A Former NBA Coach Argues That Coaches Are Not Responsible for Outcomes

(Photo: Steve A Johnson)

The coaching carousel continues to spin in the NBA.  In recent days, the Los Angeles Clippers – coming off the best season in franchise history – have decided not to bring back Vinny Del Negro as head coach.  The Phoenix Suns — coming off their worst season since they were in expansion team in the late 1960s – have decided to turn to Jeff Hornacek to lead their team back to respectability.   And the Atlanta Hawks – who were essentially average this last season – have turned to Mike Budenholzer to lead the team next year.

These are hardly the only teams to make a change.  Since the end of the 2012-13 NBA season, the Brooklyn Nets, Charlotte Bobcats, Cleveland Cavaliers, Detroit Pistons, Milwaukee Bucks, and Philadelphia 76ers have all decided that the person who coached the team at the end of this past season shouldn’t be around for the next season. In all, at least nine of the 30 NBA teams will have a new coach next year. 

These changes – as I have argued before –will probably not make much difference.   A study published in the International Journal of Sport Finance (full PDF here) – which I conducted with Mike Leeds, Eva Marikova Leeds, and Mike Mondello – found that most NBA coaches across a sample covering 30 years did not have a statistically significant impact on player productivity. And in other sports, we also have evidence that coaches cannot systematically change outcomes.

One would expect that coaches would disagree with these studies.  Certainly a team is not likely to commit millions to a coach who claims the coaches really don’t matter.  That is why I was surprised to read the following from Brendan Savage of mlive.com a few weeks ago.  The article includes extensive quotes from Jeff Van Gundy, a former NBA coach (and current television broadcaster) whose brother (Stan Van Gundy) is actively seeking employment as an NBA coach. 

Without mentioning Joe Dumars by name, Van Gundy blamed Detroit’s president of basketball operations – and not Lawrence Frank, who was fired… after two seasons – for the Pistons’ failure in the past few years.

Here’s what Van Gundy had to say when asked about the recent NBA coaching changes that included Doug Collins resigning in Philadelphia and Byron Scott getting fired in Cleveland. 

“Detroit Pistons basketball slogan: When the going gets tough, we fire the coach,” Van Gundy said. “It’s unbelievable. It’s unbelievable. You know what surprises me, Chris? These new owners in Detroit have to be exceedingly bright to have made as much money as they have. And to be duped again that your G.M. tells you that the roster is good and the coach is bad … what was the problem with Michael Curry?  What was John Kuester? Now Lawrence. They run through coaches and they haven’t even begun to address their problem. They have very little talent and very little basketball character. You combine that, you’re going to be in a long rebuild.

“I’m just surprised that when everybody acknowledges it’s a player’s league – everybody would agree with that – then the most important player or person in any organization is the person that picks the players. But we don’t, as organizations, examine them. We just take the easy way out time and time again. You lose, the G.M. convinces the owner ‘We got good players. It’s the coach’s fault.’ We fire the coach; we bring a new coach in; we continue to lose. We fire that coach, saying that ‘We have better players.’ It just goes on and on. It’s typical and I can’t believe that the Detroit owners fell for it. I just can’t believe it.”

One suspects that the new owner of the Pistons “fell” for this story because this is the story coaches tell.  NBA coaches are paid millions of dollars.  To secure these contracts, one suspects that these coaches argue in the interview process that coaches are truly important.

But now we have a former coach (again, with a brother looking for work) telling a very different story.  And again, I think the data supports what Jeff Van Gundy is saying.

As Van Gundy notes, the Pistons have recently gone through three coaches. And the Pistons have not had much success after these coaches came to Detroit.  Before these coaches arrived, though, the Pistons were quite successful.  

The difference, though, doesn’t appear to be the coach walking the sidelines.  When the Pistons were successful the team simply employed more productive players.

To see this, let’s look back at the 2007-08 season.  That year the Pistons won 59 games.

When we look at Wins Produced (calculation detailed here), we can see that most of these wins were produced by a minority of the team’s roster. 

The Pistons in 2007-08 (numbers from theNBAgeek.com)

An average player produced 0.100 wins per 48 minutes (WP48).  Of the players who played at least 1,000 minutes in 2007-08, five players were above average. And these five produced 46.2 of Detroit’s wins.

But since the end of the 2007-08 season, these players departed and/or aged.  For example:

For the Pistons to remain competitive, this production had to be replaced.  And when Iverson’s contract expired at the end the 2008-09 season, the Pistons had an opportunity to acquire productive talent in the free agent market.  Unfortunately, the Pistons – and specifically, this means Dumars — decided to spend much of this money on Ben Gordon and Charlie VillanuevaAt the time I noted that neither player had ever been very productive.  And since this trade, neither Gordon nor Villanueva has been very productive. 

Unfortunately for the Pistons, not only did the Gordon and Villanueva acquisitions fail to work out, the team also had trouble finding other productive players.  

Again, the Pistons had five above average players in 2007-08.  Here are ALL the above average players the Pistons have employed since 2008 (minimum 1,000 minutes played). 




Wins Produced

Antonio McDyess




Jason Maxiell




Tayshaun Prince




Ben Wallace




Tayshaun Prince




Jonas Jerebko




Jason Maxiell




Greg Monroe




Ben Wallace




Tracy McGrady




Greg Monroe




Jonas Jerebko




Andre Drummond




Greg Monroe




The list only includes eight names.  Three of these players – Maxiell, McDyess, and Prince – were on the 2007-08 team.  As noted, these players aged and so their production had to be replaced. 

Two more names – Ben Wallace and Tracy McGrady – were both quite old when the Pistons added each to the roster.  So both could hardly be thought of as the sort of players the Pistons were acquiring to build for the future.

That leaves us just three names.  Jonas Jerebko, Greg Monroe, and Andre Drummond were all acquired in the draft.  Of the 14 players drafted by the Pistons since 2008, these are the only players that have become productive players.  So the Pistons have missed more than they have hit in the draft.

In sum, the Pistons – since 2008 – have not been adding enough productive players to contend. Joe Dumars – as Jeff Van Gundy notes – has been the person responsible for assembling these rosters. And he has also been the person who seems to believe that if he just finds the right coach, somehow these players – who generally do not have a history of producing large sums of wins – will somehow become productive NBA players.  But the data says this is unlikely.  Basketball players tend to be consistent across time, and changing coaches doesn’t often change what players offer on the court. 

The good news is that the Pistons still have Drummond, Monroe, and Jerebko.  So the cupboard is not bare.  But also, the team needs more. 

Can Joe Dumars be counted on to find those players?  Or is he currently looking for the next Gordon or Villanueva?

We don’t know which players Dumars plans on adding.  We do know that Dumars is currently looking for a coach.  But as Jeff Van Gundy has argued, that is probably not going to make a difference.  It is the players that matter.  And until Dumars finds more productive players, fans of the Pistons (and yes, I am one of these) will continue to be disappointed.  

Frank Coleman

Great article. PJ Carliesemo (sp) who coached the Nets this past season said on Mike and Mike that if you swapped the coaches of the best teams with coaches from the worst teams, the teams would still be the same.


I get the point, but it's not really fair to examine wins produced when determining the effectiveness of the head coach because the WP48 formula is only a player statistic. If coaches were to have a positive effect, it would be reflected in the players' stats.

It seems to me that it's an issue of parity (and oversupply?) of NBA coaches. One top-level coach might not be better than the next, but I have to imagine teams would suffer if you pulled a random person off the street and asked him to coach an NBA team. Perhaps the detrimental effect on the players' statistics would be written off as a product of aging another year as is often done when no one looks for any other explanation.


But that's exactly what the study did. It looked to see if NBA coaches had an effect on NBA players. And most of them didn't (and yes, the age of the players was controlled for).

As to your final point, yes, it's likely that teams would suffer if they were replaced with some random person. But that was never the argument advanced by Berri.

Andrew B

I coached youth basketball for years- some years we were undefeated, some years we had zero wins, and usually it was in between. I was the same person, although I changed the team's style based on the players, but clearly it was whether I happened to have good players that year. Parents would call me a genius or idiot (I am exaggerating) depending on how we did.

Dwight K Schrute

Same here. The only constant was my son(s) in various sports from season to season. My opinion was that if you drafted well, you would play well. My team had 1 win one season and 1 loss the following season.
At the college/pro level, I would say that a coach can lose games but won't win many on his/her own. A GM will never admit a mistake in setting the roster, so its always the coach's fault.


The Other Problem With The Conclusions From The Study

Like in ever business there are only a limited number of top of line people. In the NBA you have a limited number of excellent coaches a lot of mediocre coaches and very few bad coaches who don’t last long and are usually replaced by equally bad or mediocre coaches.

Good coaches do not go from team to team. So their team has no coaching changes. Jerry Sloan was with the Jazz for 25 + seasons. How do you measure a coaching change affect for Utah over those 25 + seasons when they had only one coach? Phil Jackson was with the Bulls for nine seasons and Lakers for 11 seasons. Would any coach have won all those rings in Chicago and Los Angeles? There is no way to tell. Pop has been with the Spurs as head coach since Santa Ana attacked the Alamo. How do you measure his impact on the success of the Spurs? So the study deals with a lot of data points of teams from among the large pool of mediocre coaches.

Yes changing one mediocre coach for another will not have much if any impact. However replacing a mediocre coach with a top tier coach of which there are few or a lousy coach of which there are few will have an impact.

For statistical people they forget the 80/20 rule and forget all about the bell shaped curve which puts 2/3 of the data points within one standard diviation of the mean. This means that 2/3 of the coaches will not make much of a difference but going from a coach at the one end of the bell shaped curve to the other end of the bell shaped curve will make a difference. Also going from one of the coaches within one standard deiviation of the mean to a coach at either end of the bell shaped curve will make a difference.

Of course even a great coach needs talent to win and great players will make any coach look better just as lousy players will make any coach look worse.



If your population includes only NBA-worthy coaches and not all basketball coaches then one can argue the curve is relatively flat such that the difference between the top and bottom is quite insignificant to determining outcomes.

For example, among the top 30 brain surgeons in the world there probably isn't much to be gained in getting your surgery done by #1 or #30.


You can take 30 scientists and throw Einstein and Newton in the group and not all 30 would be about equal. The best coaches in the world? are coaching in the NBA. One could argue that, so there will always be a few to a handful that will be exceptional and they rarely change teams.

To the poster below. There are moe than 30 NBA coaches. Even in one season there are more than 30 due to firinigs. I would say that the population of NBA coaches is more like 40-50 than 30. While they don't all coach a NBA team in one season they are constantly part of the rotating cylce of coaches that would be included in the study.

[WORDPRESS HASHCASH] The poster sent us '0 which is not a hashcash value.


Dumars has set that franchise WAY behind in the same way Isaiah did with the Knicks.

It's not all his fault. But picking Darko over (you name it, but let's just say:) Wade was a franchise-altering move. They've had bad luck and poor scouting with their picks ever since.

The Boourns

It's easy to fault Dumars for the Darko pick in hindsight when you see how well Wade, Carmelo, and Bosh all turned out, but for the most part Darko WAS the consensus 2nd or 3rd overall draft pick. It's not like Dumars stretched on selecting Darko...just about every GM and scout out there thought Darko would be a potentially better Dirk.

As for their luck since Darko...I'm not sure you can argue that Dumars has been all together bad at drafting. Stuckey at 15th overall was a solid pick, but he's been poorly handled. Afflalo in the late 1st was an absolute steal (again mis-handled). Amir Johnson was a crazy good pick, but again mis-handled by being traded away too soon just like Afflalo. Greg Monroe is most likely going to be the best big man in his draft class assuming Cousins never gets his act together and Drummond will definitely be the best big in his draft class unless he is hindered by injury. Jonas Jerebko was a solid pick up in the 2nd round and at this point, so was Kyle Singler. Dumars worst pick since Darko was Austin Daye, and it was horrendous, but Dumars' bigger issue is once he makes a smart draft pick, he either mis-handles them by putting guys like Allen Iverson on the team with a young impressionable player like Stuckey or he simply mis-values them like he did when he traded away Afflalo and Amir just so he could turn around and sign Ben Gordon and Charlie Villanueva.



Miami has a coach who throws out the basketball and watches.

Chicago has a coach who seems to motivate the players to play above their abilities.

GM's are more important then coaches, but coaches who can get players motivated are rare in the NBA. Also some coaches are better at designing a system that works for the players they have.

So, talent first (from GM) and a system that utilizes talent second (from coach). Perhaps coach is a distant second, only half the worth of a good GM.


you cant be serious with that characterization of Spoelstra. Look at the heats spacing alone. Spoelstra has a great roster and has tailored the offense to maximize the talents he has. He's a really good young coach. and i hate the heat.


Well, look at it this way - Detroit probably aren't helping themselves by firing coaches, but unless there's a payout involved they're not hurting themselves either!


So, if Jerebko was above-average in '11-'12, but not in '12-'13, one reason would be inconsistent playing time. Inconsistent playing time falls on the coach, IF, as Berri contends, "Basketball players tend to be consistent across time..."


ding ding ding! we have the winning comment.


3 above average players, and only one of them had enough consistent playing time.

A good coach figures out a way to get and keep the good players on the floor.


Isn't it the GM hiring the coaches, too? At some point doesn't the inability to pick a good coach bring the GM into question?


this could be true for football as well. Bill Belickick's record with out Brady is 43%.