We are all smiles here in Madison because our lovable and woke-as-hell Badgers toppled defending champion and top overall seed Villanova in the NCAA men’s basketball tournament yesterday. We are even getting as chesty as our Midwestern sensibilities will allow.
But as the college basketball press pieces through the wreckage of the nation’s brackets, it seems to have stumbled on a worthwhile question: Why was Wisconsin playing as an 8 seed at all? The Badgers won 25 games this year and finished second in the Big Ten. Teams that did worse than the Badgers and lost to them multiple times this year entered the bracket as higher seeds. It did not make sense.
The answer appears to be that the Badgers racked up those 25 wins against the wrong teams.
The NCAA has used something called the Ratings Percentage Index to select and seed the at-large teams for its tournaments since the 1980s. The RPI ranks the 351 men’s Division I college basketball teams on the logic of you are who you play. If you play good teams and beat some of them, you rank higher than if you play terrible teams and beat most of them. The Badgers’ played teams like Central Arkansas, Prairie View A&M and Chicago State early in the season. Its preseason strength of schedule ranked in the 300s, said ESPN. The RPI has widely acknowledged problems. Primarily, it is biased toward teams from major conferences, because those teams play a tougher schedule. This makes it more difficult for smaller programs to get invitations to the tournament. But that is the so-called objective measure the NCAA uses. Most bracket predictions before Selection Sunday had the Badgers in the 5 or 6-seed range. But ESPN said Wisconsin’s final RPI was 36, which suggests it was closer to being a nine than a seven.
Ridiculous, said those who watch college hoops obsessively (not me).
But what exactly makes it ridiculous? Is it the eye test? Is it the four seniors the Badgers start? Is it the tournament pedigree of the program? Is it taking the start of the season more seriously than the slump over the last three weeks of the season? Is it wanting activists Nigel Hayes and Bronson Koenig in the postseason as part of sports reporting’s newfound liberal ethos? (By the way aren’t we missing the hot conspiracy theory for the Badgers’ seeding? The committee did not want Nigel Hayes, who is suing the NCAA over amateurism, to find his way to the Sweet Sixteen.)
The logical call is for better metrics to be used in building a tournament bracket, and the NCAA appears to be working on that. At a meeting this winter, the debate seemed to be between using results-based metrics like RPI or predictive metrics like some of the other rating systems, or else figuring out how to combine the two systems. Understand though that the choice of metrics is a moral decision, not a statistical one. Should NCAA tournament bids be a reward for a good regular season or should they be handed out for reasons combining results and other factors? And if you say other factors should be included, what factors? Now we are discussing values. What is the good and right way of being a college basketball team?
Whatever rating system, or combination of rating systems, the NCAA selection committee uses in creating its postseason tournament has to allow it to look at 351 teams at once. There may be many ways to do this and better systems are always a good goal. Every decision, however, moves away from RPI’s biases and toward a different set of problems. If we are going to value point differentials or efficiency ratings, we are operationalizing assumptions about how basketball works at this level. We are defining the primary ways we can judge the relative quality of teams who will never share a floor (sharing a floor being the best way to compare teams). No ratings system is a value-free rendering of truth. They are numerical outcome of a set of hypotheses about what makes a team good or not. Better metrics can only get us to a perfect bracket when we agree on the perfect way to play basketball. Maybe the Golden State Warriors have discovered that. Or maybe – and this is what I suspect – there are lots of ways to play winning basketball.
What does it mean to get the bracket right anyway?
Does it mean giving the best teams the clearest possible path to the Final Four? Is an upset actually a failure by the committee? That cannot be right. I am old enough to remember Thursday night when everyone was complaining that the first day of the tournament was deathly boring had been rendered boring by the lack of upsets. Should the committee use the bracket to engineer drama in the early rounds? That sort of necessitates putting teams like Wisconsin on the 7 or 8-line, so they can get early cracks at the elite teams? That hardly seems fair to the Gonzagas and Villanovas of the world, though not crying for Goliath is the ethos of the tournament.
There are no perfect systems. Perhaps we may just take joy in the tinkering.
michaelmirer.wordpress.com 