In last week's post about coaching success in March, I primarily focused on the concept of performance vs. expectation. While I certainly feel that this is an important metric, at the end of the day, what really matters is winning games, advancing in the tournament, and cutting down nets. With this is mind, I thought that it would be fair to take a look at a few other measures of coaching success in March. Specifically, I wanted to look at overall winning percentage as well as the winning percentages as the favorite and as the underdog. In addition, I also wanted to take a look at the number of appearances per round. Perhaps more interestingly, I finally wanted to analyze the rate (appearances per attempt) of each coach advancing to a particular round. In order to keep the data set a bit more manageable, I will just report the data for active coaches with 10 tournament wins or more and all coaches with 20 tournament wins or more in the post-1979 era. The table with all this data summarized is shown below:
As a first pass, I sorted the table based on overall winning percentages, and the current standing are not much of a surprise. Coach K and Roy Williams (with their 8 titles between them) currently sit 1-2 in the rankings. Calipari, Billy Donovan, Dean Smith, and Pitino are not far behind. They all have winning percentages above 74%. That is notable because in any given tournament, if you win 75% of your games, you have at least advanced to the Regional Final. In effect, these coaches all average making it to the Regional Final, which is consistent with the fact that all 6 of these elite coaches make it to the Regional Final between 44% and 58% of the time.
In the next batch of coaches in the table, you find Jim Calhoun, Bill Self, Tom Izzo, and Jerry Tarkanian, all of which having winning percentages above 70%, just below that Elite 8 average performance. Slightly below that are the likes of Denny Crum, Jay Wright, John Beilein, Frank Martin, John Thompson, and Jim Boeheim, all who clock in at a winning percentage of around 66.7% or better. The 66.7% mark is notable, as this marks the line where a coach averages a Sweet 16 bid. As for the other 29 coaches in the table, they lie below this threshold.
After the overall winning percentage, I think that it is interesting to look at two additional metrics: each coach's winning percentage both as a favorite and as an underdog. That data is also shown in the table above. From the table, it is clear that the majority of the elite level coaches all win between 80 and 90% of their games when they are favored. Roy Williams, Pitino, Calhoun, Calipari, Tom Izzo (ranked 8th), Coach K, and Jay Wright all fall into this category. Interestingly, the top ranked coach in the table based on this metric is John Beilein. His current record as a favored seed of 17-2 (89.5%) is only behind Massmino (7-0), Brad Stevens (7-0), and P.J. Carlisimo (9-1) for all coaches with more than 4 games as a favorite. For as much as I like to poke fun at our friends from Ann Arbor, Beilein’s record as a favorite is truly remarkable. His only loses were to Ohio in 2012 in the 4-13 match-up and 8-seed Kentucky in the 2014 Regional Final, and clearly there is no shame in that. He basically never loses as the favorite.
As a first pass, I sorted the table based on overall winning percentages, and the current standing are not much of a surprise. Coach K and Roy Williams (with their 8 titles between them) currently sit 1-2 in the rankings. Calipari, Billy Donovan, Dean Smith, and Pitino are not far behind. They all have winning percentages above 74%. That is notable because in any given tournament, if you win 75% of your games, you have at least advanced to the Regional Final. In effect, these coaches all average making it to the Regional Final, which is consistent with the fact that all 6 of these elite coaches make it to the Regional Final between 44% and 58% of the time.
In the next batch of coaches in the table, you find Jim Calhoun, Bill Self, Tom Izzo, and Jerry Tarkanian, all of which having winning percentages above 70%, just below that Elite 8 average performance. Slightly below that are the likes of Denny Crum, Jay Wright, John Beilein, Frank Martin, John Thompson, and Jim Boeheim, all who clock in at a winning percentage of around 66.7% or better. The 66.7% mark is notable, as this marks the line where a coach averages a Sweet 16 bid. As for the other 29 coaches in the table, they lie below this threshold.
After the overall winning percentage, I think that it is interesting to look at two additional metrics: each coach's winning percentage both as a favorite and as an underdog. That data is also shown in the table above. From the table, it is clear that the majority of the elite level coaches all win between 80 and 90% of their games when they are favored. Roy Williams, Pitino, Calhoun, Calipari, Tom Izzo (ranked 8th), Coach K, and Jay Wright all fall into this category. Interestingly, the top ranked coach in the table based on this metric is John Beilein. His current record as a favored seed of 17-2 (89.5%) is only behind Massmino (7-0), Brad Stevens (7-0), and P.J. Carlisimo (9-1) for all coaches with more than 4 games as a favorite. For as much as I like to poke fun at our friends from Ann Arbor, Beilein’s record as a favorite is truly remarkable. His only loses were to Ohio in 2012 in the 4-13 match-up and 8-seed Kentucky in the 2014 Regional Final, and clearly there is no shame in that. He basically never loses as the favorite.
The compliment to the coach’s record as a favorite is
their record as an underdog. In the table above, there are only 3 coaches with winning records as the underdog. Billy Donovan and Leonard Hamilton are both 6-3 (66.7%) as underdogs, but the sample size of nine games is relatively small. In third place is Tom Izzo will a stunning 14-13 (52%) record as the underdog, and he still holds the record for the most wins as the lower seed in tournament history (14). Jim Beoheim is only one win behind with 13, but is overall sub 500 at 13-15 (46.4%). No other active coach has more than 7 wins as an underdog (Beilein, Few, and Gregg Marshall), and only Lute Olson (11) and Rollie Massimino (11) have more than 10 total in history.
That all said, the real measure of a coach'e prowess in March, one could and perhaps should argue, is advancement in the tournament itself. As such, the final block of data in the table above shows first the total number of appearances in a given round for each coach, with the rate of appearance shown in parenthesis. For example, Coach K has advanced to 24 Sweet 16s in his 34 tournament appearances (24/34 = 71%). In order to aid in the visualization of this data, I prepared a set of histograms for appearance rate data for each round. As an added bonus, I also made a histogram for the number of times a coach gets upset in the first round, which is more interesting than simply the rate of making it to the 2nd round. Those histograms are shown below.
In the case of the 1st round upset frequency, essentially all of the modern elite coaches get upset less than 20% of the time, and some (Roy Williams, Jay Wright, and Brad Stevens) have actually never been upset in the 1st round. Not surprisingly, some of the coaches with a bit of a reputation for choking in March (Sean Miller, Lute Olson, Gene Keady, and Rick Barnes) are all found to lose at a clip greater than 20%. Perhaps the most surprising number are Bob Knights 39% clip in the modern era and Kevin Stallings shocking 50% rate. Tom's Izzo' rate of 18% (3/17) first round upsets is slightly high relative to some other coaches, but is in the most common bin in the histogram.
At the end of the day, the only prize that really matters is the National Title, and by this metric the truly elite coaches are Coach K, Billy Donovan, Jay Wright, and Jim Calhoun, all of whom have multiple titles and all of whom have won a title in 13-14% of their tournament appearances. In the batch after that are coaches who win titles 9-11% of the time, including Denny Crum, Roy Williams, and Rick Pitino. Tom Izzo is back in the pack who have won the title in 4-6% of their tournament appearances, including Bill Self, Calipari, Lute Olson, and Jim Boeheim. But, as we all know, winning the whole thing is tough, and this chart is by far the one that is influenced more by individual plays. For example, if one were to eliminate the Villanova buzzer beater in the 2016 title game, Jay Wright and Roy Williams would essentially trade spaces on this chart. So, maybe it is enough just to be on this chart in the first place. That seems fair to me.
As an added bonus, it occurred to me that it should be possible to identify a subset of elite coaches based on the histograms above. I decided to focus on coaches with performance rates in the top 10% of each category above. The data set of coaches with 10+ tournament wins is about 130 coaches, so in each case I pull the top 13-15 coaches. Based on this analysis, it seems like the best cut-off for elite status is being in the Top 10 in at least 3 categories. The following table summarizes this result:
Based on this analysis, there is a clear Top 11 coaches of all time (in the modern era) that can be considered elite, and the list seems to have every major coach. Three total coaches (Dean Smith, Roy Williams, and Rick Pitino) scored in the Top 10 of all five categories, while additional coaches (Tarkanian, Calipari, and Coach K) appeared in 4 of 5. The remaining five coaches (Jay Wright, Izzo, Crum, Donovan, and Larry Brown) round out the Top11.
That all said, the real measure of a coach'e prowess in March, one could and perhaps should argue, is advancement in the tournament itself. As such, the final block of data in the table above shows first the total number of appearances in a given round for each coach, with the rate of appearance shown in parenthesis. For example, Coach K has advanced to 24 Sweet 16s in his 34 tournament appearances (24/34 = 71%). In order to aid in the visualization of this data, I prepared a set of histograms for appearance rate data for each round. As an added bonus, I also made a histogram for the number of times a coach gets upset in the first round, which is more interesting than simply the rate of making it to the 2nd round. Those histograms are shown below.
As for the Sweet 16 appearance frequency, the distribution is a bit more bell-curve shaped. Dean Smith is the champion of this specific analysis, having made 15 Sweet Sixteen in 19 attempts in the modern era (79%). Not far behind are Coach K and John Calipari, who are both over 70%. Tom Izzo's rate of 62% is good enough for 6th place among active coaches, although interestingly Sean Miller's rate (64%) is slightly higher.
Regarding the Regional Final round, here John Calpari and Rick Pitino set the pace at 58% and 57%, respectively, while Bill Self and Billy Donovan are close behind at 50%. If nothing else, this is an indication of the overall success that all four of those coaches have in the Round of 16. In contrast, Jim Boeheim's rate drops to a mere 23% in this round due to his historically poor performance in the Sweet 16. Tom Izzo's rate of 43% is certainly very strong and is in fact good enough for 5th place among active coaches.
As for the Final Four, Larry Brown's 3 Final Fours in only 8 tournaments leads the pack at 37.5%, but Coach K (35%), Izzo (33%), Pitino (33%), Roy Williams (32%), and John Calipari (32%) are all right there at the top of the list, and they all have coached in at least twice as many tournaments. The mere fact that all five of those elite coaches have such similar statistics is notable itself. Most other Final Four coaches are lucky to reach a Final Four 20% of the time, so this is a clear demarcation of the truly elite.
At the end of the day, the only prize that really matters is the National Title, and by this metric the truly elite coaches are Coach K, Billy Donovan, Jay Wright, and Jim Calhoun, all of whom have multiple titles and all of whom have won a title in 13-14% of their tournament appearances. In the batch after that are coaches who win titles 9-11% of the time, including Denny Crum, Roy Williams, and Rick Pitino. Tom Izzo is back in the pack who have won the title in 4-6% of their tournament appearances, including Bill Self, Calipari, Lute Olson, and Jim Boeheim. But, as we all know, winning the whole thing is tough, and this chart is by far the one that is influenced more by individual plays. For example, if one were to eliminate the Villanova buzzer beater in the 2016 title game, Jay Wright and Roy Williams would essentially trade spaces on this chart. So, maybe it is enough just to be on this chart in the first place. That seems fair to me.
As an added bonus, it occurred to me that it should be possible to identify a subset of elite coaches based on the histograms above. I decided to focus on coaches with performance rates in the top 10% of each category above. The data set of coaches with 10+ tournament wins is about 130 coaches, so in each case I pull the top 13-15 coaches. Based on this analysis, it seems like the best cut-off for elite status is being in the Top 10 in at least 3 categories. The following table summarizes this result:
Based on this analysis, there is a clear Top 11 coaches of all time (in the modern era) that can be considered elite, and the list seems to have every major coach. Three total coaches (Dean Smith, Roy Williams, and Rick Pitino) scored in the Top 10 of all five categories, while additional coaches (Tarkanian, Calipari, and Coach K) appeared in 4 of 5. The remaining five coaches (Jay Wright, Izzo, Crum, Donovan, and Larry Brown) round out the Top11.
Comments
Post a Comment