Retention Knowledge
Pundits claim that roster consistency wins basketball games. Are they wrong?
Last season’s preseason top 25 was littered with teams bolstered by the analysis that they return most, if not all, of the starters from previous teams. While some of these were solid choices, like Purdue and Marquette teams returning the core of top ten squads from the prior year, others didn’t quite pan out.
Michigan State rocketed up the AP rankings to the #4 spot, even netting a single vote for #1, after a Sweet Sixteen run excited media voters. The media should’ve paid a little more attention to that number, as the Spartans finished 2023-2024 sixteenth in KenPom and fell in the second round of the NCAA Tournament. The voters must’ve been unaware that the solid core they were banking on finished 26th the year prior.
Florida Atlantic received an even more meteoric rise from their shock Final Four run, jumping into the top ten as tourney stars shocked the nation by remaining in Boca Ration despite numerous lucrative offers elsewhere. The players might’ve been better served taking those up, as the Owls regressed back to 47th in KenPom during an erratic campaign. AP voters must not have noticed that FAU would’ve finished the previous season 24th in BartTorvik without a controversial call on the last play of their opening round bout.
Finally, a Texas A&M team that finished the 2022-2023 regular season on a hot streak carried that momentum into a preseason #15 ranking in the AP Poll, despite having finishing 33rd nationally with nearly the exact same team. The Aggies finished 2023-2024 at the #35 spot; The dummy writing this article had them 13th.
While returning production isn’t inherently bad, it serves mainly as a strong indicator that a team is likely to finish near where they did the year prior, with a strong improvement only slightly more likely than a sharp decline. Correlation between final rank in KenPom’s analytical standings and the court time a team returns reached only about .2, below the threshold for significance.
That is, until you get to the extremes. While there were outliers, the 10 teams with the highest returning minutes rose an average of 27 spots in the KenPom rankings between 2022 and 2023, while the 10 squads with the least returning minutes dropped a precipitous 61 ranks in KenPom’s rankings.
This trend continued last season, as the returners gained an average of 44 spots, while the greenest squads fell 39.
While there are dangerous enough outliers to avoid drawing major conclusions without analyzing the individual cases, here are the candidates for these rises and falls this season:
This year will be a fascinating test for the theory of extremes. Louisville, DePaul, and even lowly IUPUI (now going by IU Indy, to the disappointment of Ooey-Pooey fans everywhere) have significantly better coaches and rosters. California has reloaded for Mark Madsen’s second season with a solid transfer class of power conference pieces. UTSA brought in an incredibly talented but potentially volatile cast of transfers above their weight class. Kentucky hired the architect of America’s most fun offense last season and gave him everyone from Lamont Butler, who hit the buzzer beater that sent SDSU to the 2023 title game, to Ansley Almonor, the stretch five that broke Zach Edey in FDU’s 16-1 upset. While UMES is a complete MESS, they don’t exactly have much further to fall in the rankings (Here is where the data science nerds can rightly yell at me for using ordinal data.) Something has to give; either extreme roster turnover isn’t negative every year, or a few of these teams will have truly shocking seasons.
Returning to the teams expected to compete for championships next season, only Gonzaga and Houston find themselves amongst the extremes of roster retention. Houston clocks in at #12, with the one player lost being All-American guard Jamal Shead. While this has thrown some off the scent, I won’t be fooled, as the same concerns were raised about the departure of star guard Marcus Sasser the year prior. The AP Preseason Top 25 hasn’t been released yet, but CBS Sports’ Gary Parrish continues to update his “Top 25 and 1” rankings, which give some insight into who the capital-J-Journalists are bullish on entering the year. Of these teams, a few strike me as dangerous cases of overvaluing returning minutes.
I would’ve thought everyone learned their lesson on Texas A&M last season, but the Aggies sit proudly at the #11 spot, with Parrish gleefully noting that A&M returns four of their top five starters from last season. Why 80% of a team that finished 35th last season and a transfer from SMU equals the #11 rank in the country is a mystery to me, even as someone who fell for Texas A&M’s charade last season. Providence also snuck into the rankings at the “And 1” spot of Parrish’s top 25, and I’m slightly baffled as to why. The Friars finished 59th in KenPom last season, and yet Parrish justifies his selection with the claim that “Providence [returns] four of [their] top seven scorers,” a mealy-mouthed way of saying that they lost two of their three best players and kept the rest. While Bryce Hopkins’ injury in January derailed things somewhat last season, they were around 35th with him, the 13th pick in the NBA Draft, and an All-Big East center. Now, it’s just him.
While extreme cases of player retention or loss can be some indication of a program’s health, the actual value of returning minutes in predicting team success in the transfer portal era is very limited. While this may seem like an obvious statement, virtually every broadcast of a college basketball game will feature a commentator making some statement hinting towards the contrary. While I love to gnash my teeth and decry their ignorance, digging into the data did provide some pause in my kneejerk rejection of player continuity’s overinflated value. While the trend isn’t large, it isn’t negative. There does seem to be some value to the dogma when it comes to the extreme cases, at least until Louisville, Kentucky, DePaul, California, and Ooey Pooey dominate college basketball this season.