Who are college basketball's great overachievers on the sidelines?
A misguided attempt to answer the unanswerable
Let’s start with the obvious: there is no right answer to this question.
I have had better introductions to posts in world history, but I figure that’s a disclaimer we should get to right off the top. There is no perfect way to quantify a coach’s impact or lack thereof on or off the court. These days, there’s about five million things that go into a team before they even touch a basketball court for game one of a potential 40-game season. There are times in history, and still presumably schools in existence, where you can be good at off-court or on-court but not both. At the highest level, that hasn’t been the case in a long time.
So! With that out of the way, here’s the real story. This is a question I’ve been pondering since spring 2022, with various achievements in attempting to find the right answer. Ultimately, I came to the conclusion you see in the opening sentence of this piece: there is no right answer. There’s several answers, because you can look at a lot of different stats or qualifiers to get you the answer you think you’d like.
I’d roughly break down the idea of the Great Overachiever title into four categories:
Has the most Wins Above Bubble, meaning total wins adjusted for schedule difficulty;
Has the highest average KenPom/Torvik rating over a given period of time;
Gets the most out of the talent on his roster, using a metric available on Bart Torvik’s site;
Does the best job at beating preseason expectations.
There are many, many other categories one could use, including record versus the spread (available at TeamRankings in team format) or the ultimate decider, How Did You Do in the NCAA Tournament. That one seems to be everyone’s favorite at a very specific time of the year, so I suppose you can count Dan Hurley as the best coach in America if you put Eric Musselman #2. It’s only fair, right?
In the end, I’ve done the work necessary to produce a decent group of answers. (A list of what I guess you could call anti-answers, or coaches who’ve done the least will be available in a B-side article later this week.) Depending on what you personally prefer, you can take the answer you’d like out of that group. Or, if you don’t like the numbers, find your own. I’m sure that with enough work and guile, one can twist any set of statistics into supporting the coach they want into being #1 in the nation.
Lastly, here’s the qualifiers I’ve used.
Must have coached in every season from 2017-18 through 2023-24, at least as their teams were eligible to do so. The Ivy League schools as well as some scattered others (eventually, eight other schools opted out at some point of the year) sort of get grandfathered in here. It wasn’t their fault that there was a pandemic, obviously, so it’s pretty hard to exclude them. Coaches of those teams in a theoretical 2020-21 season got included here. Why 2017-18 through 2023-24? It gives us three full(ish), normal seasons on both sides of the COVID coin. This eliminated John Gallagher (205 games coached) through no fault of his own, thanks to the incompetence of Hartford.
Except for the COVID exceptions listed above, must have coached 200+ games at the Division I level in the last seven seasons. Pretty much everyone should meet this because everyone plays 30 games per season now at minimum. Those who didn’t qualify but coached in at least 100 games/three seasons will get name-dropped in the B-side article mentioned. By a hair, this means Chris Beard (201 games coached in 6.5 seasons) DID make the cut.
Must still be coaching a Division I team in 2024-25. Well, what would the point of the article be if we were talking about someone who isn’t coaching in college hoops anymore? Legends like Cliff Ellis (200 games) to go with names like Steve Henson (200), Reggie Witherspoon (201), Edward Joyner (207), Dan Monson (207), Scott Pera (208), Keith Dambrot (211), Jeff Linder (212), Mike Davis (213), Jeff Jones (213), Travis Ford (217), Dan D’Antoni (219), Jerod Haase (222), Mike Hopkins (224), and Mike Boynton (228) will all be exempt from this list.
That leaves you, dear reader, with a list of 125 coaches who made the cut. These 125 will constitute our initial list that we’ll break down category by category. I’d even argue that I’ve done the job of giving you headliners and supporting acts. #1 and #2 are almost entirely going to be made up of coaches at high-major jobs, because they have a lot of money and win a lot of games. #3 and #4, conversely, will still give you some of those same names but should highlight plenty of overachievers from the mid-major levels that you may be underrating.
Is this an accurate reading of the situation? Maybe, maybe not. But I’ve wanted this to exist for years now, so might as well be the fool and open my mouth. On we go.
1. Are you winning?
THE PITCH: Of the four categories here, this is the easiest to sell. The only thing the average fan cares about is winning, which is supposedly the entire point of sports once you hit a certain level. Instead of using pure win-loss, which doesn’t adjust for strength of schedule, we’re using Wins Above Bubble, a publicly-available metric that looks at your wins and adjusts them for how the average bubble team (typically around 50th in a KenPom/Torvik system) would have been expected to perform.
THE RESULT: Of the four categories here, this has the least surprising results. Your WAB Champion, from 2017-18 to today, is Bill Self of Kansas. Self’s teams have won an astonishing +51 games above what would be expected of the average bubble team over the last seven seasons. What this means: the average bubble team would have gone 137-106 against Kansas’s schedules. Kansas went 188-55. Only one other coach in America even touched +40.
THE OTHER RESULTS: The remainder of your top five went as such.
2. Mark Few, Gonzaga (+40). For all the hate Gonzaga gets from uninformed morons - sorry! - they’re actually exactly as good every year as everyone claims. The average bubble team would have gone 171-70, not 211-30.
3. Matt Painter, Purdue (+39). Sexy basketball? Infrequent. Winning basketball? Basically every year now. Purdue’s COVID-excluded average of about +6 WAB a year would’ve ranked sixth-best this past season. That’s the average, meaning that for the last seven years Purdue basketball has functionally won at the rate of a 1 or 2 seed on average each season.
4. Kelvin Sampson, Houston (+37). Everyone whined about Houston being in the AAC for so long, so they went to the best conference in America and won it with ease on their first try. Seems good.
T-5. Scott Drew, Baylor and Tony Bennett, Virginia (+31). Both have won a title in this stretch; both have also had some really ugly seasons. The problem with both is that neither is exactly interested in media coverage nor with being conventionally attractive on the court, which means a pretty consistent run of quality ball may go overlooked versus other contemporaries of their time.
Honorable mentions: Rick Barnes, Tennessee (+29) and Tom Izzo, Michigan State (+26). No one else above +22.
WHY IT’S NOT PERFECT: Some key and obvious names are being left out here, for one. Dan Hurley, Nate Oats, Greg McDermott, and several others didn’t really come close to the cutline for one reason or another. It also means that, by virtue of inexplicable close game luck in one specific season, Greg Gard can be considered a superior coach to Dan Hurley. Perhaps not.
The other problem is that the coaches who rank as the worst in WAB often have really good excuses for it. Dead last is Wayne Tinkle at Oregon State, who has the lowest basketball budget (which will get lower) of any 2023-24 Big Six school. The next-lowest are Lamont Paris and Shaheen Holloway, who spent the majority of their time in this sample at mid-major schools and have taken on hard jobs. I don’t think anyone would call either underachievers.
2. Are you winning consistently?
THE PITCH: This one is pretty simple. Over the last seven years, how consistent were you year-over-year? The best way to do this is using Torvik’s publicly-available tools, which allow you to sort over the last seven years to see the coaches who’ve overseen the most consistently efficient, excellent programs. With little surprise, perhaps you’ll guess the top five with…fair ease.
THE RESULT: More than anyone else in the last seven years, the most consistently nasty program to deal with on a nightly basis has been Mark Few and Gonzaga. By a hair, they bested Kelvin Sampson and Houston for pure year-over-year dominance. The comparison between the two is a delight. From 2017-18 to now:
Number of top-three finishes: Gonzaga 4, Houston 3
Highest finish: Gonzaga #1 (twice), Houston #2 (thrice)
Lowest finish: Gonzaga #12 (2023-24), Houston #18 (2017-18)
NCAA Tournament wins: Gonzaga 17 (#1 in America), Houston 14 (T-#2 with Kansas and Duke)
eFG% rankings: Gonzaga #1 offensively (58.2%), Houston #1 defensively (43.7%)
Like it or not, by this metric, these have been the two most consistently high-end coaches over a seven-year period. It’s to the point that they’ve even bested some now-gone names like John Beilein (#3 if eligible) and Coach K (#4).
THE OTHER RESULTS: The rest of your top five is unsurprising.
3. Matt Painter, Purdue. Same reasons listed above. Four top-10 finishes in seven seasons, zero outside of the top 25.
4. Bill Self, Kansas. The Kansas discussion is really interesting because if you tune into any random December game against, say, North Dakota State, Kansas will probably only show interest for about 20 minutes of it before doing whatever they have to do to reach the finish line. It’s against top-50 competition where Self pays the bills. Since 2017-18, Kansas is a mind-boggling 83-47 against top-50 teams, per Torvik. No other coach in America even cracks 70. The efficiency hasn’t always been there outside of 2019-20 and 2021-22, but the wins absolutely have.
5. Rick Barnes, Tennessee. This is the reverse of the Self argument. Barnes and crew have an unearned reputation as a group that demolishes lesser competition but struggles against the elites. This hasn’t really been that true - Tennessee is 26-31 versus top-25 Torvik teams since 2017-18, which places them seventh in total wins versus that group - but it is true that since 2017-18, no coach has been meaner to Quadrant 4 competition than Barnes. Tennessee is a flawless 34-0 and ranks #1 in overall efficiency versus Q4 comp. Barnes also has given Tennessee their four best KenPom finishes in school history, which should count for something.
Honorable mentions: Scott Drew, Baylor; Tom Izzo, Michigan State; Chris Beard, Texas Tech/Texas/Ole Miss.
WHY IT’S NOT PERFECT: I think it works and it doesn’t. Coaches who spent a year or two in the mid-major ranks prior to taking a larger job are unfairly penalized here, because their previous schools had hard-set ceilings they couldn’t crack. Would anyone rate John Calipari (11th) out as a superior coach to Dan Hurley (12th) at this moment in human history? What about Chris Holtmann (16th) over Shaka Smart (17th), Randy Bennett (18th), or Brian Dutcher (20th)? Nice as it is on paper, this just feels like a starter more than the whole thing.
3. Are you winning consistently and reaching your roster’s best outcome?
THE PITCH: Something most people won’t know about because it’s only visible on the Team Table of his site is Torvik’s Talent Rating, which more or less measures on a 0-100 scale how talented your roster is. It’s largely based on recruiting rankings, so perhaps you can give or take how much this means to you. (Given the +0.7 correlation between talent and efficiency over the last seven years, which qualifies as a significant positive relationship, it should probably mean a lot.)
I became really interested in this in late 2022, when I was attempting to find ways to quantify a coach’s impact on a roster. The coach (and staff) should receive credit for getting the players in the first place. The coach’s job is then to get the most out of said players, which is not always easy to do. There’s a real difference between what Bill Self has done with deeply talented rosters at Kansas versus what Andy Enfield has done at USC, as an example.
The best way to do this, after two years of research, ended up being pretty simple.
Take a team’s Torvik Talent Rating and turn it into a Z-score, aka how much more or less talented they were than college basketball’s mean.
Do the same for their Barthag, which is the Torvik version of a KenPom rating.
Multiply the Barthag’s Z-score by two and substract the Talent Rating’s Z-score. That gives you the Total Over Expectation.
For context, a +14 Total Over Expectation (+2 per year) in a seven-year period looks a lot different if you’re doing it at Vermont (John Becker) or St. Bonaventure (Mark Schmidt) than if you did it at Houston (Kelvin Sampson). Also, the average was +5.8; the top 20 coaches were all at +12.73 or better.
We’ll do this answer in two parts: the overall top 5, then the best by Talent Rating ‘group’ (i.e., the best coach among top 50 recruiters, top 100, 101-200, and 201-infinity).
THE RESULT: Credit to DataMizzou on Twitter, who guessed this one. The biggest overachiever versus roster talent over the last seven years is Randy Bennett at Saint Mary’s, who turned in a +17.3 Total Over Expectation, or merely +8 above any other coach in the WCC. Context is heavily dependent, obviously, but think of it this way: Bennett has never had a consensus top 100 recruit on his roster at any time. Not ideal! And yet, Bennett and Saint Mary’s are 167-60 since 2017. They rank as the 21st-best program of the last seven years, per Torvik. That’s ahead of starlets like UConn, Illinois, Wisconsin, Marquette, and numerous more.
THE OTHER RESULTS: The remainder of your top five are as follows.
2. Bob Richey, Furman (+16.6). Richey recruits at roughly the 240th-best rate of any program in America, which makes it all the more impressive the lowest Furman finish in KenPom of his tenure is last year’s 141st. That came after six straight top-100 finishes, all of which rank as the best in school history. It is beyond baffling he has not found a new job yet, though the Three Man Weave boys believe he’s holding out for Clemson whenever that comes.
3. Kelvin Sampson, Houston (+15.41). Sampson and Houston have recruited over the last three years at the level of a top-20 or top-25 program, which makes it that much crazier Houston has finished 18th or higher seven straight seasons and has three straight second-place finishes. Basically, this is a program that recruited like a Minnesota or Colorado and now recruits/builds like a Tennessee. Scary!
4. Leon Rice, Boise State (+15.4). The great, consistent Rice has simply plugged away at this job for what feels like 25 years and always delivers. Rice has never had a top-100 recruit, but Boise’s got three straight top-50 finishes despite theoretical recruiting disadvantages and having the eighth-best basketball budget in the 11-team Mountain West.
5. Grant McCasland, Texas Tech (+15.39). My favorite college basketball coach. “Really?” Sure, why not. It’s not that serious. McCasland has delivered over and over again at multiple jobs now, first elevating North Texas to new heights, then taking over a Texas Tech program in disarray and putting them to a 6 seed in year one. We’ll get more into him in the next section, but McCasland is elite at identifying hidden gems and at building a coaching staff. Anyone smart enough to hire Dave Smart and Jeff Linder has the goods.
Honorable mentions: Porter Moser, Oklahoma/Loyola Chicago (+15.38); John Becker, Vermont (+14.9); Mark Schmidt, St. Bonaventure (+14.7); Brian Dutcher, San Diego State (+14.5); Chris Jans, Mississippi State/New Mexico State (+14.4).
THE OTHER OTHER RESULTS: The top five coaches by recruiting category were as follows.
TOP 50 RECRUITERS: Matt Painter, Purdue (+13.2); Scott Drew, Baylor (+12.7); Chris Beard, Ole Miss/Texas/Texas Tech (+11.4); Bruce Pearl, Auburn (+10.9); Rick Barnes, Tennessee (+10.6).
51-100 RECRUITERS: Randy Bennett, Saint Mary’s (+17.3); Kelvin Sampson, Houston (+15.41); Brian Dutcher, San Diego State (+14.5); Mark Pope, Kentucky/BYU/Utah Valley (+13.7); Nate Oats, Alabama (+13.3).
101-200 RECRUITERS: Leon Rice, Boise State (+15.4); Grant McCasland, Texas Tech/North Texas (+15.39); Porter Moser, Oklahoma/Loyola Chicago (+15.38); John Becker, Vermont (+14.9); Chris Jans, Mississippi State/New Mexico State (+14.4).
201-INFINITY RECRUITERS: Bob Richey, Furman (+16.6); Mark Schmidt, St. Bonaventure (+14.7); Casey Alexander, Belmont/Lipscomb (+13.9); Ritchie McKay, Liberty (+13.7); Brian Wardle, Bradley (+13.2).
WHY IT’S NOT PERFECT: It’s not a complete measurement. This is a useful list, perhaps, of coaches who get the most out of the roster on the court. It’s not at all a useful list of coaches who make the roster in the first place. In fact, if you were to try and rank the coaches based on something like two parts value to one part talent - not how I’d do it, personally - the sport’s top coaches would be a mess. You’d go Mark Few, Bill Self….then John Calipari. That’s to say nothing of theoretical top-five coach Mick Cronin.
As such, I actually find this most valuable for identifying guys that could deserve a better opportunity. It had McCasland nailed as early as 2019-20 as someone who needed a better job, and it had Pope by the COVID year, too. (Maybe not KENTUCKY level big, but hey.)
4. Are you winning consistently, reaching your roster’s best outcome, AND always beating the preseason expectations of the sport’s leading metrics site?
THE PITCH: Wordy! But it’s got a point. Ken Pomeroy, who you may have heard of once, releases his preseason projections sometime in October every year. He’s done them since the 2010-11 season; lucky for Ken, we only need data from 2017-18 to now. For all of the hate KenPom gets from people who don’t seem to get how it works, it’s proven to be remarkably accurate over the years. The in-season version of it more or less sets the Vegas lines for games, excluding situational ones or those with heavy injury involvement. It even has the most accurate read on the preseason now.
This one is really simple: how well, or how poorly, did a coach’s team perform versus the preseason projections of the most accurate metrics site in college basketball? This one obviously won’t be perfect (Ken presumably would say this himself), but if there are guys who overperform or underperform every year, it would be quite notable. In theory, no one should be able to ‘beat’ the system, but as I found out, some coaches are really good at it.
THE RESULT: So, uh, remember that part about “what if someone beats this every year?” There were three who did it. All three have beaten Ken’s preseason ratings by an average of +6 points or above. In our 125-coach sample, the average coach beat the preseason ratings by 5 points or more about 1.72 times each, or 1-2 times every seven years. Fair enough. These three coaches are responsible for 14 of those alone.
Your KenPom overachievement champion by a hair is Grant McCasland, formerly of North Texas, now at Texas Tech. McCasland has beaten the KenPom projections every single year of his eight-year coaching career by an average of +6.7 points per season. It says something that a +2.2 point overperformance last year with a patchwork roster was the second-worst outing of his career. Thrice, McCasland beat the projection system by 9 points or better. Only one other high-major basketball coach, a guy named Nate Oats, has also done it three times.
Second and third place have very different careers and styles. One of which has been mentioned a ton in this article; the other is getting his first mention of the entire piece. Second place is Kelvin Sampson at Houston, who has done the unusual of being projected to be really good every season and still routinely pantsing the projection system by an average of +6.5 points a year. The only coaches in his stratosphere among top-100 recruiters, unsurprisingly, are Nate Oats, TJ Otzelberger, and Dan Hurley.
Third place is controversial beyond belief. It’s new Louisville and former Charleston/Winthrop HC Pat Kelsey, who has beaten the projections seven straight years by an average of 6 points a season. In fact, over the last three seasons, no coach in America was better at beating the KenPom preseason projections than Kelsey, at an average of +8.6 a season. Considering most public computer systems have Louisville somewhere in the 60-80 range for 2024-25, a 6-point overperformance in Year One probably means an 8 seed.
THE OTHER RESULTS: No real surprises here. Fourth and fifth were Nate Oats at Alabama and TJ Otzelberger at Iowa State at +4.8 and +4.4 respectively. A quick note here: he came up all of five games short, but Michigan’s new head coach and Florida Atlantic’s former one Dusty May would have been fifth at +4.7.
The rest of the top ten: Mark Pope, BYU (+4.38); Dan Hurley, UConn (+4.33); Leon Rice, Boise State (+4.29); Craig Smith, Utah (+4.25); Niko Medved, Colorado State (+4.15).
THE OTHER OTHER RESULT: Of the 125-coach sample, only one coach in the entire sport failed to top a preseason projection even once across a seven-season span. It may or may not surprise you to learn that this is Mike White, formerly of Florida, now of Georgia. White is the fourth-worst performer versus the sample overall, and only once (2018-19) did he come within a point of at least pulling even with the preseason projection. On average, he’s about four points worse. Sorry, Mike.
WHY IT’S NOT PERFECT: Two obvious reasons. First, you can overachieve every year and still not be that great or notable at your job. McCasland’s highest career KenPom finish is 31st, which is not bad at all but does put a temporary limit on how well he can be viewed. Secondly, if you’re projected to be really good every year, this probably isn’t a great measurement for you. Bill Self ranks as the 12th-worst coach at beating these projections and only did it twice: 2019-20 and 2021-22. Considering that in a non-pandemic world he could’ve won two championships, he probably doesn’t care.
5. Are you able to find the answer?
The answer, purely based on number of mentions in this article, would be any of these eight: Sampson, Self, Oats, Hurley, Few, Randy Bennett, Barnes, or McCasland. (My personal pick: Sampson, who was the only coach mentioned in all four categories.) I’m not sure that I personally get all the way there for any of them, and it still leaves out several good cases that I like a lot. So: I leave it up to you. What’s your pick here? Which of the four formulas, if you can call them that, comes closest to your personal truth? The comments are open, which could either be a mistake or a gift. Your call.
The B-sides article, with information that didn’t fit in this article (including smaller-sample all-stars!), will be available Friday.
I have missed reading Will’s work, this was a fun summer treat
As a Texas Tech Fan, of course my choice is the metrics that boost McCasland. This has given me a lot of faith in the hire but also makes me curious of how his career will continue at a school with a bigger budget and higher expectations. Now that 1/3 recruiting prowess will have to show through.