Dr. Self and Mr. Hyde

The data seems pretty clear to me after some in depth evaluation over the last 24 hours. First, the conclusion...Coach Self is second to none in regular season victories and especially in conference victories but he is totally mediocre in NCAA tourney victories. Now read what I said again. This analysis does not portend to judge his coaching only the resulting victories or losses. To wit; I have posted a table below that is a bit complicated. It measures NCAA tourney success in two different ways. First it scores a win over a higher (better) seeded opponent as +1 and a loss over a lower seeded opponent as -1. All expected wins or losses are scored as 0. The second method is to assign odds for each seed matchup (which I have) and for a win you add that fractional game and for a loss you subtract. For instance, if a 1 seed beats a 6 seed the probability of that happening is 0.665. So the 1 seed coach will get 1-0.665 or 0.335 and the 6 seed will get -0.335. This is actually a little more accurate than just calling it an upset or an expected. So with that here are the definitions of the table below.
Games - The number for each category, wins, losses and total.
Points - the odds derived points for wins, losses and net total.
Upsets - The 1 point per upset win and -1 1 for upset loss.
Net Upset% - The net upsets (wins-losses) divided by total games.

Where Self ranks - the Good
Wins - Self ranks 8th in wins since the tourney went to 64+ teams. he has played in 18 tourneys and won 2.2 games per tourney which is tied for 12th.
Winning% -- Self has won 70.2% of his tourney games which puts him 13th. (Incidentally, Larry Brown is 2nd).
Points -- Self has accumulated 9.8 win points, good for 13th.

Where Self Ranks - the Bad
Self has accumulated -9.99 points in losses good for 7th worst (of course this is also a function that he has been to so many tourneys)
Net Upsets -- Self has upset a higher seed 4 times and been upset by a lower seed 12 times giving him 8 net upsets good for 5th worst. Coach K is actually far and away the worst in this category. Bob Knight and Bob Huggins are also worse than Self as is Rick Barnes. Rollie Massimino is the best with 9 net upsets followed by Tom Penders with 8 and Tom Izzo with 7.
Net Points - Here is where the rubber meets the road. Net points essentially measures if a coach has played to seed or played better or worse than seed. A score of zero means a coach is playing exactly to seed. A positive score means better than seed and a negative obviously worse than seed. Bill Self, while far from being bad, has not played to seed. This is sort of the acid test in my mind. A coach that plays to seed is okay. Self actually has a slightly negative score of -0.18. Coach K is the best exceeding seed by 8.73 net points, followed by Pitino at 7.77, Roy Williams at 7.66 and Tom Izzo at 7.40.

Any given year the tourney is a crapshoot. But as you start compiling many years on your resume' these numbers mean more and more. Bill Self has done a fantastic job of getting high seeds in the tourney. But in 18 seasons he has been upset by a lower seed 12 times. Maybe it is just luck of the draw and maybe it is just the way the ball bounces. Clearly Self is one of the greatest coaches in the game. But based on the numbers, his tourney success is purely mediocre as compared to expectations.

My wife and I were talking about this today. So I'll ask you the same thing. Would it be better if Self had worse regular seasons and got lower seeds? Then KU would be the team upsetting the higher seeds in the tourney sometimes. Would it be better if KU didn't make the tourney some years?

I can appreciate the amount of work you put into these numbers. It's amazing.

I knew that Don would eventually get the statistical analysis that supports what we all believe to be true. I've sort of sensed it for a few years, and we all talk about it constantly. I certainly haven't done any analysis that meets any real world standard, but I have a theory, or perhaps several interrelated theories as to why Self's team tend to under-perform in the NCAAs.

First of all, Self is a defense first coach. Yesterday we were exceptional on defense, yet lost anyway. And the single biggest variable in putting together a defense is whether you focus on preventing 2s, or 3s. Self has always focused on cutting off the 2s, and taking what comes from the 3 point line. And sometimes a team gets really hot, and kills us from 3, and we lose.

Second, Self is a 'system coach' on offense, but he must have the players willing and able to execute his system for the entire 40 minutes. He prefers combo guards, yet the best teams we've had had true point guards running them. And our best Self team, 2008, had 2 point guards, with one starting out on the bench. In 2012 Taylor was our PG, and had become quite good, and as a result, we became quite good as well. We lost to Michigan because we didn't have a true PG. I could go on, but you get the point. Self prefers combo guards, but ironically his offense runs better when we have a true point guard to run it.

Third, his offense is complex, and requires not only intelligence but discipline, and we haven't exactly struck gold in those areas for the most part under Self. This year's team didn't become really successful until Self inserted Lucas into the starting lineup. Lucas is both smart and disciplined, and combined with Perry, also smart and disciplined, our interior offense flourished. And for most of the season, our guards played with intelligence, and most of the time, with discipline. Yesterday, and on a few other days at he end of our seasons, our guards did not play with discipline. I re-watched the game. Yes, I'm insane, but I would rather watch us lose than watch UNC win, and what I saw was something of a surprise. Frank Mason missed an open Perry Ellis 4 times in the second half, and each time it would have given us a better chance to score than how the sequence ultimately played out. Ellis did not have a good game because he was 'off' mentally, as was Selden. To some degree both played 'hero' ball, with little success, and it cost us more than a few possessions. But in the past this has happened in games just like this, the result of individuals ignoring Self's desired approach and 'free lancing' instead. It has cost us at least 3 shots at the Final 4, IMHO.

Last, Self's system requires time and practice to implement fully and correctly. In league play our guys figure it out after a while, and we get into a rhythm by the 5th or 6th game. And Self knows who we are going to play, so he can plan effectively, and create practices designed to attack each defense most successfully. In the NCAAs, we lose when we don't have 4 or 5 days to get ready for our next opponent. For some reason Self can't get his guys fully prepared in 2 days, and that's when we lose. It is not a coincidence. If we had to play Villanova with 4 days to prepare, I believe we would beat them most of the time, and definitely would have yesterday. But 2 days isn't enough for Self's system to work optimally. I don't know what to suggest, but that's my theory.

I think that Self is simply better at defense, yet ironically we lose because we don't score enough points in these games. Perhaps he's alter his approach. I don't know, but I can hope.

I think we'll see next year. Definitely a rebuild.

Peter, I largely agree with your assessment. I think the reasons we have lost so many games we should have won in the tourney are legion. I think you caught the general principle. I would double down on one point you made and that is Self has a strong tendency to stick with the plan even when the plan is not working. I still believe that 90% of the KU stalwarts saw that Elijah Johnson was not up to the task against Michigan yet Self stuck with him and the rest is history. Some would say, what alternative did he have and that point is well taken. I completely agree that Mason was off his game against Villanova and largely the entire tourney. He made at least 6 clear mistakes against Nova that I would say are uncharacteristic. I struggle with how much to put on Self in any particular instance but when you look at the 18 year run, Self unarguably has been mediocre in the tourney as compared to his elite of elite in regular season. As to your point guard point, I agree but an astute observer would point to the 2005 team with Aaron Miles and the flame out against Bucknell. But despite the amazing college talent on that team, that team was dysfunctional. It was the Self guys and the Williams guys and they did not mesh and Self did not have control of that team. That was an immensely talented team and had no business losing to Bucknell. And that was with the consummate point guard. But generally I agree with your assessment. Self's system of inside out requires PG's who get the ball inside. Over a season we do that better and better. But a good coach can confound that and if we do not adapt it becomes a crapshoot or a loss.

i felt the need to rereview your analysis on this point corpus. first let me say that i love the time and effort you put in on all your posts and when i visit the board i am delighted when i see youve posted. its always interesting and thought provoking. on this one thread i must take issue and pls understand i have no desire to offend in any way. in this case i think your argument is simply tautological. self has almost always earned a one or two seed. he then frequently underperforms that ranking. that is absolutely bound to happen given the amt of chance in a one and out tournament. it is impossible to outperform as a one seed and very hard to do as a two. this is math not opinion. i think the right analysis would be to establish an index over a long period of time for each seed ( ie all ones, all twos, etc) to see where the avg finish is for each overall seed and then compare it to selfs' finishes vs seeding. does he index hi or low vs that avg. what at least that wld produce is a more precise assessment of performance vs expectation. my guess is he still wont be top quartile on that basis, but certainly not bottom. whatever the outcome tho i am more in the camp that the tournament is a new season and a coach shldnt be perversely penalized for high expectations. in that model all wins are positive in which case self is obviously ranked at or near the top. but, more importantly, whether one takes my approach or yours i just dont find it logical to then jump to a list of reasons why he "fails" in the tournament. getting there every year as a high seed, winning sometimes and losing others, doesnt at least in my view support a position that he needs to coach or recruit differently. i realize this may be one of those debates that will exist on this board forever and i guess thats what makes it fun. thx again for all your cogent contributions.

sorry corpus meant to add just one more thought on this. what you've done essentially is to handicap the results based on expectation. and i realize you and others probably mean to do that. but i wld just ask in what other field of endeavor do we that? not even golf, except between duffers i guess. cerainly not in stocks. or academics. or any human endeavor i can think of. so why here except as a way to salve our collective dissapointment and/ or to find a bogeyman even if he is a coach most of us on both sides of this issue adore.

Big g, I appreciate your response and I assure you I am not offended. I do get a bit frustrated when I spend hours doing some pretty involved analysis and someone posts a response that is clearly not addressing my analysis and even worse, going beyond to imply things that I have not only not said but don't think. Believe me, the analysis is anything but tautological. I plead guilty to the sin of brevity in trying to lay the groundwork for my analysis but that would get very boring in the 2nd of about 30 paragraphs. So I try to explain it concisely for brevity. I assume if you are not a stats guy you move on and skip the post. If you are a stats guy you probably get by with the brevity. Again, I assure you that what you say is necessary s exactly what I have done. Fist, if getting a high seed makes it impossible to over perform, I can show you with the analysis that that is not only not true it is not the case in the actual outcomes. Icross checked my methods against every seed in every year. In fact, you could make a case that a high seed has a better chance to outperform. But the fact is that each game is assigned a probability based on several factors, the largest being actual tournament history. So for instance, this year Self just about broke even going three and one. That is not bad nor is it good. It is mediocre. Izzo lost as a 2 seed to a 15. That is bad. Getting to the final four actually puts you ahead for the year. Again, when you say "absolutely bound to happen given the amount of chance in a one and done tourney" that is exactly what my points system is meant to normalize against. "This is math, not opinion." I do not take offense at that but I did take it as you do not think my math is not right. Fine, I can send you the program and algorithm and you can test it out and show me where the math is wrong. But to say the math is clearly proof I am wrong is....well let's just say Idisagree.

"I'm more in the camp that the tournament is a new season and the coach shouldn't be penalized." 2 points here. I am fine with you or anyone else being in that camp. I am in that camp in many ways as well. I think each person should get their enjoyment from their fanship in the way that helps them enjoy. I am totally good with that and support it 100%. But again, that has nothing to do with my point. I am not implying that to accept the analysis as I have put forth means you have to somehow look with a jaundiced eye at the tourney or at Self. I love the guy and I love the tourney. I love that fact that he is among the top in the tourney in wins, win%, wins per year and getting to the tourney at all and especially with consistently high seeds. I love all those things. But again, they have nothing to do with my point. And I certainly do not believe he has failed in the tourney in some sort of general overarching sense. He has a specific statistic that puts him solidly in the mediocre category. If you or I or anyone else has a standard of something much closer to perfection then I guess you could classify this statistic as failing. But the analysis simply says he is mediocre at performing in the tournament as compared to expectations. And his being a consistently high seed poses a challenge for normalization that I believe I have met in my normalization algorithm. As to coaching or recruiting differently, your point is well taken and valid. Nothing in my analysis gives any meaningful input to his coaching or recruiting. Any comments I have made in that regard are fair game and I accept that any disagreement is perfectly acceptable and reasonable. I am hammering this point a little bit out of frustration for losing and my way of burning off steam is to go do some complex analysis. I appreciate your thoughtful and reasoned input. Yet I do want to make sure I have defensibly averred my point in the face of potential misunderstanding.

thx corpus, appreciate the clarification. i now accept that your model produces fair handicapped results. our difference ( if there is one) is therefore wholly philosophical and not answered by any statistics as you indicate. have a great off season

hi corpus, since you seem to have so much data in a massageable format, for those of us valuing consistency over all else it would be very interesting to see a rank ordering of all head coaches for yrs since the 64 team tournament started showing just number of years coached, number of tournament wins, and ranked on the third col which wld be tournament games won per yrs coached. feel free to resist, defer or ignore of course. just thought it wldnt hurt to ask. thanks!

Big g, right now my database goes back only to the 1985 tourney when the field expanded to 64 teams. I'm sure I can find the data back farther and get it downloaded. I am always happy to crunch any numbers. I used to be an engineer but since I have been sentenced to cease and desist the fun stuff some 25 years ago I get precious little opportunity aside from my hobby. I'll see what I can do. Thanks for asking.