KenPom, Memphis, and Preseason Projections
Analyzing Sunday's brand new KenPom preseason rankings
With yesterday’s release of the KenPom preseason rankings, today’s Hoop Vision Plus post is unlocked for ALL to enjoy. Hoop Vision Plus is a premium newsletter combining X’s and O’s, analytics, and coaching to help you watch the game smarter. To receive more college basketball analysis like this all season long, subscribe to HV+ today!
If you were on Twitter yesterday, you probably saw discussion about the overnight drop of the 2019-2020 KenPom rankings.
KenPom.com is one of the go-to stats sources for college basketball fans and coaches alike. For many coaches, it’s become an essential part of the scouting process; Colgate assistant coach Dave Klatsky even had Ken Pomeroy on his Statchat podcast last February to discuss the applications of the site for coaching and scouting.
Today, we take a look at the sites’ most popular feature for fans: the rankings.
What teams are projected to make the biggest jumps forward this season?
How have teams projected to make a similar jump forward in the past done by season’s end?
How much predictive value do the updated preseason rankings provide over last year’s rankings?
What should we make of Memphis being ranked #50? Are there any historical comps for the Tigers?
Biggest jumps: teams expected to improve
Ken explained the updated methodology behind his preseason rankings back in 2016. The big change that season was the addition of transfers into the equation.
With that in mind, it probably shouldn’t be a huge surprise that the team projected to make the biggest jump in adjusted efficiency margin is UTEP.
UTEP (-11.3 —> +2.9)
UNC Asheville (-23.7 —> -10.6)
Arizona (+7.3 —> +17.9)
Towson (-10.7 —> -0.2)
Albany (-9.9 —> +0.5)
In year one of the Rodney Terry era, UTEP went just 8-21. But Terry has dipped heavily into the transfer market as he aims to turn things around in year two.
The Miners beat Texas Tech earlier in October in a charity exhibition game — and first-year transfers combined to play 62% of available minutes in that game. Eric Vila (Texas A&M), Bryson Williams (Fresno State), Souley Boum (San Francisco), and Daryl Edwards (LSU) all recorded between 24 and 35 minutes against Texas Tech.
What should we expect from the teams projected to take a big jump?
We now have an eight-season sample size for the KenPom preseason projections. The top projected jumps from those past seasons are listed in the table above.
Score one for Ken’s preseason rankings. Every team in the table finished closer to their preseason projected AdjEff than to their previous season’s AdjEff.
If we filter for just high-major conferences, we have the following teams projected to take the biggest steps forward this season.
Arizona (+7.3 —> +17.9)
Wake Forest (-0.9 —> +8.7)
Georgia (+4.1 —> +13.4)
Xavier (+10.6 —> +19.2)
Illinois (+8.1 —> +16.5)
Sean Miller and Arizona lead the way among the high-majors. The Wildcats bring in star freshmen Nico Mannion (RSCI #6) and Josh Green (RSCI #13) — in addition to two D1 transfers Stone Gettings (Cornell) and Max Hazzard (UC Irvine).
Combine that recruiting class with a strong historical track record for Sean Miller, and Arizona is Ken’s top ranked Pac-12 team.
Beyond simply the teams projected to take the largest jumps, how much predictive value do the preseason rankings add?
The graph on the left features the correlation between the previous season’s adjusted efficiency and the end of season’s adjusted efficiency. The graph on the right features the correlation between the preseason adjusted efficiency and the end of season’s.
The previous season actually does a reasonable job of projecting year two. This might be a surprise given the narratives about roster turnover surrounding college basketball. That being said, Ken’s preseason projections do add value to the projection (R increases from 0.82 to 0.88).
The difficulties of projecting Memphis
In year two at Memphis, Penny Hardaway has delivered on the promise of bringing in big recruits. The incoming class includes five players in the RSCI top 100:
James Wiseman (RSCI #1)
Precious Achiuwa (RSCI #14)
Rejean "Boogie" Ellis (RSCI #36)
D.J. Jeffries (RSCI #45)
Lester Quinones (RSCI #60)
If we go back to Ken’s 2016 explanation of his preseason rankings, he explained:
Finally, there is no change on incoming freshmen. The top 30 or so have an impact on a team’s rating and beyond that the computer is mostly blind to newcomers. That’s not to say it can’t make some guesses, though. In fact, it’s kind of a fun challenge to predict the impact of recruiting classes without any information on the recruiting class itself. Things like basketball budget, conference affiliation, recent performance, and whether the coach is returning handle some of this. But history says you can also glean some information from what kinds of players have left a team.
In Ken’s system, only the top 30 or so recruits in the country have an impact on a team’s rating. And, knowing Ken, it’s safe to say there is a reason for that — players beyond that point likely have limited predictive power.
But the second part of Ken’s point is particularly pertinent to Memphis. Even if recruits outside the top 30 don’t have direct impacts, incorporating things like basketball budget, conference affiliation, and recent performance tend to help account for them indirectly.
With Memphis, they are in fairly unprecedented territory. We have seen recruiting classes similar to what Penny did this year, but mostly from the blue bloods of the world — not a school in the AAC.
This, of course, makes the Tigers difficult to project. To find potential comps for Memphis, we went back and identified young teams (an average of less than one year of experience per player) with low roster continuity from the previous season (less than 25% of minutes coming from returners).
A graph of the top rated teams to fit the young and inexperienced criteria is below:
Teams above the line exceeded their preseason expectations
Teams below the line fell short of their preseason expectations
In all, eight of the 18 instances which met the young and inexperienced criteria belong to Kentucky, Duke, and Kansas. None of the 10 non-blue blood teams finished in the end of season top 25.
The two closest comps to 2020 Memphis are 2016 Washington and 2012 Texas.
Washington brought in a recruiting class of Dejounte Murray (#45), Marquese Chriss (#52), Noah Dickerson (#75), and Matisse Thybulle (NR). By no means was that recruiting class as highly-touted as the Memphis class, but — with the advantage of hindsight — it now looks very strong given the final results of the those players.
Lorenzo Romar’s team started the season ranked just #134. They finished the season ranked #66 — and Murray and Chriss were both selected in the first round of the NBA draft.
Texas brought in a recruiting class of Myck Kabongo (#13), Sheldon McClellan (#48), Julien Lewis (#68), and Jonathan Holmes (#70).
All four freshman played key minutes for Rick Barnes during the season. The Longhorns opened the season ranked #30 and finished the season ranked #32.
For Memphis, James Wiseman is where those two comparisons fall (somewhat) apart. Both Washington and Texas had strong recruiting classes, but not a player ranked top 10 nationally — let alone number one.
We have seen elite freshman talent not necessarily translate into wins (Ben Simmons and Markelle Fultz), but the hope for Memphis fans is that an elite talent plus the rest of a very strong class can put the Tigers closer to the Duke/Kentucky/Kansas grouping than the Texas/Washington grouping.