Thursday, March 31, 2011
Should Managers Be Sacked Mid-Season?
The investigators performed a highly detailed and complex statistical analysis of German Bundesliga matches. The period covered 1963 through 2009 and represented a total of 14,018 matches. During this time period, there were 361 mid-season coaching changes.
Goal differential (GD) was used as the measure of “team fitness” for most of their analyses. The researchers discussed using points but argued that GD was a more sensitive and stable measure of team performance. To determine if the coaching change affected team fitness, they examined the GD of 10 matches preceding the dismissal of a coach and 10 matches after his dismal. In all, 154 teams played at least 10 matches prior to and 10 matches after dismissal of the manager (some dismissed their coach with only a few games remaining in the season). For each dismissal team, a control team was selected from the same season. Control teams had the same GD for the 10 matches played prior to the date of the dismissal team’s manager sacking. However, their coach was not dismissed. Thus, the dismissal and control teams showed roughly the same performance during the matches leading up to dismissal date with only one coach being dismissed.
The statistical analysis compared the team fitness (GD) between the control and dismissal teams before and after the date of the coach dismissal. They also compared the change in team fitness as a result of dismissal.
The results of the analysis are striking. First, the GD over the course of the season varied very little. For teams that retained their manager for the duration o the season, the GD remained constant at about -0.50 per match. The same held for the dismissal teams.
Second, the researchers found that changing a manager in mid season had almost no effect on team fitness. With a new manager in place, GD improved by, at most 0.08 per match. From a statistical standpoint, such a change is negligible. Therefore, replacing the manager with at least 10 games remaining in the season had no effect at all on team fitness. Further expanding the analysis to include more than 10 games prior to and after sacking gave the same results - no improvement in performance.
Third, the researchers noticed that there was a noticeable drop in GD during the two matches prior to the coach dismissal (this may have prompted the dismissal). Immediately after dismissal, the GD returned to its previous level. Thus one could argue that the new coach reversed the downward trend in form. However, the researchers compared this trend to other teams that showed a similar drop in form. Essentially, many, many teams experience similar declines in team fitness and rebound after a two match drought. Thus, it is unlikely that changing coaches accounted for the improvement in fitness. It is more likely that this represents the common ebb and flow of the season rather than a coaching change by itself.
Finally, because wins, losses and table position are the ultimate marker of team performance, the researchers re-analyzed the teams using points instead of goal differential. They found essentially the same results. After dismissing the manager, the points earned over the next 10 matches was nearly identical to those earned during the preceding matches.
What about replacing the manager during the off-season? While the analysis of this concept is more complicated, the investigators found that changing managers under this condition affected a team’s performance by, at most, 15%. In fact, this value may be an overestimate and that changes made to the coaching staff as well as player movement probably had the greater influence.
Based on this study, sacking a manager in mid season seems to have almost no effect on the team’s performance. Also, replacing him between seasons seems to have little impact. The study does not say that replacing a coach cannot impact a team’s play. There are instances of team performing much better after a coaching change. What this study shows is that on average, across nearly 45 years of Bundesliga seasons, dismissing a coach has had little influence on a team’s performance.
Why does sacking a manager have so little impact on team performance? Are managers important? It is possible that managers working at the elite professional level (such as the Bundesliga) are so highly qualified that one performs his duties as well as another. Given this, the researchers conclude that the immediate impact of a coach as a “fitness producer” is very small. Compared to other factors such as the team’s wage budget and playing personnel, they estimate that a new manager’s influence is on the order of 15% or less.
As for the second question, it should be pointed out that success and failure can be separated by a fine line. The average number of points earned by the 15th and 16th place teams in the Budesliga (the relegation line) for the past five years is 33.2 and 31.6, respectively. That equates to a 5% difference in performance. Thus a small influence can, in some cases have a sizable effect.
Should managers be sacked mid-season? The answer is obviously very complicated. The analyses performed in this study suggest that noticeable short-term (10 games) improvements are not likely to be realized. It would be very rare indeed for a team to transform itself based on a coaching change alone. Then again, an extra point or two earned could mean the difference between relegation and remaining in the league.
Heuer A, Müller C, Rubner O, Hagemann N, Strauss B (2011) Usefulness of dismissing and changing the coach in professional soccer. PLoS ONE, 6:e17664.