Returning Possession Minutes and Mid-Majors

Why hasn’t someone hired Steve Prohm away from Murray State yet? That’s the question I’ve come to lately.

Continuing my research into returning possession minutes (RPM) and mid-majors, I pulled down the data for the teams in the leagues ranked 12th (Horizon League) through 24th (NEC) last season on KenPom for the past four seasons. There’s a total of 550 team seasons included now in my data set.

When I ran a regression to determine if RPM was a significant predictor of team success the answer was certainly, yes. I used RPM and the previous season’s pythagorean rating to “predict” next season’s pythagorean rating. The regression ultimately has an r^2 of 0.522 and all the pieces are significant. Of course that r^2 means there is a lot of wiggle room. Part of this is because no prediction system in college basketball is going to be perfect. Players get hurt, coaches move around, teams unexpectedly find the perfect rhythm or lose it. Dan Hanner has done some awesome work around using a lineup model to get better predictions, but RPM and last year’s pythagorean rating at least gives us somewhere to start.

It also leads to some big outliers. Teams that performed poorly even though they brought back a number of players are of course the biggest conundrums. One of the biggest head scratchers from the past four seasons  was front and center during Ivy League play. The Penn Quakers brought back basically their entire rotation and mightily underachieved, finishing 8-20 (5-9 in Ivy League) and ranked 266th in KenPom. That performance has a lot of people wondering why Jerome Allen is still employed by the Quakers. It’s understandable. Considering the RPM and 247th ranking in pythagorean record the season before, the expectation was the team would improve, not drop to the bottom of the standings saved only by Cornell’s gross incompetence. Here are the ten teams that performed the worst against their expectations during the past four seasons:

My (albeit crude) model expected Penn to be competing for second place in the Ivy League. What a terrible assumption! They finished seventh. But this model isn’t the only projection that thinks Penn woefully underachieved last season. Nope. The Ivy League preseason poll picked the Quakers second as well! Three of these teams (North Texas in 2013, Chicago St. in 2011, and Colgate in 2012) were a coach’s first season on the bench. Obviously there was an adjustment period. Injuries (Kyle Vinales) and suspensions (Matt Hunter) help explain CCSU’s underperformance last season. But now I’m basically making excuses.

Enough of the negativity though. What teams have outperformed expectations during this same period? Glad you asked.

Steve Prohm come on down! The Murray St. head coach has been leading the Racers for three seasons and two of them appear on this chart. The 2012 team is one you might’ve heard about. They started the season 23-0 and finished 31-2 after losing to Marquette in the NCAA tournament. This table also shows why I thought Mike Brennan deserved serious consideration for Coach of the Year. The first-year head coach took American, which was picked ninth out of 10 teams in the Patriot League’s preseason poll, to 20 wins and an appearance in the Big Dance. Stony Brook also appears on this list. While the 2013 Seawolves mightily overachieved, finishing 25-8 and 52nd in KenPom, they were eliminated by Albany in the semifinals of the America East tournament. The loss of Tommy Brenton and the new hand-checking rules decimated SBU this season and ceded the top spot in the conference to Vermont.

I’ll be continuing to refine this model during the offseason. Right now though it’s pretty obvious that bringing talent back can help a mid-major team explode. We can dig deeper into the outliers (including the teams that were projected to be the best in the data set and just didn’t quite get there) and hopefully learn more about how to predict just what’s going to happen in 2014-15.