET rankings. Since the NCAA publicly updates the NET daily from early December through Selection Sunday, WAC teams will know every day where they stand in their seeding scenarios for the conference tournaments.

Values of each game will change as teams' performances change and will lock at the conclusion of the regular season.

"It can be transparent because we can see exactly what each point value is worth," Speraw said. "Whereas a win versus whoever -- you may gain or lose points in the NET based upon a game in February that you play against a bad opponent -- you may be winning, you go down two spots (in the NET). In this in this scenario, winning is still winning. It still provides you a positive momentum forward, regardless of the opponent."




Based on Pomeroy's algorithm, a win over the top-ranked team would be awarded thusly:

Road win: +.978 points
Neutral win: +.960
Home win: + 930


A loss against the No. 1 team doesn't carry nearly as much negative weight as a loss to a team in the 200s or 300s. A loss to No. 1 by a WAC team equates to:

Road loss: -.022 points
Neutral loss: -.040
Home loss: -070


But a home loss to a team ranked 300? That would be -.862, a much heavier hit. All of these pluses and minuses will be tallied day by day and reflected in the league's official standings page. The higher the number, the better the seed.




"You obviously want people to schedule good games, you want people to play good games and win good games, but how do you incentivize them for doing so?" Thornton said. "Our thought was, we can positively incentivize them, because now we're taking into consideration what your seeding is in the tournament based on the quality of games that you play and the quality of games that you win across the board. To me, that's a huge difference."

Had this been in place last season, the WAC standings would have been different, though not drastically. The WAC had a three-way tie for first place in the league standings, but retroactively applying the formula would have changed it to look like this:

screen-shot-2022-07-10-at-7-03-50-pm.png

Courtesy of the WAC
Thornton said the WAC also updated its conference tournament format; previously the top seeds earned auto-byes into the semifinals. Now the top four teams will earn byes into the quarterfinals only.




This idea is the latest bold push from Thornton and Speraw. The duo is also the driving force behind the to-be-determined February nonconference scheduling concept that could boost college basketball's viability heading into the NCAA Tournament. (Thornton said plenty of leagues are still on board and ready to go, but there is still some discussion being had with others before this ambitious idea can take its next significant step.)

"The 'wildcard week' concept is something that was kind of part of our strategic plan, conversations, and would work well with this," Thornton said. "If that goes down, that would work well because it gives teams two more good games if you're good. So that would work hand-in-hand in that piece and kind of cutting down on the amount of games that you actually have to physically schedule yourself."

This idea is sure to capture the attention of other traditional single-bid leagues in college basketball, and if it proves fruitful, the WAC might not be the only conference using this approach a few years from now.
 
ADVERTISEMENT
ADVERTISEMENT