The Wages of Wins Journal

Brad DeLong is once again trumpeting the career of John Kenneth Galbraith.  Yesterday he posted a link to a Washington Post article (Friends Who Fit Together Smartly: Galbraith and Schlesinger, Like-Minded Neighbors) examining the connection between Galbraith and Arthur Schlesinger.  The article also provides further details on Galbraith’s life and work, and asks whether any current economist can match his intellectual stature. This is a difficult question to address.  In sports stars of the past tend to look better than present day performers, simply because people remember the triumphs and minimize a past player’s missteps.

One of the many triumphs of Galbraith was his ability to write for a general audience.  This is a skill that is not emphasized in graduate school and not often rewarded in academia.  Yet if economists cannot communicate with people outside of our field, what is the value in our research? Let me make a very strong statement: What we do matters only to the extent that it matters to the 6 billion non-economists living on the planet. Galbraith could communicate his ideas to non-economists, hence he was able to make his ideas matter.  Today we have writers like Paul Krugman and Steve Levitt following in the giant footsteps of Galbraith.  Will today’s writers ever be thought of as highly as Galbraith?  My guess is fifty years from now people will be asking: Why can’t today’s economists be more like Krugman or Levitt?

- DJ

JC Bradbury, currently an economist at the University of the South, but soon to be sports economist at Kennesaw State University, has a wonderful site called Sabernomics.  Today he posted a great essay detailing why home runs have increased in baseball.  Bradbury dismisses the issue of steroids and instead builds upon the work of Stephen Jay Gould.  Gould’s work is detailed in our book, and also was referenced heavily in a paper Marty and I published in Economic Inquiry entitled “On the Evolution of Competitive Balance.”  It is nice to see a fellow sports economist also note the work of the late evolutionary biologist.

In addition to noting the work of Gould, Bradbury also referenced a paper, entitled “Steroids, Home Runs and the Laws of Genius” by Art De Vany.  De Vany debunks the role steroids have played in baseball via a variety of methods; including a discussion of how taking steroids would not actually help a player hit more home runs. The role steroids would play in hitting home runs is outside my field of expertise, but it would be interesting to see how sport scientists would react to De Vany’s arguments.  If a player took steroids would he be able to hit more home runs?  Conventional wisdom says yes, but De Vany is clearly saying this is wrong.  Anyone else care to comment?

- DJ

A couple of weeks ago I joined The Sports Economist, a blog led by the one and only Skip Sauer. I have posted a few comments at this site examining the ability of payroll to explain wins in Major League Baseball, the possibility that we would be responsible for future strikes in professional sports, and the behavior of the Memphis Grizzlies and LA Clippers at the end of the 2005-06 season. For the people who only know our work through this blog – and these two people know who they are – these posts would have gone unnoticed. So for these people, I have re-posted what I wrote for The Sports Economist here. If you wish to see the many responses to these blogs, you need to go to The Sports Economist.

- DJ

This was originally posted at The Sports Economist on April 19. 

Like all sports fans, I want to know the future. I don’t want to wait around until October to see who will win the World Series. I want to know right now.

Well, what do we know right now? We know how each team did in spring training. And there is indeed a statistical link between how each team has done in the spring over the past three seasons and where the team finished in the corresponding fall. Unfortunately, it is not a very strong link. Only about 4% of the variation in final standings can be explained by spring records.

Although 4% isn’t much, is it enough to give us some insight into who makes the playoffs? Of the 24 teams who made it to the playoffs from 2003 to 2005, 10 looked like playoff contenders in the corresponding spring. From this we see that only 42% of the time spring training performance correctly identified playoff participants in the fall.

What about the identity of the World Series champion? If spring training mattered, the last three World Series titles should have been won by the Royals (2003), Twins (2004), and Angels (2005). This year the Marlins, based upon their Grapefruit league performance, should be planning a parade. Okay, I may not know much, but I know the Marlins won’t be champions this fall.

A quick confession: before looking at the link between spring training records and final outcomes I suspected there wasn’t much of a relationship. The link between spring and fall is important to note, though, because it is similar to the link between team payroll and final records.

Yes, there is a statistical relationship between what teams pay and how much they win. Like spring and fall records, though, the link between pay and wins is not very strong. As I show in a forthcoming book with Martin Schmidt and Stacey Brook, The Wages of Wins, team payroll only explains about 18% of the variation in team wins in baseball.

If we just think about playoff participants, from 1995 to 2005, the amount teams spent correctly predicted 51% of baseball’s finalists. So all in all, payroll is a bit better than spring records, but only a little bit.

What about the World Series? The leading payroll teams in each league this season are the Yankees and Mets. So a subway series is in our future? Well, if payrolls truly predicted final outcomes, 2006 would be our fourth consecutive subway series. And the Yankees would be playing for their eighth consecutive World Series title.

So do we know the future today? If all we know is spring records and payroll, it looks like we are still guessing. Unless, of course, you are a Marlins fan. If that is the case, you should have had your parade in April, because there won’t be any parade this fall.

This was orignally posted at The Sports Economist on April 29.  

The NHL lock-out of 2004-05 was the seventh labor dispute in the past 25 years to cost North American fans access to the games they love. And this incident was the biggest, leading to the cancellation of an entire season. People ask each time one of these events occur, will the fans come back? In the NBA and NFL, each time games were cancelled fewer fans came back, although the decline did not have statistical significance. The story in hockey is a bit different. In 1994-95 more than 40% of the season was lost to a labor dispute. The fans reacted in 1995-96 by setting a new attendance record. In 2005-06 history repeated itself. An entire season of games was lost and fans reacted by setting a new attendance record. This leads one to ask, does taking hockey away from NHL fans make these fans happier?

Not being much of a hockey fan, I do not know. What I do know has been detailed in The Wages of Wins and in the papers I published with Martin Schmidt. In essence, after seven disputes the data keeps telling the same story: fans do not hold a grudge when games are taken away. Of course, this is not the story the media tells. The media tells us that strikes and lockouts threaten the future of sports in North America. In fact, one could argue that the potential player strike in baseball in 2002 was avoided because both sides were convinced that a strike would threaten the survival of Major League Baseball.

We have now written a book that is bound to be read by perhaps dozens and dozens of people. In this book we argue that the story the media tells about labor disputes is not true. Although telling the truth is important, there may be a negative externality associated with our honesty. If labor and management in professional sports believe that strikes and lockouts do not threaten future attendance, will these events occur with even more frequency?

It is a frightening thought. Even more frightening is the thought that people might think future labor disputes that cancel games are our fault. So perhaps my mother was wrong. Maybe honesty is not always the best policy.

- DJ

This was originally posted at The Sports Economist on May 2nd.  

One of the best recent articles written in the field of sports economics was “Losing to Win: Tournament Incentives in the National Basketball Association.” by Beck Taylor and Justin Trogdon. This paper, which appeared in The Journal of Labor Economics in 2002, offered evidence that before the institution of the draft lottery, non-playoff bound NBA teams actually played to lose towards the end of each season.

Although the NBA denies that teams ever tried to lose to secure better draft picks, the league did institute a draft lottery to lower a team’s incentives to try and lose games. Although the draft lottery might change the incentives of non-playoff teams today, the current playoff format offers a new problem. When the Memphis Grizzlies faced the LA Clippers on April 18, the winner secured the 5th seed in the Western Conference playoffs while the loser was seeded 6th. The better seed, though, was not much of an award. The 5th seed earned the right to play the formidable Dallas Mavericks while the loser got to face the less impressive Denver Nuggets.

In the game both teams appeared to do their best to act on their incentives. The Grizzlies sat their best player Pau Gasol. Gasol led the team in Wins Produced, producing 28% of the teams wins in 2005-06. Memphis also reduced the minutes of Eddie Jones, who produced 8.3 wins this past season. The Clippers countered by sitting Chris Kaman – 8.6 Wins Produced – and Sam Cassell – 7.1 Wins Produced. Furthermore, Elton Brand, who produced 17.7 wins this past year, only played 22 minutes. In the end, the Clippers managed to sit more talent than the Grizzlies and LA lost the game by six points.

As expected, Memphis reaped its reward in the playoffs. Last night the Grizzlies were swept out of the playoffs by the Mavericks. Meanwhile the Clippers eliminated the Nuggets last night, moving on to the second round for the first time in 29 years.

So this all begs the question: Should Memphis have tried harder to lose to the Clippers on April 18? If Memphis started all of its bench players, would the Clippers have tried to play their cheerleaders? Given the behavior of these teams, and the results we saw last night, clearly the NBA needs to re-think their playoff format if it wishes to avoid teams playing to lose in the future.

- DJ

My post regarding competitive balance in the NBA generated a few responses at The Sports Economist blog.  I posted a response at the Sports Economist which I thought I would make available for vast millions – okay, one or two – people reading this blog.

The responses to my post center primarily on the issue of how to measure competitive balance. Let me respond by noting that I agree that focusing only on playoff outcomes is not the best way to get at competitive balance. My purpose in taking this approach, though, was to note the relative predictability of playoff outcomes in the NBA, or from the other side of the coin, the unpredictability of playoff outcomes in the other sports. Frankly I just think that is a neat story.

As it turns out, when we look at “better” measures of competitive balance, such as the Noll-Scully — which compares the standard deviation of winning percentage to the ideal standard deviation that would exist if the league were balanced — the same story is told. The standard deviation of regular season winning percentage in the NBA has been 2.7 times the ideal over the last three decades. In the National Football League, National Hockey League, American League, and National League the ratio of actual to ideal has been less than two over this time period. One can tell the same story if you look at the last 10 years or last 20 years.

What is neat about the Noll-Scully metric is that it allows one to compare balance in different leagues. This is important because for one to say competitive balance is “good” or “bad” one needs to establish a reference point. With the Noll-Scully one can say competitive balance in the NBA is not as good as what we observe in the other major North American sports. That is also the same story I told by looking at how regular season performance predicts the playoffs. Again, I think the playoff story is still neat, but I fully acknowledge metrics like the Noll-Scully are better measures of balance.

Once one has a measure of balance, you next need to address whether balance and attendance are linked. In our published research, as well as work by Brad Humphreys, a statistical link between attendance and balance has been uncovered. The economic impact of balance on attendance, though, is quite small. In other words, it is not clear that fans truly care about the levels of balance they have observed in baseball and basketball. These findings, already reported in the sports economics literature (and –pardon the shameless self promotion — our forthcoming book The Wages of Wins) are also consistent with what I was saying in my post about recent attendance records and balance in the NBA.

In sum, if we compare baseball to basketball, it is hard to argue baseball has a competitive balance problem. Furthermore, even if you could argue that there was a problem in baseball, it is not clear that fans actually care very much. So I come back to my original point: If fans do not care that much about balance, then aren’t attempts to fix this “problem” just another excuse to take money away from the players?

- DJ