For baseball analytics professionals, the fact that a baseball manager has less impact on his team’s chance of winning has been a notion gaining traction for a number of years. A number of high-profile sabermetricians have attempted to quantify the value of baseball managers, but research thus far suggests little connection between the success of a team and the on-field actions of its manager. In James Click’s Baseball Between the Numbers, the author notes that even the best managers often wind up costing their team victories by the end of the season.
What impact does this have for college baseball? Should a team that changes managers—regardless of the reason—expect to experience a notable change in performance?
Taking a look at a three-year sample set—in this case D-I schools that made a coaching change in the summers of 2009, 2010, or 2011—we arrive at 35 separate cases. Setting aside for a moment the particular circumstances of the changes, we can begin by simply taking a look at the average change in RPI from the last year with the previous manager to the first year with the newly hired manager. In the year following a coaching change, the average change in RPI on the preceding year turns out to be a mere minus 1.34—essentially no change on the preceding year.
One might note that a new manager simply needs more than a single year to “change the culture,” or instill his values (though a baseball manager isn’t installing a zone defense or teaching the press). However, looking at year-to-year performance change should offer some degree of insight into what a baseball manager is capable of effecting in a talent group similar to that of his predecessor. Furthermore, in the second year after hiring a new coach the change in RPI goes up to plus 8.8, an improvement, although still less than a 3% change in a pool of 300 teams.
All manager changes are not created equal, however. Whereas some schools may look to maintain a stable salary level with a new hire, other schools might look to significantly increase, or contrarily, to slash salary expenditures. Perhaps a correlation between the change in salary and the change in RPI would shed some light on program direction. If we call maintaining salary level to be a range of roughly +/- $10,000, a decrease being more than – $10,000, and an increase as more than + $10,000, we have three groups for comparison.
Key:
Decrease | |
Maintain | |
Increase |
Salary Analysis:
School | Last Yr Salary | First Yr Salary | Salary Diff | Year |
Cal St Fullerton | 3—,384 | 1—,184 | -1—,200 | 2011 |
Arizona St | 2—,424 | 1—,000 | -1—,424 | 2009 |
Nebraska | 3—,122 | 1—,000 | -1—,122 | 2011 |
Alabama | 3—,000 | 2—,000 | -1—,000 | 2009 |
Houston | 2—,000 | 1—,000 | -6—,000 | 2010 |
Ohio St | 3—,000 | 3—,100 | -5—,900 | 2010 |
Cal St Long Beach | 1—,000 | 9—,708 | -5—,292 | 2010 |
Old Dominion 1 | 8—,486 | 6—,000 | -2—,486 | 2010 |
Maryland Baltimore | 6—,806 | 4—,327 | -2—,479 | 2011 |
S. Illinois | 6—,228 | 5—,697 | -1—,531 | 2010 |
Cal St Northridge | 1—,764 | 1—,000 | -1—,764 | 2010 |
Ball St | 7—,999 | 6—,500 | -1—,499 | 2010 |
Air Force | 9—,971 | 8—,338 | -1—,633 | 2010 |
NC Central | 6—,000 | 5—,500 | —-,500 | 2011 |
Southern Miss | 1—,500 | 1—,116 | —-,384 | 2009 |
Nicholls St | 5—,000 | 5—,000 | — | 2010 |
Sam Houston St | 9—,594 | 9—,000 | — | 2011 |
Akron | 6—,230 | 7—,000 | — | 2011 |
W. Michigan | 6—,022 | 7—,500 | —,478 | 2010 |
Winthrop | 9—,676 | 1—,000 | —,324 | 2010 |
S. Dakota St | 5—,000 | 6—,000 | —,000 | 2011 |
Cal State Sacramento | 1—,056 | 1—,100 | —,044 | 2010 |
Central Kansas | 5—,500 | 6—,500 | —,000 | 2010 |
North Carolina A&T | 4—,695 | 5—,350 | —,655 | 2011 |
New Jersey Tech | 4—,434 | 6—,000 | 1—,566 | 2010 |
W. Illinois | 3—,770 | 5—,803 | 1—,033 | 2009 |
W. Kentucky | 1—,416 | 1—,000 | 2—,584 | 2011 |
Illinois St | 7—,388 | 1—,000 | 2—,612 | 2009 |
S. Alabama | 1—,713 | 1—,200 | 2—,487 | 2011 |
UNLV | 8—,788 | 1—,000 | 2—,212 | 2010 |
N. Florida | 7—,713 | 1—,500 | 3—,787 | 2011 |
UC Davis | 8—,324 | 1—,822 | 5—,498 | 2011 |
Washington | 1—,946 | 1—,008 | 7—,062 | 2009 |
Old Dominion 2 | 6—,000 | 1—,000 | 9—,000 | 2011 |
Tennessee | 3—,128 | 4—,000 | 1—,872 | 2011 |
RPI Analysis:
School | Last Yr RPI | First Yr RPI | RPI Diff | Year |
Cal St Fullerton | 15 | 20 | -5 | 2011 |
Arizona St | 6 | 1 | 5 | 2009 |
Nebraska | 69 | 92 | -23 | 2011 |
Alabama | 39 | 11 | 28 | 2009 |
Houston | 84 | 63 | 21 | 2010 |
Ohio St | 99 | 148 | -49 | 2010 |
Cal St Long Beach | 95 | 65 | 30 | 2010 |
Old Dominion 1 | 221 | 127 | 94 | 2010 |
Maryland Baltimore | 288 | 289 | -1 | 2011 |
S. Illinois | 168 | 157 | 11 | 2010 |
Cal St Northridge | 121 | 219 | -98 | 2010 |
Ball St | 119 | 238 | -119 | 2010 |
Air Force | 248 | 259 | -11 | 2010 |
NC Central | 300 | 280 | 20 | 2011 |
Southern Miss | 31 | 61 | -30 | 2009 |
Nicholls St | 151 | 150 | -1 | 2010 |
Sam Houston St | 78 | 49 | 29 | 2011 |
Akron | 257 | 275 | -18 | 2011 |
W. Michigan | 272 | 189 | 83 | 2010 |
Winthrop | 98 | 136 | -38 | 2010 |
S. Dakota St | 159 | 276 | -117 | 2011 |
Cal State Sacramento | 223 | 234 | -11 | 2010 |
Central Kansas | 215 | 162 | 53 | 2010 |
North Carolina A&T | 284 | 278 | 6 | 2011 |
New Jersey Tech | 289 | 293 | -4 | 2010 |
W. Illinois | 280 | 282 | -2 | 2009 |
W. Kentucky | 88 | 171 | -83 | 2011 |
Illinois St | 170 | 82 | 88 | 2009 |
S. Alabama | 99 | 158 | -59 | 2011 |
UNLV | 208 | 87 | 121 | 2010 |
N. Florida | 112 | 109 | 3 | 2011 |
UC Davis | 211 | 160 | 51 | 2011 |
Washington | 133 | 68 | 65 | 2009 |
Old Dominion 2 | 127 | 223 | -96 | 2011 |
Tennessee | 116 | 120 | -4 | 2011 |
There were a total of 12 schools that reduced salary, 12 schools that maintained salary, and 11 schools that increased salary. The average change in RPI for these three groups was +.013, -14.5, and +7.27 respectively. Each group included either one or two outlying cases which significantly altered the average. 1
In removing the two outliers from our reduced salary pool, we now find the average RPI change to be +11.58. Likewise, removing the one outlier from the maintained salary pool—as well as the two outliers from the increased salary pool—yields average RPI changes of -7 and -14.28 respectively. Moving into the second year the RPI differentials are positive, but still modest, with reduced, maintained, and increased salary pools yielding +4.8, +5.25, and +17, respectively.
Essentially, there does not seem to be a significant connection between either an increase or a decrease in salary expenditures and the resulting performance of the baseball team. If there is less gain in terms of RPI by increasing salary expenditures, then what might be the best way to maintain or reduce salary in the event of change? Perhaps this data can inform the process of making a coaching change—we can examine the hidden value of hiring talented assistant coaches, rather than making splash with an established hire.
In the summer of 2011 one of the most successful baseball programs in the country, Cal St Fullerton, found themselves in need of a new baseball manager. Although head coach Dave Serrano was handsomely compensated, a larger athletic department was able to hire him away. The University of Tennessee, fresh off of facilities renovations, hired Serrano from Fullerton and significantly increased his salary.
Coming off a successful 2010-2011 season that saw Fullerton finish with an RPI of 15, the inclination must have surely been present to hire a proven head coach in order to prevent a backslide. Fullerton, however, chose to move in a different direction. Rick Vanderhook had been a lifelong assistant college baseball coach, and had spent the previous year as the associate head coach at UCLA. Fullerton was able to sign Vanderhook to a salary of $146,184—a $184,200 decrease on the previous year’s salary expenditure. For revenues of approximately $1.05 million, that represented a savings of 17.5% in the salary of the head coach alone.
Despite hiring an assistant coach and slashing salary, Fullerton maintained their RPI standing in the 2011-2012 season, finishing at 20. In this past year the team once again performed well, finishing with an RPI of 6. The coaching change did not significantly alter the on-field performance of the team.
Indeed targeting assistant coaches was a common theme for schools that maintained or reduced salary during a coaching change, as only 3 schools out of 24 chose to hire an established head coach. Conversely, 6 of the 11 schools that increased salary did hire head coaches. While there obviously is nothing wrong in hiring the known quantity of an established head coach, it would appear that there is also much to be gained in giving some attention to the assistant coaching pool during a personnel turnover.
On average, an assistant coach makes a lower salary than the average head coach. The average salary for a DI head coach is $160, 679, whereas the average salary for an assistant (or associate) coach is a mere $55,975. That indicates that salary considerations will generally start at a lower tier when hiring a new head coach whose previous experience only entails assistant positions. Additionally, the pool of available assistant coaches is simply much larger than that of head coaches. There are 299 DI head coaches versus 1024 assistant (and associate) coaches. More choices, at reasonable cost, makes for greater flexibility and control for the school making the hire.
If one is willing to remove traditional concerns about head coaching experience—at least in the sport of baseball—it would seem that in making a change, assistant coaches should be a consideration for any athletic department. While a head coach can offer a definitive track record of performance, if that track record turns out to have little predictive capacity regarding future performance, at least in comparison to an assistant coach, that value is diminished.
Between availability, cost, control, and return on investment, the pool of collegiate assistant baseball coaches offers many reasons for an athletic department to take note of assistants.
References:
- These outlying cases saw a wild plus or minus change of around 100 spots in the RPI, significantly skewing the final average. Certainly any significant change should be factored into our calculation, but there is reason to note these wild swings in RPI as outlying cases. In all five such examples, the wild change in RPI in the first year was immediately followed in the second year by a return in the following year to previous norms. For example, one school that saw a positive increase from an RPI of 208 up to an RPI of 87 in year one immediately returned to an RPI of 215 the following year. Conversely, there is not one example of a school making a significant jump and then maintaining that level of performance. ↩