Probing Jesse Newell's critique of Self a little more deeply

I think everyone on here knows how big of a fan I am of Jesse Newell, the former KU beat writer at the LJW, TCJ, and now the KU beat writer at the KC Star. I love his scouting reports, post-game analyses, his deeper dives into game film, and his willingness to push the envelope with interesting quantitative analysis.

At the same time, everyone on here also knows that I'm one of KU's harshest critics when it comes to falling short of the high post-season goals that the 2nd winningest college program all-time sets for itself (or should set for itself, IMO).

With that by way of background, I've thought more and more about this KC Star piece of his , which I linked in this thread yesterday , and I must say that I have some serious issues with it.

As full disclosure, he and I have been discussing it a bit today on Twitter--or as best as one can in trying to have a meaningful discussion on a site that allows for only 140 character drive-by posts. Because I cannot possibly have a reasonable and informed discussion while handcuffed with such artificial boundaries, I've decided to post my thoughts here--both so that Jesse can better see where I'm coming from (should he care to visit our friendly home) and for the benefit of the Flock here...should you have perspectives to add. If you do, whether positive or negative toward either Jesse's approach or my well-intentioned pushback toward it, I feel that's the most sane way to proceed if the goal is to engage in an informed give and take.

______________

I'll cut right to the chase by positing that I think the analytic approach he chose has some notable flaws--first, in terms of the numbers it seeks to employ; secondly, with respect to some important (and perhaps more straightforward) success/failure indicators it chooses to ignore; and thirdly, in terms of an apparent double-standard that applies when many compare Self (who has a bad tourney rep nationally) with Tom Izzo (who enjoys an outstanding tourney rep nationally).

The first flaw is a doozy, and it centers on the small sample size of data that ironically (and damningly) starts in 2009--the year after Self won it all. Oops! Jesse admits to this being a shortcoming, and it's entirely inadvertent in as much as that's as far back as the KenPom win-probability data goes. But whether it's unwitting or not, it's still a huge flaw with the model--and therefore, IMO, with the conclusions.

The second problem centers on this line from the piece, which I find to be very telling in terms of national reputations and narratives: "Coach K ranks second, while Tom Izzo — hurt by this year’s early exit — comes in third despite a strong postseason reputation." (Bolded text mine, for emphasis)

The problem I have here is that the piece dives deep into some particular numbers and quantitative analysis for Self, yet then falls back to a rather lazy perception or meme--widely perpetuated and magnified nationally--about Tom Izzo and his recent tourney performance. At the same time, the piece doesn't account for other measures (which some might argue are more concrete, defining, and straightforward) to assess relative post-season success.

Here are just a few that come quickly to mind, focusing on 2003-present (which is a very clean and valuable starting point, in as much as Jesse's focus is on KU and Self, and it was then that Self as a coach began operating at a program that could consistently compete with the likes of Izzo and Michigan State):

* Since Self took over KU in '03, he has gone 30-12 overall in the Dance (for a 71% winning percentage) against Izzo's 27-13 (a 67% winning percentage).

* As for deep runs? Self has taken KU to 6 Elite 8s, while Izzo has taken MSU to 5. Izzo does have the edge in Final Fours during that time, however: 4 to Self's 2.

* But in the deepest of runs that matters most of all (the National Championship game), Izzo has made only 1 over those 13 seasons while Self has made 2.

* We also know full well that Self has taken KU to the promised land with a national championship in 2008, while Izzo has failed to get a ring during that time frame. (Yes, he did get a ring one time in his Michigan State tenure, but that was many moons ago--back in 2000)

When you see the absolute numbers presented like that, who, I would ask, has the stronger post-season record, as opposed to just falling back on "strong postseason reputation" (Jesse's description of Izzo's stature)?

___________________________________

Here are some other interesting facts that a piece like this, focusing on comparative tourney performance, doesn't mention--again, in this case, focusing on Self vs. Izzo:

Michigan State just flamed out as a 15 versus a 2 in this current tourney--an ignominious feat that has never happened to a KU team in the long sweep of its post-season history, much less a Self-coached team.

-- It bears noting that Sparty's game with Middle Tennessee State wasn't even close.

-- As it pertains to performance against expectations, this is the same MSU team that many commentators thought should be a 1 seed...and a team that was most favored to win it all not only by the national commentators but also the most credible probability-based forecasting sites (such as fivethirtyeight.com). This acute level of "failure" by Michigan State (relative to expectations for its success) surely has to count for something, in terms of strong weighting for underperformance--as compared with KU's tourney performance over Self's tenure. (Self's worst flameout at KU definitely sucked, but sucked far less, in that it came as a 3 seed playing a 14 in 2005, and in that game KU only lost by 1 point.)

The other thing that a purely numbers-based approach fails to do is account for key mitigating factors, which don't show up in the numbers but in KU's case contributed to two of its recent, premature exits under Self: Embiid's devastating injury in 2013-14 (no way KU exits early if he's healthy) and "Cliffgate" last year (when he couldn't suit up to play).

-- A healthy Embiid playing with Wiggins and the rest of that KU team in 2014 arguably had the potential to win it all; instead, the loss of their critical rim protector left them highly vulnerable and unable to adjust in such a short time prior to the Dance. Such factors cannot be dismissed lightly.

The last thing I would point out--in addition to Self's greater number of tourney wins than Izzo since joining KU, his higher tournament winning percentage during that tenure, and his inarguably greater success in reaching the pinnacle of the national championship game (and winning it one of the two times while there)--is that Self has absolutely run circles around Izzo when it comes to high-level, consistently elite performance facing P-5 competition in the regular season.

-- Izzo's worst team during the period in question was 19-15 (56% winning percentage) and arguably should've gone to the NIT. Self's worst team was 23-7 (for a 77% winning percentage), and he's never come close to not making the tourney at KU--something, I might add, that Calipari, Williams, and K cannot say for themselves.

-- Looking more broadly at that level of success, since he joined KU in 2003, Self has amassed 385 regular season wins (for an 82% winning percentage), while Izzo has only garnered 335 wins (for a 72% winning percentage). Given Self's exceptional and uber-consistent regular-season performance, it is perhaps inevitable that KU has fared more poorly than those schools and coaches who have not garnered similar regular-season success--and thus had an opportunity to play with house money and punch above their seed in the Dance.

-- In other words, even while many of us (myself definitely included) hit Self hard (and Roy before him) for not taking KU to the promised land as much as we'd like or expect, it certainly can be argued that Self has set himself up for more failure by dint of how good his teams are in the regular season. This often results in a very high seed, with the resultant, very lofty expectations...and bitter disappointment when they don't go far enough. For teams that don't enjoy such regular season success and garner lower seeds, they can play as upstarts and exceed expectations far more easily.

At the end of the day, there are different ways to skin this cat. No hard-core KU fan (myself very much included) is happy with our post-season performance since 2012 (or with those two early exits in 2004-06). And yet it's also true that well-respected coaches such as Tom Izzo have more than their share of "warts" and issues when it comes to consistently elite performance. Roy's post-season record at UNC? That's a different topic altogether. But when it comes to comparing Self with a high-quality coach like Izzo, with the frame of reference being the period during which Self has coached at KU, the results are more mixed than what the popular narrative conveys...and what this KC Star item presented.

File this in the "for what it's worth" category. And a big shout-out to Jesse for always pushing the envelope, providing unique analytic angles, and not shying away from provocative positions.

Rock Chalk...

Yep, and that's the point I was making with this:

Self has set himself up for more failure by dint of how good his teams are in the regular season. This often results in a very high seed, with the resultant, very lofty expectations...and bitter disappointment when they don't go far enough. For teams that don't enjoy such regular season success and garner lower seeds, they can play as upstarts and exceed expectations far more easily.

Two additional points here:

* If we use that KenPom win probability model, Self would fare better if KU sucked during the regular season, then overperformed in the Dance. Say, lose 12 games, draw a 6 seed, and win a couple in the Dance--then he'd look great. But would that make us feel awesome as fans?

* Even though he needs to do better in the post-season to become elite, it's also true that, ironically, he's a victim of his own success--namely, by coaching up sometimes relatively average teams during the regular season, then gaining a high seed (raising expectations), then often faltering in the Dance.

The most obvious case in point was our 2014-15 team. By the time we got to the Dance, we were in the midst of "Cliffgate," Perry was hurt, and we looked mediocre. I knew that team was headed NOWHERE in the tourney, and yet we got a very high seed (a 2) based on the regular season. Then the moronic NCAA committee hands us way-underseeded Wichita State in the 2nd round. A recipe for disaster, meaning a very early bounce. In that case, given how the team limped into the tourney, numbers alone can't do the analysis justice. Yet Self takes a big hit for that failure in the Jesse/KenPom model. Again, that team was going nowhere in the post-season, then got daggered by the committee with WSU, and the disjunction between the high seed (based on regular season performance) and actual results drags Self down.

2010 is not as good of an example but still salient. You were in OKC along with me, so we both drank the kool-aid of high expectations with that team. It was ranked #1 overall with a 1 seed--massively high expectations. But if we're being brutally frank, that team was Sherron, Cole, and the "rest." If one or both of those guys had an off night, our margin for error went down to zero. Tyshawn was young and out of position as a shooting guard. The Morris twins were coming along, but still sophs who would need one more year to fully blossom. Xavier Henry was trying to live up to his Dad's ridiculous expectations, gazing out the window at the NBA, and often disappearing on the court. We had to rely on Sherron over and over and over again that year to pull our chestnuts out of the fire. Even more so down the stretch and into the tourney when Cole got that weird bug or whatever that led to asthma-like problems. By the time we got to OKC, the guy still couldn't run up and down the court consistently, and our vaunted high-low game suffered greatly. Throw in UNI catching lightning in a bottle with Faroshkockahoochee and that big ogre of theirs hitting the only 3s he'd ever hit in his life, and you had the biggest letdown in many a year for KU. But UNI was a very good, underseeded team as well, and KU had "issues" heading into the tourney (most notably, Cole) that Jesse's purely numerical model can't account for.

Anyway, enough from me on this. But as long as KU does well in the regular season, it's going to land higher seeds. That, coupled with its heightened "win probability" calculation (per Jesse's model off of KenPom), is going to result in higher expectations for all observers, whether regular fans or stat nerds. When we don't make deep runs, expectations will be dashed and much angst will result. (And I'm part of that dynamic--I'm like you, I want and expect elite success from a program that takes a back seat to no one!) Meanwhile, of course, other teams will earn seeds at the level of 4 through 10 and have a much greater opportunity to "over-perform." (See: Syracuse this year--Boeheim is getting a huge boost relative to the likes of Self from his deep run with a mediocre team)

Rock Chalk...

The horse can't be dead when you leave with this:

No doubt the majority of fans prefer the way it has played out with 12 straight Big 12 titles and limited success in NCAA.

I actually think the vast majority are like you and I. Elite success in the NCAA is first priority (NC crowns). Regular season elite success comes right behind that. But both are wanted, hoped for, and (yes) expected. It's tough for any coach to live up to that, but that's why Self, in his first-ever press conference as the new KU coach, joked that his (Jayhawk-labeled) seat was "hot." At KU, you have pressure (and expectations) to be elite. He's tasted the fine wine of elite success (and so have we, with him), in '08 and again in '12 versus a psuedo-NBA team in the NC game. He'll get there (still a relatively young pup)...or so we all hope and pray.

PS: Amongst KU friends and fans, I'm brutally honest about our program--and our need to close the deal more with NC crowns. But when an outsider or someone in the MSM prints something that enemy fans and foes of KU will pick up on to bash us ("Rock Choke Jayhawk" and other such nonsense), I circle the wagons and let it rip. We have nothing to apologize for to wankers who love to hate KU. (Jesse is not one of those, btw--he eats, sleeps, and drinks the Hawks, just like us) Speaking just amongst ourselves, however, we need to do better--both to shut the yaps of the detractors but more importantly to close the one remaining "gap" in our elite, second-to-none historical program's resume.

Now, the horse is officially buried.

Rock Chalk...

Corpus--Good stuff. Thanks for the response, which is precisely what I was hoping for (that you'd weigh in at some point).

RE your opening:

K, with all due respect I will disagree on a couple of your points. On the point of ignoring the first 5 years, that is clearly a faux pas. No way can you ignore part of your dataset. That is a cardinal sin. I understand that he did not have KenPom probabilities back beyond 2009 but he should have used some other probability source. I could provide him mine which are derived largely from the actual tourney results but adjusted a little based on various things I felt made them more accurate.

It's a little confusing how you wrote that, because you start by saying you disagree with me, but then you cite a cardinal sin with Jesse's dataset. To a casual reader on here, the way you wrote it, it sounds like you're taking me to task here, when in reality you're agreeing with me about a major flaw I see in Jesse's dataset (the arbitrary cutoff date and the ensuing very small sample size).

After that opening, though, it's clear you see eye to eye with him and differ with some of my thoughts. I think the source of these disagreements goes well beyond numbers and to the yin-yang/qualitative-quantitative nature of measuring things like "expectations" versus performance. I have a LOT to say on this, and I think it's both interesting and that it matters, in as much as it goes to the core of how non-casual sports fans view any given team, its performance, and measure results.

I'll elaborate more on this later on, but, alas, I'm swamped with work and I really need to get moving out on my taxes. That little March Madness excursion to Louisville threw me off on both fronts. So...I'd humbly and graciously ask you to stay tuned, check back in at some future point here (probably sometime over the weekend), and we can carry on further with this. For anyone else, please don't let my temporary absence deter you from weighing in also. Just know, however, that I am not done putting forth some perspective on the topic.

Thanks, Corpus!

Rock Chalk...

Wouldn't it be great if we had answers to such question as why UConn? Kind of like finding the answer to the meaning of life.

Perhaps the success of UConn has to do with recruiting great guards and otherwise solid college players in support and having a great bench coach in Calhoun. His disciple Ollie showed he could win with Calhoun's players, but we'll see if he can with his own.

By the way, most people I talk with today are fine with "that one guy". Some are so misguided by still liking him as to prefer he (not his team so much as he) wins Monday night. Others, like me, are fine with him, and are even former fans of the program he took over before the fiasco that led to him taking it over. But that "former" is in concrete, and the academic fraud revelations have only cemented my decision to never, ever again be supportive of that program winning against anybody, not even Duke -- with the possible exception of the Lexington school, at least as long as sleazeball seafood is still there. That would represent a true tossup for me, and a time to call for the meteor.

Still lots of angst, won't get better if the tarred ones win, either. Hang in there.