The coach efficiency in the NBA 1993-2014: Have been fair the NBA coach of the year award?

Julio del Corral (Universidad de Castilla-La Mancha, @jdelcorraltm)

Andrés Maroto (Universidad Autónoma de Madrid, @jazzandmar)


One of the most successful leagues around the world has just started this week: the NBA. An interesting question is how efficient the coaches have been in the NBA so far. Year after year, the NBA has the best basketball players of the world playing in the league. This fact and the huge professionalism of sports in the USA result in NBA teams having oversized budgets. However, displaying a roster with exceptional players or a big budget is not always a guarantee of success. There are other key variables, such as managing egos within a team, the psychology and motivation, the adaptation to changes and conjuncture problems as injuries, long travels…, which might explain why some teams tend to win more than others in the NBA . Thus, the head coach is an essential factor for the teams to win a Championship or to reach the play-offs, in the case of teams predicted to have worse results.  . Following this reasoning, the efficiency, or how efficiently a main coach manages his resources, is important when analyzing basketball and more precisely the NBA league.

We calculate the efficiency as the inverse of the probability of getting more victories than the number of victories obtained from betting odds. From a more detailed explanation of the methodology, please read this post.

In the following chart we summarize the efficiency of the present main coach from the season 1993/1994. The results show that the best coach during these past seasons would be Steve Clifford. He got the second highest number of wins in a single season in the Charlotte Bobcats history (Hornets at the present), reaching the play offs and being sweept only by the defending champion Miami Heat. Jeff Hornacek and David Joerger would follow him in the list. The ex-All Star player won the NBA Coach of the Month award in December 2013 and improved by 23 victories upon the previous season’s record, leading Phoenix Suns to a 48-34 record in his first season as the main coach. David Joerger, who previously won five times the D-League Champioship, led Memphis Grizzlies to a 50-32 record in his first season as the main coach. A remarkable result is the 0.65 obtained in 18 seasons by Gregg Popovich, all of them in San Antonio Spurs. It is obvious that the higher the number of seasons the more difficult to obtain efficiencies close to one. This result was already highlighted in an article in fivethirtyeight.
Coach
Team
Efficiency
Victories
Matches
Seasons
Mike Budenholzer
Atlanta Hawks
0.52
38
82
1
Brad Stevens
Boston Celtics
0.10
25
82
1
Lionel Hollins
Brooklyn Nets
0.67
214
415
7
Steve Clifford
Charlotte Hornets
0.99
43
82
1
Tom Thibodeau
Chicago Bulls
0.81
205
312
4
David Blatt
Cleveland Cavaliers
Rick Carlisle
Dallas Mavericks
0.67
569
968
12
Brian Shaw
Denver Nuggets
0.26
36
82
1
Stan Van Gundy
Detroit Pistons
0.58
371
579
8
Steve Kerr
Golden State Warriors
Kevin McHale
Houston Rockets
0.66
172
324
5
Frank Vogel
Indiana Pacers
0.72
167
267
4
Doc Rivers
Los Angeles Clippers
0.55
644
1142
15
Byron Scott
Los Angeles Lakers
0.46
416
937
13
David Joerger
Memphis Grizzlies
0.89
50
82
1
Erik Spoelstra
Miami Heat
0.60
314
476
6
Jason Kidd
Milwaukee Bucks
0.69
44
82
1
Flip Saunders
Minnesota Timberwolves
0.61
638
1164
16
Monty Williams
New Orleans Pelicans
0.48
128
312
4
Derek Fisher
New York Knicks
Scott Brooks
Oklahoma City Thunder
0.72
293
463
6
Jacque Vaughn
Orlando Magic
0.10
43
164
2
Brett Brown
Philadelphia 76ers
0.42
19
82
1
Jeff Hornacek
Phoenix Suns
0.96
48
82
1
Terry Stotts
Portland Trail Blazers
0.55
202
447
6
Michael Malone
Sacramento Kings
0.21
28
82
1
Greg Popovich
San Antonio Spurs
0.65
967
1410
18
Dwane Casey
Toronto Raptors
0.51
158
352
5
Quin Snyder
Utah Jazz
Randy Wittman
Washington Wizards
0.39
191
520
8
Will they be able to repeat these successes in a second season? Will Greg Popovich, who was awarded with the NBA Coach of the Year last season, receive the award again? Will Steve Kerr or Derek Fisher success as they did as players? Will David Blatt lead the renewed Cavs to the Championship after winning brilliantly the Euroleague last season? We must wait for answering these and other interesting questions.  

In a recent post it was evaluated the efficiency of the coaches in the 2013-2014 regular season; and by doing so, we have analyzed the accuracy of the “NBA coach of the year” award. According to our analysis the best coach was Steve Clifford from Charlotte Bobcats, 43 victories out of 82; however he was only classified as fourth. On the other hand, Gregg Popovich, San Antonio Spurs’ coach, was awarded with the NBA coach of the year. San Antonio Spurs was the first team classified in the regular season with 62 victories and eventually won the NBA. Hence, the easiest explanation for these results is that in the NBA coach of the year award there is a bias towards conceding the prize to coaches from good teams such as San Antonio Spurs rather than from others of a lower level as Charlotte Bobcats. Still, there is a consensus among the economists about what means to be the most efficient coach: being the coach that has got the most out of the resources. In this way, betting odds provide information about which is more meritorious, whether obtaining 62 victories with the Spurs or obtaining 43 victories with the Bobcats.  In this case, betting odds said that it is  slightly more meritorious to obtain 43 victories with the Bobcats. We use the efficiency of each coach in each season to estimate the following probit models.

Model 1 estimates a probit model where the dependent variable takes the value of one if the coach has been awarded the NBA coach of the year award in that season and zero otherwise. The independent variables are the ratio between the number of victories in the season and the number of matches coached in that season; and the difference between the efficiency of the most efficient coach of the year according to our estimations and the efficiency of that coach. The complete list of efficiencies can be viewed here.

If the award was conceded without bias the ratio between the number of victories and the number of matches would have to be non-significant; however it is significant at any confidence level. Thus, it can be said that in the NBA coach of the year award there is a statistical bias towards the coaches from the best teams.
To build upon this analysis it is estimated Model 3 in which the dependent variable takes the value of one if the coach has been the most efficient according to our methodology. However, some variation in the difTE variable has to be included in order to be able to estimate this model. The results show that the coefficient of percentage of victories is non-significant and even negative. Model 3 was estimated in order to control that the variation included in difTE does not change the results significantly.

It is also worthy to indicate that we have constructed a dummy variable for ex ante famous coaches but this variable was found to be non-significant. Therefore, in the NBA coach of the year award there is a bias towards conceding the award to coaches from the best teams. Nevertheless, this bias should be avoided as far as this award is really prestigious and it should be conceded according to the merit of the contestants.