Peaking For NCAAs: Which Coach Does It Best?
Peaking For NCAAs: Which Coach Does It Best?
By Andrew SpeyAsk any Division I NCAA wrestling coach, and they will tell you the college season is all about March. Winning at the NCAA tournament is the m
By Andrew Spey
Ask any Division I NCAA wrestling coach, and they will tell you the college season is all about March. Winning at the NCAA tournament is the most important goal for both the individual and the team. Coaches meticulously plan their year so that every wrestler on the team is at his maximum athletic potential at that point.
Fans across many sports commonly refer to this phenomenon as peaking. And while there is no doubt that athletes train in cycles to be in the best possible shape at particular times of the year, there is still debate as to whether some coaches have the ability to get more out of their athletes than others.
There are nine individuals who have been both head coach of a Division I program for the last 16 years, and shepherded no fewer than 100 wrestlers through the NCAA tournament during that timeframe. Those coaches were our subjects, and we examined the performance of their national qualifiers to test our hypothesis.
Test Subjects:
John Smith - Oklahoma State - 147 NQs
J Robinson - Minnesota - 137 NQs
Rob Koll - Cornell - 127 NQs
Joe McFarland - Michigan - 117 NQs
Mark Manning - Nebraska - 116 NQs
Brian Smith - Missouri - 115 NQs
Tim Flynn - Edinboro - 112 NQs
Tom Borrelli - Central Michigan - 109 NQs
Barry Davis - Wisconsin - 100 NQs
No seeding is perfect, but with a large enough sample size, those over-seeded should average out with those under-seeded; thus, our analysis should not be adversely affected by seeding. Our analysis ignored higher seeds, as examining the results of national qualifiers receiving the highest seeds would have influenced biased results. With no place left to go but down, coaches with a preponderance of top-seeded wrestlers would be punished with a disproportionate number of "negative" results.
A chart of all our test subjects’ programs and the number of wrestlers who received 12-9 seeds over the last 16 years can be found below.
Next, we filtered the results by NCAA finish, and assigned point values to each wrestler’s specific placement on the following scale:
A round of 12 finish would be "wrestling to seed," and worth zero points on our scale. Three negative points are assigned to each wrestler eliminated in the first consolation round; minus two for a round of 24 finish; and so on.
This is somewhat arbitrary, but the results are more or less the same no matter what numbers you use as long as they scale, and we prefer to have point values that resemble the NCAA tournament team scoring system.
There is little difference to a team if a wrestler finishes in the round of 24 versus the round of 16; however, the point differential between an eighth and first-place finish is quite significant.
In any event, it’s the scale we’re using, and debate on the topic has been prorogued for now.
Or for those who prefer a more comprehensive table, the entire placement chart is below.
So, case closed? JRob and John Smith are the best at peaking? Well, not quite. While we now know whose wrestlers with 12-9 seeds most overperformed over the last 16 years, we still at least need to compare our test subjects to the general population.
If the range of scores for all programs over the same timeframe is greater than -2 to +14, then it is likely that the peaking ability of our subject coaches can be explained mostly by the naturally occurring random variation of scores in the program population.
Retesting the Data
To make this comparison, we can run the same point-scoring exercise with every NCAA program that had any wrestlers seeded 12-9 in the past 16 tournaments.
Those results can then be plotted on a scatter chart to see where our subject coaches with positional longevity compare with regards to other programs of both lesser and equal caliber. That chart is below.
The total number of 12-9 seeded wrestlers over the last 16 years is plotted on the horizontal X axis and the total ‘score’ of the programs is on the vertical Y axis.
As we can see, four programs (PSU, Ohio State, Iowa and Illinois) all scored higher under the tutelage of multiple coaches than did our two highest scoring, continuously coached programs over the same period of time.
There are also a host of programs that scored lower than our lowest scoring long serving coaches. This would indicate that the natural range of results comfortably envelopes all of our test subjects.
This is partly due to the nature of our scoring rubric. If we had assigned greater negative point values for rounds of 32, 24 and 16 finishes, and did not adjust our scores for All American finishes, our trendline would be flatter, and our scatter chart would appear even more random than it currently does.
It is also a possible indicator that the stronger a program, the more likely a wrestler from that program is to overperform his seed. As this is a study of coaches, and not programs, the exploration of that thread will have to wait for another time.
We must also keep in mind that our initial results point only to a correlation, and not a causal relationship, between coaches and NCAA success (or failure). That is, to confirm that coaches have a skill known as peaking, and that they are able to apply that skill on a consistent basis, we must demonstrate that the skill has predictive power.
To provide such a demonstration, we could take our recent results and see if the trends continue. If, say, over the next five years, JRob (assuming he continues coaching, a topic we’ll skip for now) and John Smith, produce positive ‘scores’, and if coaches Koll, McFarland and Smith, produce ‘negative’ scores, we would have slightly more ammunition to support our hypothesis of the existence of a peaking skill for coaches.
But it would be flimsy buttressing at best, and that evidence is 5 years away, at any rate. However, before we give up on our hypothesis entirely, there is at least one more test we can do.
As expected, it follows the same pattern as our ‘scoring’ analysis. If this data represents the coaches’ peaking skill, then we should expect to see similar results when we look at national qualifying wrestlers seeded 5th, 6th, 7th & 8th from those same programs that do NOT attain All American status.
This will produce a ‘Failure Rate’ (for lack of a better term). Those results are below, displayed as: school - total number of 5-8 seeds - non-AA from those seeds - ‘Failure Rate’.
Our results do not pan out as expected, or perhaps they do, depending on your point of view. What they do not show is a correlation between ‘Success Rate’ and ‘Failure Rate’.
For peaking to be a skill, we would expect coaches that posses it to be as adept at getting their wrestlers to over-perform as they are at keeping their wrestlers from under-performing.
The evidence in this test is a clear rejection of our hypothesis that peaking is a coaching skill.
For instance, Mark Manning does a commendable job getting fully one third of his 12-9 seeded wrestlers on to the podium, good for 3rd best among our test subjects. Conversely, he is at the bottom of our rankings with regards to ensuring his 8-5 seeded wrestlers do likewise.
For another look at the lack of correlation between success and failure rates, below are the schools ranked by those measures, with the higher success rates and lower failure rates at the top of the chart.
For while there are some factors within the control of the coach, e.g. how well he runs his practices and teaches technique, there are other, arguably much more important factors, that are primarily the onus of the wrestler himself, such as his natural talent level and how much work he puts into to achieving his goal of becoming an All American.
And then there are the factors beyond anyone’s control, such as injury defaults and forfeits in the brackets or suffering from food poisoning the day before the tournament.
Those and countless other factors make it difficult to find evidence for a coach having the discernable skill of getting his wrestlers to peak at the NCAA tournament. Like clutch hitting in baseball, it is easy to identify after the fact, but nigh impossible to predict who posses the skill before the competition begins.
There are also undoubtedly other metrics which can be used to test the same hypothesis. There is likely more to be gleaned by digging deeper into recent NCAA results and other sets of data. All suggestions regarding further areas of exploration are most welcome.
But Wait, What About Cael and Tom?
Excellent question. One of the unfortunate limitations of only considering head coaches who have been in charge over the last 16 consecutive seasons is that we miss out on two superstars of the wrestling universe, Cael Sanderson and Tom Brands.
As no true study on current NCAA wrestling coaches would be complete without them, we will change the parameters of the analysis for a subsequent article to make sure we consider the skippers of arguably the two biggest programs in college, Penn State and Iowa.
Stay tuned for the follow up, coming your way soon, with 100% more Sanderson and Brands.
Popular Right Now:
California Bringing A Hammer Squad To Junior Duals
VIDEO: Spencer Lee And Gavin Teasdale Drilling
BJ Futrell Defeats Olympic Champion
Video: John Smith Teaches Signature Leg Lace
Ask any Division I NCAA wrestling coach, and they will tell you the college season is all about March. Winning at the NCAA tournament is the most important goal for both the individual and the team. Coaches meticulously plan their year so that every wrestler on the team is at his maximum athletic potential at that point.
Fans across many sports commonly refer to this phenomenon as peaking. And while there is no doubt that athletes train in cycles to be in the best possible shape at particular times of the year, there is still debate as to whether some coaches have the ability to get more out of their athletes than others.
The Hypothesis
We suspect certain Division I NCAA wrestling coaches help their wrestlers perform better at the NCAA tournament than other coaches. Our analysis hopes to ascertain which coaches command this mastery and exercise it most effectively.The Data
Our study looked at coaches who have enjoyed both success at NCAAs and long tenures at a single school. We wanted as large a pool of data as possible to gird against small sample size, and we needed to keep the head coach and programs consistent to discern the coaches’ hypothetical skill.There are nine individuals who have been both head coach of a Division I program for the last 16 years, and shepherded no fewer than 100 wrestlers through the NCAA tournament during that timeframe. Those coaches were our subjects, and we examined the performance of their national qualifiers to test our hypothesis.
Test Subjects:
John Smith - Oklahoma State - 147 NQs
J Robinson - Minnesota - 137 NQs
Rob Koll - Cornell - 127 NQs
Joe McFarland - Michigan - 117 NQs
Mark Manning - Nebraska - 116 NQs
Brian Smith - Missouri - 115 NQs
Tim Flynn - Edinboro - 112 NQs
Tom Borrelli - Central Michigan - 109 NQs
Barry Davis - Wisconsin - 100 NQs
The Analysis
To see how well coaches are able to extract peak performances from their wrestlers, we examined the results of their national qualifiers seeded 12th, 11th, 10th or ninth at the last 16 NCAA tournaments. Competitors seeded at that level are theoretically All-American contenders, but not favorites.No seeding is perfect, but with a large enough sample size, those over-seeded should average out with those under-seeded; thus, our analysis should not be adversely affected by seeding. Our analysis ignored higher seeds, as examining the results of national qualifiers receiving the highest seeds would have influenced biased results. With no place left to go but down, coaches with a preponderance of top-seeded wrestlers would be punished with a disproportionate number of "negative" results.
A chart of all our test subjects’ programs and the number of wrestlers who received 12-9 seeds over the last 16 years can be found below.
Next, we filtered the results by NCAA finish, and assigned point values to each wrestler’s specific placement on the following scale:
| R32: | -3 |
|---|---|
| R24: | -2 |
| R16: | -1 |
| R12: | 0 |
| 8th: | 1 |
| 7th: | 2 |
| 6th: | 3 |
| 5th: | 4 |
| 4th: | 5 |
| 3rd: | 6 |
| 2nd: | 7 |
| 1st: | 8 |
This is somewhat arbitrary, but the results are more or less the same no matter what numbers you use as long as they scale, and we prefer to have point values that resemble the NCAA tournament team scoring system.
There is little difference to a team if a wrestler finishes in the round of 24 versus the round of 16; however, the point differential between an eighth and first-place finish is quite significant.
In any event, it’s the scale we’re using, and debate on the topic has been prorogued for now.
The Findings
The resulting ‘scores’ of each programs’ wrestlers seeded 12 through 9 over the last 16 years are below.| Team | Score |
|---|---|
| Minnesota | 14 |
| Oklahoma State | 13 |
| Nebraska | 9 |
| Edinboro | 6 |
| Central Michigan | 1 |
| Wisconsin | 1 |
| Cornell | -1 |
| Michigan | -1 |
| Missouri | -2 |
So, case closed? JRob and John Smith are the best at peaking? Well, not quite. While we now know whose wrestlers with 12-9 seeds most overperformed over the last 16 years, we still at least need to compare our test subjects to the general population.
If the range of scores for all programs over the same timeframe is greater than -2 to +14, then it is likely that the peaking ability of our subject coaches can be explained mostly by the naturally occurring random variation of scores in the program population.
Retesting the Data
To make this comparison, we can run the same point-scoring exercise with every NCAA program that had any wrestlers seeded 12-9 in the past 16 tournaments.
Those results can then be plotted on a scatter chart to see where our subject coaches with positional longevity compare with regards to other programs of both lesser and equal caliber. That chart is below.
The total number of 12-9 seeded wrestlers over the last 16 years is plotted on the horizontal X axis and the total ‘score’ of the programs is on the vertical Y axis. As we can see, four programs (PSU, Ohio State, Iowa and Illinois) all scored higher under the tutelage of multiple coaches than did our two highest scoring, continuously coached programs over the same period of time.
There are also a host of programs that scored lower than our lowest scoring long serving coaches. This would indicate that the natural range of results comfortably envelopes all of our test subjects.
Preliminary Conclusions
There is a clear, upwardly sloping trendline, indicating that programs with more total wrestlers seeded 12 through 9 in the NCAA tournament typically score higher using our formula than schools with fewer wrestlers of the same seeds over the same timeframe.This is partly due to the nature of our scoring rubric. If we had assigned greater negative point values for rounds of 32, 24 and 16 finishes, and did not adjust our scores for All American finishes, our trendline would be flatter, and our scatter chart would appear even more random than it currently does.
It is also a possible indicator that the stronger a program, the more likely a wrestler from that program is to overperform his seed. As this is a study of coaches, and not programs, the exploration of that thread will have to wait for another time.
We must also keep in mind that our initial results point only to a correlation, and not a causal relationship, between coaches and NCAA success (or failure). That is, to confirm that coaches have a skill known as peaking, and that they are able to apply that skill on a consistent basis, we must demonstrate that the skill has predictive power.
To provide such a demonstration, we could take our recent results and see if the trends continue. If, say, over the next five years, JRob (assuming he continues coaching, a topic we’ll skip for now) and John Smith, produce positive ‘scores’, and if coaches Koll, McFarland and Smith, produce ‘negative’ scores, we would have slightly more ammunition to support our hypothesis of the existence of a peaking skill for coaches.
But it would be flimsy buttressing at best, and that evidence is 5 years away, at any rate. However, before we give up on our hypothesis entirely, there is at least one more test we can do.
Additional Analysis
By reexamining the 12-9 seeds of our long tenured coaches, and dividing the number of 12-9 seeds that reach All American status by the total number of 12-9 seeds over the last 16 years, we can calculate a ‘Success Rate’. The results are below, showing: school - total number of 12-9 seeds - AA by those seeds - ‘Success Rate’.| 'Success Rate' | |||
|---|---|---|---|
| Minnesota | 22 | 11 | 50.00% |
| Oklahoma State | 28 | 10 | 35.70% |
| Nebraska | 21 | 7 | 33.30% |
| Central Michigan | 25 | 8 | 32.00% |
| Cornell | 14 | 4 | 28.60% |
| Michigan | 18 | 5 | 27.80% |
| Missouri | 17 | 4 | 23.50% |
| Edinboro | 13 | 3 | 23.10% |
| Wisconsin | 13 | 2 | 15.40% |
This will produce a ‘Failure Rate’ (for lack of a better term). Those results are below, displayed as: school - total number of 5-8 seeds - non-AA from those seeds - ‘Failure Rate’.
| 'Failure Rate' | |||
|---|---|---|---|
| Cornell | 20 | 2 | 10.00% |
| Central Michigan | 13 | 2 | 15.40% |
| Edinboro | 8 | 2 | 25.00% |
| Minnesota | 41 | 12 | 29.30% |
| Missouri | 23 | 7 | 30.40% |
| Michigan | 23 | 8 | 34.80% |
| Wisconsin | 17 | 6 | 35.30% |
| Oklahoma State | 23 | 9 | 39.10% |
| Nebraska | 22 | 15 | 68.20% |
Our results do not pan out as expected, or perhaps they do, depending on your point of view. What they do not show is a correlation between ‘Success Rate’ and ‘Failure Rate’. For peaking to be a skill, we would expect coaches that posses it to be as adept at getting their wrestlers to over-perform as they are at keeping their wrestlers from under-performing.
The evidence in this test is a clear rejection of our hypothesis that peaking is a coaching skill.
For instance, Mark Manning does a commendable job getting fully one third of his 12-9 seeded wrestlers on to the podium, good for 3rd best among our test subjects. Conversely, he is at the bottom of our rankings with regards to ensuring his 8-5 seeded wrestlers do likewise.
For another look at the lack of correlation between success and failure rates, below are the schools ranked by those measures, with the higher success rates and lower failure rates at the top of the chart.
Final Conclusion
To explain the lack of evidence supporting our hypothesis, we can consider the myriad factors that determine the final placement of a wrestler at the NCAA tournament.For while there are some factors within the control of the coach, e.g. how well he runs his practices and teaches technique, there are other, arguably much more important factors, that are primarily the onus of the wrestler himself, such as his natural talent level and how much work he puts into to achieving his goal of becoming an All American.
And then there are the factors beyond anyone’s control, such as injury defaults and forfeits in the brackets or suffering from food poisoning the day before the tournament.
Those and countless other factors make it difficult to find evidence for a coach having the discernable skill of getting his wrestlers to peak at the NCAA tournament. Like clutch hitting in baseball, it is easy to identify after the fact, but nigh impossible to predict who posses the skill before the competition begins.
Further Study
There is potentially additional insight to be gained by shortening the timeframe of the study, which would permit the casting of a wider net of D1 NCAA coaches. For example, reducing the length of tenure to 7 years would allow for the inclusion of Cael Sanderson and Tom Brands, among others.There are also undoubtedly other metrics which can be used to test the same hypothesis. There is likely more to be gleaned by digging deeper into recent NCAA results and other sets of data. All suggestions regarding further areas of exploration are most welcome.
The Bottom Line
While we will always be able to determine if a team was coached into delivering a peak performance after an NCAA tournament has concluded, ascertaining if a coach will be able to generate such a performance before the tournament is wrestled so far eludes our grasp.But Wait, What About Cael and Tom?
Excellent question. One of the unfortunate limitations of only considering head coaches who have been in charge over the last 16 consecutive seasons is that we miss out on two superstars of the wrestling universe, Cael Sanderson and Tom Brands.
As no true study on current NCAA wrestling coaches would be complete without them, we will change the parameters of the analysis for a subsequent article to make sure we consider the skippers of arguably the two biggest programs in college, Penn State and Iowa.
Stay tuned for the follow up, coming your way soon, with 100% more Sanderson and Brands.
Popular Right Now:
California Bringing A Hammer Squad To Junior Duals
VIDEO: Spencer Lee And Gavin Teasdale Drilling
BJ Futrell Defeats Olympic Champion
Video: John Smith Teaches Signature Leg Lace