1 on 1 with David Berri: Part 1


David Berri

David Berri

In Part 1 of a two-part interview with Stumbling on Wins author David Berri, the Southern Utah econ professor who uses economics to discover inefficiencies in sports, Berri discusses where the NBA statistical movement is at, why the Suns have gotten more value out of their roster than any other NBA team in the past six years and whether chemistry really explains the success of the 2009-10 Suns. Also be sure to check out Part 2, in which Berri discusses whether the Suns have adequately replaced Amare Stoudemire.

Michael Schwartz: If an economist were to take a job as GM of an NBA team and run the team rationally using the principles discussed in Stumbling on Wins, how would they do?

David Berri: A team that was aware of the research in economics would be able to make better decisions.  For example, such a team would look beyond scoring totals in evaluating a player (and consider the factors reported below), avoid considering such factors as Final Four experience or an inch of height in evaluating draft picks, and avoid giving minutes to high draft picks who have already shown after one or two years in the league that they are not productive.

Although understanding this research is helpful, I think people need to know there is more to the job of a general manager.  A general manager must be able to effectively communicate to an organization and also interact with the other teams in the NBA.  And if you don’t have the skills to accomplish these tasks, you will probably not be that successful. In other words, Bill James and Billy Beane both understood that baseball teams were making mistakes in the 1990s.  But I am not sure James would have been the right person to lead the Oakland A’s.  It does help for a person to have more skills than just knowing the latest statistical research.

Schwartz: From where you’re sitting, where is the NBA statistical movement at? Do you think there will be a day when we’re all talking about win shares and PER?

Berri: From the perspective of teams, I think we have reached a point where many people recognize that one cannot evaluate a player’s contribution to team wins by just watching the players.  In other words, people in the NBA know that some statistical analysis is necessary.  But teams are struggling to figure out which statistical approaches are really helpful. For example, there are significant problems with both Adjusted Plus-Minus and the Player Efficiency Rating (PER).

By the way, for a comment on the problems with adjusted plus-minus see:

http://dberri.wordpress.com/2010/03/07/john-hollinger-dean-oliver-and-some-other-people-comment-on-plusminus/

For a comment on the problems with PERs see:

http://dberri.wordpress.com/2006/11/17/a-comment-on-the-player-efficiency-rating/

Beyond the problem of understanding which statistical models are more useful than others, there is also the problem of fan interest.  Think about baseball for a moment.  People familiar with baseball statistics understand that batting average is not a particularly good measure of performance.  But if you watch a broadcast of a game you don’t hear much about the “advanced” baseball stats.  Decades after it was shown batting average isn’t very good, it is batting average that is most likely to be noted by broadcasters.  That is because most baseball fans are familiar with batting average.

The “advanced” stats in baseball are not nearly as well understood.  I suspect the same is true in basketball.  I don’t expect we are going to see most NBA fans move away from discussing players in terms of points scored per game.  In other words, the “advanced” stats in basketball — just like we see in baseball — are not likely to be seen in a typical future broadcast. And we should note an additional reason why that’s the case.  Batting average can be mentioned by a broadcaster without any additional explanation.  If you use a measure that most fans have never heard of, then you have to spend time during the broadcast explaining the “new” measure.  Such an explanation has to be offered every time the “new” measure is discussed.  Offering this explanation would prove to be tedious to the broadcasters, and probably not that interesting to most sports fans.  Hence, the broadcasters probably conclude that most people are better off with the more simple measures.

Schwartz: According to this story on your site on value proposition, the Suns have gotten more bang for their buck over the last six years than any team in the NBA. Why do you think that is?

Berri: This was a story written by Arturo Galletti for the Wages of Wins Journal.  Arturo found that from 2004-05 to 2009-10, the Suns and Spurs were the only team to get more wins than they paid for in each and every season. For the Spurs, one might credit the team’s leadership since it stayed essentially the same across the entire time period examined. The Suns, though, have had a change of leadership.  The results, though, stayed the same.

So that leads us to ask … what’s up with Phoenix?  And the answer is … I don’t know.  All we can see is that Phoenix has done quite well across the past six seasons.  Yes, they have failed to win a title.  But given the money spent, Phoenix has given their fans more wins than one would expect.

And since the Suns wisely chose to avoid giving Amare Stoudemire $100 million (a point I will expand upon in Part 2), it is possible the Suns will continue pay less than other teams for the wins the team achieves.

Schwartz: In Stumbling on Wins you discuss how scoring is sorely overvalued in the NBA both in terms of player salaries and voting for the All-Rookie team. You use the Isiah Thomas story to show how he failed in New York by putting together the best scorers he could find although they lacked proficiency in other areas of their games. For my readers who may not have read the book but are interested, why is this and what should teams really be looking at when evaluating players?

Berri: What we report in the book is evidence that decision-makers in the NBA primarily focus on scoring in evaluating basketball talent.  Whether one looks at free agent contracts, the NBA draft, the coaches’ voting for the All-Rookie team, and the allocation of minutes; the evaluation is driven by points scored.  Points scored, though, can be manipulated by the player.  Specifically, the more shots a player takes the more points he will score. So players have a clear incentive to focus on their shot attempts.

Wins in the NBA, though, are not strictly about points scored.  Wins in the NBA are about shooting efficiency and factors associated with gaining and maintaining possession of the ball (i.e. rebounds, turnovers, and steals).  These factors, though, do not often drive player evaluation.

So why do teams primarily focus on scoring? Scoring is the event that stands out in watching a game.  We suspect that teams consider many more factors. But we also suspect that not enough effort is made to formally evaluate the relative impact of the various factors considered.  And the informal effort made to rank all the factors considered tends to place the most dramatic factor (i.e. scoring) at the top of the list.  In sum, we don’t think teams only consider scoring.  But we think the decision-making process leads scoring to dominate the player evaluation process.

Schwartz: One of the biggest story lines of the 2009-10 Suns was their amazing chemistry. Is there a way to consider this intangible when analyzing a team from an economic sense?

Berri: Chemistry is a term often used in discussing sports. But it’s not often defined. What I tend to see is that a team exceeds expectations.  Consequently people argue the team has “good chemistry.” Or a team — like the Knicks under Isiah Thomas — fails to meet expectations.  Then the team is described as having “bad chemistry.”  In Stumbling on Wins, though, we present evidence that the Knicks under Isiah didn’t have “bad chemistry.”  They just had a collection of players who were not quite that good.

Let’s take the same approach offered to discuss the Knicks in Stumbling on Wins and use this to discuss the Suns in 2009-10.  The following table reports what the Suns did in 2009-10.  As one can see, the team’s Wins Produced sums to 54.0 (the team actually won 54 games).  When we look at performance of these players in 08-09, though, we would only have expected about 40 wins.  So the team was about 14 wins better than expected.

Phoenix Suns Wins Summary

Phoenix Suns Wins Summary

Looking at the individual players we can see who was responsible for this improvement. About 11 of these extra wins can be linked to Robin Lopez, Goran Dragic, and Steve Nash.  Lopez and Dragic are young players who didn’t play much in ’08-09.  So we should not be surprised both got better. Nash is old and should be declining (and someday he will).  But he actually reverted back to what we saw in ’07-08.

Had we known about Lopez, Dragic, and Nash, then we would have expected the Suns to win about 50 games.  Because we didn’t know about these three players, though, this team exceeded expectations by a wide margin.  And consequently, we are now talking about chemistry.  In reality, this improvement can be explained by just looking at just a handful of players the Suns employed.  Most players on this team did about what we would have expected.  Or to put it another way … if there was a “chemistry” issue in Phoenix this past offseason, it didn’t impact most of the players on the team.

Be sure to also check out Part 2, in which Berri delves into Amare Stoudemire’s value and analyzes how the Suns have chosen to replace him.

Tags: David Berri Stumbling On Wins Wages Of Wins

  • http://brightsideofthesun.com Alex

    Hey Michael, this is a great interview. Thanks for the insight, and the referenced to some analytics. Hope you don’t mind I linked to this article on BS0tS this morning.

  • Al

    Earl Clark was LAST!!! Time to trade him for Jason Thompson. Suns could have drafted him in 08 Draft.

  • suns68

    The problem with statistics in basketball is that the game is so fluid, it’s difficult to impossible to reduce it to meaningful mathematical measures.
    In baseball, for example, there are a lot of discreet, measurable events. The player either hit the ball or he didn’t. He caught the ball or he didn’t. Batting average or fielding percentage means something.
    But in basketball, you can say Player A hit the shot, but what you can’t measure is that there are four other guys on the floor making stuff happen to make that shot possible, or easier.
    Looking at the chart above, can anyone seriously make an argument that Channing Frye and Jared Dudley were costing us games? Compared to what?
    I watched probably 75 regular season games and all of the playoffs, and I thought both those guys had breakout seasons.
    What I saw on the court was that Frye”s unsuspected talent for the three ball pulled big men away from the basket, opening up driving lanes for Amare. And Dudley’s never-say-die hustle seemed to inspire everyone around him, except maybe Earl Clark, to play their best basketball.

  • duds

    So according to this statistic – which I’ve never heard of before now, and I’ve heard a lot about other advanced statistics like win shares and PER – Lou Amundson is better than Goran Dragic and Robin Lopez.

    Yeah, this is why you still need to watch the games.

  • Pingback: A Q&A with Valley of the Suns « The Wages of Wins Journal

  • Sun Stats

    Suns68, I think you’re reading the table wrong. Frye and Dudley didn’t cost the Suns games. They just didn’t produce as many wins as last year hence the negative in the difference column.

  • mike

    Not “Better”, more effective at producing plays that affect a win.

    To understand WP48, you need to understand that Basketball is a game of efficiency & net possessions: the more possessions a team has, and the more effectively it uses them, the better.

    The reason people miss this is that there are so many possessions, and “efficiency”, the ability to use possessions wisely, is less noticeable than a behind the back pass for a dunk. But is a solid box out and rebound less valuable than a flashy pass that fails 3 out of ten times?

    It’s the same reason people gamble: they bias for things they remember, not for a solid statistically repeatable model. And how often did you, or does anyone, watch Lou vs Goran?

  • Mike Meez

    Great article. Looks like an interesting, complex way to measure efficiency. And it certainly seems like a better method than focusing on PPG. However, I agree with suns68 about the nature of the game defying. Having Channing Frye as a 3 point threat was huge and that won’t always show up in any statistics. Defensive ability does not always show up in statistics either. If a guy locks an offensive player down so that the guy on offense has to pass, that is solid D but doesn’t get taken into account in stats. Having Robin Lopez down low to matchup with Bynum and Duncan means that those teams won’t pound the ball inside as much so plays that would have occured without Lopez now don’t happen at all. There’s not a stat for that either but obviously it’s very important. So even as intricate as the method may be for this table, I’m not ready to buy wholesale that it is completely accurate as to how many wins a guy contributed. Maybe I’ll get the book and see how he comes up with his figures though.

  • http://www.valleyofthesuns.com Michael Schwartz

    @Meez I would definitely recommend this book as well as Berri's first book, Wages of Wins. I'm kind of a closet econ nerd and I was fascinated by their analysis of inefficiencies in sports and some of the models they used to evaluate players and outside factors such as attendance.

  • Pingback: 1 on 1 with David Berri: Part 2| Valley of the Suns

  • JR

    I would add one comment on chemistry. I think his chart shows "chemistry" even though he does not recognize it. I would define as a team being more than the sum of its parts. In statistics that would show up as players playing better together than they do separately. A number of players did better (according to the Berri's criteria) than the previous year – including Nash. When a large portion of a team "over performs" – I would call that chemistry.

  • ghost of Darko

    when are the Mensas who come up with these "advanced stats" gonna correct for the deficiencies in their advanced stats? and I'm sure Suns fans are thrilled at their team's bang for the buck performance. sure makes the 2 finals appearances in 42 years along with watching all that talent walk out the door in the last 6 years taste so much better.

  • Pingback: Reconciling Wins Produced with team chemistry| Valley of the Suns

  • Pingback: A Phoenix Suns playoff Wins Produced analysis | Valley of the Suns