Responding to Zimbalist

Today, Maury Brown has posted Andrew Zimbalist’s review essay of The Baseball Economist along with Vince Gennaro’s Diamond Dollars. This essay is to be featured in the next issue of the Journal of Economic Literature.

I am open to criticism, and I feel that my book is subject to many reasonable objections. After all, if I thought everyone would agree with the entire content, there would have been little reason to write it. I have had many meaningful conversations with readers about the content of my book. Even I have some concerns about things that I have previously written. However, Zimbalist’s critique is lacking and borderline dishonest in several places. Believe me, I take no joy in dismantling a negative review, but several of Zimbalist’s incorrect claims force me to respond.

Below, I detail several parts of Zimbalist’s review of my book (Vince can defend his own work) and add my responses. In many cases it is clear that Zimbalist failed to even read the book. In other criticisms, he is just wrong. I follow the order of his review, and though I do not respond to everything, I cover enough to show most of his concerns are without merit.

On umpire rent-seeking, Zimbalist writes:

Bradbury’s chapter on whether it pays for a manager to argue with the umpire is provocative, but unsatisfying….It is rent seeking because there is no net gain, no output increase, just a transfer of marginal calls from one team to another. Meanwhile, the fans, according to Bradbury, have their utility lowered because they have to spend a few extra minutes at the game due to these fits of managerial distemper. Well maybe, but it is also possible that the fans enjoy managerial protests both because they are amusing and because it vicariously vents their own frustration at bad umpire calls.

Fair enough, it’s not my favorite chapter, either. The point of the chapter was to teach the concept of rent-seeking using baseball instead of typical boring classroom examples. I don’t think it’s a policy issue of great importance, and there is no doubt that it is my own preference that arguing be removed from the game.

On the big-city-versus-small-city discussion, Zimbalist writes:

Bradbury’s simple regression finds that variance in city size accounts for 40 percent of the variance in win percentage over a period of years. This seems to indicate a rather substantial impact of city size. Further, the author fails to consider the interactive effect of city size and a team owning its own regional sports channel (RSN), the number of large corporations in the market, or the size of MLB’s assigned team television market – three factors that would have reinforced the effect of city size.

The regression I report is simple because of the numerous potential control variables that I included, did not show any effect. And with only 30 teams, my degrees of freedom were approaching/surpassing the minimum. I looked at income, split the markets with multiple teams in two, and included multiple team dummies. For this reason, I did not report the results. I admit to not including a dummy variable or interactive term for whether or not a team owned its own RSN. This is endogenous—having an RSN to generate money that leads to wins is something that any team could do, regardless of its size. Better-managed teams won out over those that did not. I have no doubt that TBS helped the Braves improve their financial standing beyond what Atlanta ticket-holders produced, but that is the point: population wasn’t the cause. I’m not sure how this cripples the regression estimates or affects the supporting material in the chapter that doesn’t rely on the regression estimates to support the point that big-market advantage is overblown.

Zimbalist continues:

Along the way, Bradbury misapprehends the functioning of the amateur draft and overlooks the unequalizing effect of the posting system with Japanese baseball.

I’m not sure what I missed with the draft, nor how the posting system drastically alters what I reported. But, as you will soon see, Zimbalist likes to drop these bombs and run.

Zimbalist doesn’t like my application of the prisoner’s dilemma to explain the incentives for steroid use in baseball.

His chapter on steroids in baseball employs game theory to model the choice that a player makes whether or not to indulge. He argues that when every player chooses to use steroids it is a Nash equilibrium. This result, however, appears to depend on his arbitrarily chosen values for the supposed productivity gain and the health costs (only $500,000) from indulgence. Bradbury’s analysis ignores the enormous uncertainty that surrounds this choice for players.

As it happened with the rent-seeking chapter, Zimbalist forgets that I am using an issue in baseball to introduce readers to an important economic concept. The example I use includes hypothetical, not actual, payoffs to help elucidate the incentives faced by players. Of course the payoffs affect the choice. I designed the matrix to explain how a prisoner’s dilemma operates. This is very clear in the text. If he has an argument as to why this is inappropriate to model this decision as a prisoner’s dilemma game, I would like to hear it.

I ignore uncertainty? How does this affect the outcome in a relevant way? As Zimbalist ought to know, any game theory model can be tweaked with numerous assumptions that drastically alter the results. I’m not sure what type of uncertainty here is going to allow players to unilaterally decide to quit using steroids. I’m all for seeing alternate models, but I think the standard PD model works well.

On the invention of new statistical methods for scouting baseball players, Zimbalist writes:

On page 147, for instance, Bradbury writes: “Old ways and old scouting methods may disappear, but the end result is a good one for the fan: better and cheaper baseball.” No team is contemplating the elimination of traditional scouting methods, nor is one likely to in the future. New statistical methods have been employed to supplement, not supplant, traditional scouting.

Here, Zimbalist takes my words out of context and ignores my direct commentary on the issue.

TBE, p. 144.

Am I predicting an end to traditional scouting? I sincerely doubt that any baseball team, including any team run by Billy Beane, will ever abandon personal scouting. No matter how far the knowledge of scouting progresses, there will always be things that the stats will miss. No matter how far the knowledge of predicting talent progresses, there will always be things that the stats will miss. Teams can learn from statistics what information they really need from scouts, and statistics can make scouts more effective…. [Skipping over a detailed example of how pure stats scouting will fail]…Statistical analysis does not eliminate the need for on-site scouting, but the role of the scout may be reduced and modified to focus on different things.

I then go on to discuss that the development of new statistics is just one area where some teams innovated, and I point out that scouting-centric organizations like the Twins, Marlins, and White Sox have also experienced success on small budgets.

The quote that Zimbalist plucks from the end of the chapter is in reference to a discussion of “creative destruction”—the replacement of old and inefficient methods with new technology. I am explaining why the change is a good thing, and that if and when some scouting methods become obsolete that this is all part of a positive process.

On clutch hitting, Zimbalist asserts that I am not backing up my claims and that my arguments are inconsistent:

“The problem is that hitting with RISP is not a skill … but a statistical anomaly.” (p. 155) “If hitting with RISP is something a hitter can purposely alter, I have a hard time believing he is holding something back in non-RISP situations” (p. 156). There you have it – there is no such thing as clutch hitting.

Yep, “there you have it.” That is it, I just state this and keep on moving…What? Oh wait, there is an endnote (67) listed at the end of the sentence quoted on page 155? Why, whatever could it say?


TBE
, p. 320.

67. Statisticians Jim Albert and James Bennett find minimal evidence for clutch ability in these situations, and any observed effect is very weak (Curve Ball, 2001).

There you have it! Though I didn’t go into great detail on the existence of clutch play in athletics—I decided against diverting the chapter from its main goal to discuss “clutch” play—the endnote that I provide is sufficient.

Zimbalist continues:

This is an awfully linear, materialist view of the world where a player’s emotions and his state of physical depletion over a 162-game season play no role.

Linear? Maybe. Materialist? This doesn’t even make sense. Are we referring economics or metaphysics? As best as I can figure, I think he’s referring to the kind that is opposed to idealism. Still, I can don’t see how being a materialist denies clutch hitting.

Zimbalist continues:

Further, Bradbury is being inconsistent. In his chapter about the on-deck batter, he asserted that a pitcher can ramp it up and pitch more carefully and effectively to the current batter when a strong batter is on deck. So, in Bradbury’s world, pitchers can focus and pitch in the clutch, but hitters can’t turn the same trick.

There is nothing inconsistent about this position, which I do hold. In fact, I have discussed it many times with many people. Pitching and hitting are very different skills. Pitchings requires a player to regulate his effort to remain in the game. A batter swings the bat then sits down for a while.

Zimbalist doesn’t like my analysis of a curious finding in Moneyball.

Later in the chapter, Bradbury endorses the proposition from Moneyball that on-base percentage (OBP) is three times as important as slugging percentage (SLG). He arrives at this outcome by running a multiple regression of runs on batting average, OBP and SLG. The coefficient on OBP is almost three times that on SLG. The problem here is not only that the arguments are collinear and the coefficients are less reliable, but that SLG (it counts a homerun as four hits, a triple as three, etc.) is a much higher number than OBP. The coefficient, therefore, will necessarily be smaller on SLG. If elasticity is used instead of the estimated coefficient, OBP is 1.8 times greater than SLG.

The argument is the relationship between each “point” of OBP is worth nearly three-times as much as a “point” of SLG. The elasticity, which I report, is irrelevant to this sidebar to the main discussion. I understand that SLG is greater than OBP for most players, but the debate is how much different is the ratio than 1.5–1.8.

And please, no lecture on econometrics here. What does collinearity affect? The standard errors. Because both coefficient estimates are statistically significant it is important to keep both in the model. And even if this was inappropriate, I am merely trying to replicate something done by someone else. There are cases where collinearity is so extreme that two variables should not be included in the same regression: this is not one of those cases.

Also, let’s view this in light of his critiques of my econometric methods elsewhere. When I exclude something, I am creating omitted variable bias. But, when I include something I have multicollinearity. It’s hard for me to win here.

On my measurement of pitching skill, Zimbalist isn’t a believer in DIPS:

Bradbury also discusses the assessment of pitching skills in this chapter. The main argument here is that a pitcher’s ERA from one year to the next is highly variable, but that a pitcher’s walks, strikes and home runs allowed are more stable over time. The inference is that ERA depends more on outside factors, such as a team’s fielding prowess, and, hence, is a poor measure of the inherent skills of a pitcher. While there is something compelling to this logic, it seems caution is in order. First, a pitcher’s skills may actually vary from year to year, along with his ERA, as other factors change, such as, his ballpark, his pitching coach, his bullpen, his team’s offense, the angle of his arm slot, his confidence level, etc. This variability does not mean that the skill is spurious.

I disagree. It most certainly does indicate a lack of skill, especially since the other metrics do not exhibit this same variability. If he believes this, I suggest that he write up a response to my paper in this month’s Journal of Sports Economics in which I go into the full details. On page 170, endnote 71 points to the citation of this paper on p. 321. Though the paper was not published until recently, I would have been happy to provide a copy if asked. I presented the paper at the Southern Economic Association Meeting a few years back and no one there mentioned this particular objection. Neither did an anonymous referee at JSE. I’m going to see more than this “does not mean the skill is spurious.”

Zimbalist continues:

Second, if all we consider is strikeouts, walks and home runs, what are we saying about sinkerball pitchers who induce groundballs or pitchers who throw fastballs with movement or offspeed pitches that induce weak swings and popups?

I am saying that pitchers have almost no effect on hits on balls in play, and that sinkerball and offspeed pitchers are good because of their strikeouts, walks, and home runs, not because of any effect they have on balls in play.

TBE, p. 167–170.

“Everyone knows that Greg Maddux is so good because his pitches produce easily fielded balls.” …[Insert lengthy rebuttal of this argument]…It turns out that the real reason Greg Maddux is so good is that, though he is not an overpowering strikeout pitcher, he rarely walks batters or gives up home runs.

That is why DIPS is so controversial. Of course, then I argue that pitcher do appear to have some effect on balls in play, but this impact is captured in those stats. DIPS has been widely debated in the sabermetrics community since the late-1990s. I make numerous references to this in the chapter, including a variant of the argument introduced by Bill James in 1987. Again, Zimbalist seems to have glossed over my explanation.

Zimbalist feels my marginal revenue product estimates are fatally flawed.

Next, Bradbury offers a chapter on the worth of a ballplayer. He gets off to a bad start here by misrepresenting the functioning of the players’ market and the terms of the collective bargaining agreement. He then misspecifies his team revenue function, leaving out RSN ownership, the number of large corporations in the host market, the size of the team’s assigned television territory, among other factors.

I’m not sure where I’ve erred in my understanding of the CBA. I laid it out there pretty simply, so maybe I sacrificed some precision for expediency in getting to the argument. But, since Zimbalist doesn’t say, I can’t really respond.

All revenue estimates come from the Forbes Business of Baseball Report. Are these estimates perfect? No. The best option would be to have detailed financial data provided by all 30 teams, but teams are a more than a little reluctant to provide this stuff. Do they tell us something about the earnings teams take in? That is an open question, to which I believe the answer is “yes.” I think, though I am not sure, that he is suggesting that I need some more control variables for RSN ownership, but this is just weird. If winning leads to more revenue as it comes through RSNs, the the coefficient estimates ought to reflect this impact. Are omitted-variable distortions possible? They always are, and I see no obvious reason that the absence of RSNs—especially, since most teams now have them—will bias the estimates.

In the end, this is a simple model, necessitated by the availability of the data. The test of any model is how well it predicts, relative to alternative models. Recently, Zimbalist told a local Atlanta writer that it was “utterly preposterous” for Scott Boras to expect Andruw Jones to get $20 million/year. Well, he got $18.1 million, even after tanking in 2007—Boras’s $20-million/year request was made before the 2007 season. I think I did pretty well with Forbes data and a few assumptions.

Zimbalist continues:

But the fatal problem is that Bradbury’s methodology unwittingly identifies a player’s average revenue product, not his marginal revenue product. By his reckoning, all of a team’s revenue is attributed to the players, leaving nothing left over for front offices expenses, stadium expenses, minor league operations, or profits. Given this misstep, it is not surprising that Bradbury finds players at all levels (under reserve, arbitration eligible and free agents) are paid less than what he estimates they are worth.

This is just incorrect. My estimates most certainly are marginal. Those multiple-regression coefficients are partial derivatives, so I’m not sure how to respond to the statement that these are not MRP estimates.

I clearly acknowledge that there are other inputs to winning—after all, I did write the Mazzone chapter. How should we solve the fact that OBP and SLG contain contributions by coaches and players? There is no simple answer, but it is one that I address in the text. Given that players earn far more than coaches, I don’t think it’s too much of a simplifying assumption to initially give all of the credit to the players, then extrapolate from the estimates. At the end of the chapter, I state that these gross MRP estimates do not account for resource costs, which include these other factors.

TBE, p. 197.

If teams are willing to pay player salaries equal to the value produced on the field minus other training costs, then our gross MRP estimates should be above actual player salaries.

In conclusion to this particular argument, I will put my estimates up against Zimbalist’s method in his book Baseball and Billions (1992) any day of the week. These are the ones where he takes Gerald Scully to task for using the strikeout-to-walk ratio to measure pitchers—I show that Scully is right and Zimbalist is wrong in Chapter 12. These estimates also measure hitter value using a measure known as “PROD”. PROD is the sum of OBP and SLG—yes, that is OPS—yet Thorn and Palmer (1984) are not credited. [Zimbalist disputes this last point, see this post below and the comments section.]

Now onto the monopoly section.

Zimbalist feels that I have mis-stated the history of the jurisprudence regarding the origin of the antitrust exemption. I do not have time to investigate this at the moment. Without going into all of the detail, I think it is pretty clear that any mistakes that I did make (I’m not saying he’s right, either) are not relevant to the main argument. There is an exemption that arises from legal precedents, and I think its effect is minimal.

Zimbalist continues:

Bradbury then distorts the record further by asserting (p. 205): “At the heart of the argument that MLB acts like a monopolist is the existence of the antitrust exemption.” He cites no sources for this claim, because there are none.

He wants a source for this? Watch any news report discussing the business aspects of baseball. And several sports economists strongly advocate removing the exemption, so it is clearly an issue.

Zimbalist continues:

Each team sport league is a monopolist because it is the sole producer of its product and has no close substitutes.

Chapters 15 and 16 of TBE are devoted to rebutting this claim. Chapter 15 is not discussed and I address his problems with Chapter 16 below.

More Zimbalst:

While the value of baseball’s exemption today is not what it used to be, there is still a good case to be made that MLB’s minor leagues and perhaps its amateur draft could not exist in their present form were it not for the exemption.

The NBA, NFL, and NHL all have amateur drafts, yet lack the antitrust exemption.

Zimbalist on contestable markets:

Bradbury’s last essay argues that the market for top-level professional baseball in the United States is contestable. If this were true, then the earlier question about whether or not MLB is a monopoly might be moot. Here Bradbury makes two points. First, if there is an aspect of the industry that is not a natural monopoly and, hence, constitutes an artificial barrier to entry, it is the subsidies from local governments that teams receive for the construction of their stadiums. But, he avers, this is not really an issue because (p. 220) “the public does not seem averse to subsidizing major sports teams from leagues other than the dominant existing league.” It is clear that Bradbury has never been involved in starting a new or non-dominant league.

He’s got me there, I have never been involved with starting up a new sports league. Maybe by the time I am 60 it will happen. I’m not sure why this is relevant except that maybe Zimbalist fondly remembers his own involvement in the United Baseball League in the mid-1990s. This was a league that tried to start up at the time MLB was half-way to expanding the league by four teams. I admit that it is difficult to start a new league when the old one is expanding: that is the gist of my argument. In this chapter, I argue that MLB must expand to meet the needs of fans or another league will rise up, and that after years of responding to entry the league now acts to prevent entry. This seems to prove my point. I guess I should thank Zimbalist for reminding me of this, I’ll be sure to include the example in the next edition.

Zimbalist continues:

His notion that politicians are not averse to providing subsidies to teams from these upstart leagues is just plain wrong.

Let’s look at the text surrounding the quote that Zimbalist offers, in which I am responding to the potential objection that citizens will be less willing to subsidize major league teams in a rival league.

From TBE, p. 219–220.

First, [the argument] does not apply to cities without current MLB teams. Teams in competing leagues may be just as successful at extracting public subsidies as MLB if they promise to bring major-league-level baseball to town. Plus, even if the costs of operating a team are higher in a city without an MLB team, the new team doesn’t have to worry about its fans migrating to a crosstown MLB rival. A slightly inferior product my still yield sufficient revenue for the owner to purchase major-league talent. This puts these teams in a rival league on competitive footing with MLB.

Second, the history of competing leagues does not reveal any public bias toward the public funding of stadiums of stadiums for new leagues. Many of the teams in the United States Football League (USFL), which competed with the NFL in the mid-1980s, played in publicly financed stadiums. In fact, many USFL teams shared stadiums with NFL teams. History shows that the public’s willingness to subsidize teams extends beyond the dominant league brand. The point is not that MLB teams could share stadiums with rival leagues—I think this would be highly unlikely—but that the public does not seem averse to subsidizing major sports teams from leagues other than the dominant existing league. It seems that as long as a new league promises to pursue top-level talent, as the USFL did in football, citizens will subsidize new teams. So public subsidization of stadiums doesn’t appear to be much of a barrier to entry in the baseball market.

I have highlighted the one sentence of two paragraphs in which I justified the quoted statement with theory and history. It is fine to disagree with me, but at least offer an explanation. Furthermore, Zimbalist also makes no mention that I offered any explanation. Previously, Zimbalist has picked nits, misunderstood my argument, or was too lazy to find my rebuttal; this is a deliberate misrepresentation of my argument.

Zimbalist also finds fault with my use of history as evidence of competitive pressure.

Second, Bradbury goes on to argue that MLB’s market is contestable. He does this by discussing the emergence of the American Association in 1882 and the American League in 1901. He further adduces what he erroneously calls the “Central League” (real name: the Continental League) forcing baseball to expand the number of its teams in 1961. Leaving details aside, the difficulty with Bradbury’s claim is that the industry’s economic structure today is very different from what it was 57 or 120 years ago.

In regard to the Central League gaffe, mea culpa. This is quite embarrassing, and it was one of the first mistakes I noticed after publication. I’m not sure how this happened, since I know the correct name, and I noted this error in the errata (link on the right sidebar) many months ago.

I hypothesize that baseball hasn’t faced any actual competition since the Continental League threat because it has chosen to expand before such threats arise. That is what constestable market theory says ought to happen. The mere threat that someone might enter is sufficient to force a single firm to behave. It is quite difficult to miss this point, especially given the title of the chapter, “Expansion and the Invisible Hand.”

It’s fine to disagree with me, but I did not ignore the difficulty using historical examples to shed light on the present. Once again, I addressed this argument in the text when Zimbalist insinuates that I did not.

TBE, p. 227.

Has the competitive pressure that fueled interleague baseball competition in the past evaporated, or does that competitive pressure lurk in the minds of owners, who fear the entry of a new outlaw league? The fear probably still lurks. We have plenty of rich men and women looking to be loved by baseball fans across the continent. I doubt that our current stock of eccentric wealthy egomaniacs could leave large quantities of money and public adoration alone. And I have a feeling that the current stock of owners who are members of this crowd are far more familiar with this than the average baseball fan realizes.

In conclusion, I have to say that I am disappointed in Andrew Zimbalist. His demeanor and thoroughness is out of step of what is normally expected in academic discourse. It is hard for me to learn anything from critiques that are largely based on misunderstanding or flat out ignorance of my arguments. The arrogant tone is unnecessary and inhibits constructive discussion of these issues.

One of my goals in writing this book was to push economic theory to extremes within the sports economics community. At the time I began writing, I felt we were settling for too many old truisms (e.g., MLB is a monopoly) that might not be completely correct. It is not that these claims are necessarily wrong, but I wanted to raise some possible objections. It was my hope that others might respond to my ideas with new insights that might confirm or reject my hypotheses. And in the end, I might feel more comfortable holding certain beliefs.

I would like to reiterate that I am fine with criticism, and I encourage it. As I wrote in the Epilogue,

Applying my professional training to the game that I love has taught me new, unexpected lessons. Notions that I held about the game turned out to be false. I thought Leo Mazzone was overrated as a pitching coach and that batters protected one another in the lineup. I’m happy to have eliminated some of my ignorance. …

Unsatisfying answers provide opportunities to expand knowledge. I hope that if you find any of my conclusions to be unsatisfying, this will motivate you to search for better ones.

It is too bad that Andrew Zimbalist did not heed my advice.

Update: Zimbalist offers a response, of sorts. On the whole it is more of a general overview of his opinion of the book than a response. I do want to acknowledge that Zimbalist asserts that he did cite Thorn and Palmer (1984). I no longer have a copy of his book, so I cannot verify this. My assertion that he did not was based on a recollection of mine. If I am incorrect, I will retract my statement.

15 Responses “Responding to Zimbalist”

  1. JHamrahi says:

    Being an economist and accountant myself, I feel compelled to defend your work JC. But I think you answered the questions Zimbalist raised quite nicely. Like you said, it seems to me that he just jumped over portions of the book.

  2. tangotiger says:

    I appreciated the Birnbaum review of JC’s book (click on my name).

  3. Cliff says:

    To find this cite took me 10 minutes or less on Lexis Nexis.

    Federal Baseball Club, Inc. v. National League of Professional Baseball Clubs, 42 S.Ct. 65 (1922).

  4. JEC says:

    Reading Zimbalist’s criticisms leads me to think he came up with his conclusions before he read the book. Then he cherry-picked sections to validate his pre-determined conclusions, which is shameful and all too common with reviewers these days.

  5. Phil says:

    The more I read Zimbalist’s arguments about DIPS, OPS, and clutch hitting, the more astonished I am at how he doesn’t seem to have the most basic grounding in what sabermetrics has been doing over the last decade or two.

    Has he even heard of sabermetrics? He seems to completely fail to understand that JC’s descriptions are indeed based on solid research findings. It’s like he thinks that the rejection of clutch hitting talent, or the DIPS theory, is just JC’s random opinion. And so he throws his own opinion out instead (and those, unlike Bradbury’s, ARE opinions, and incorrect ones), and charges forward as if Bradbury is wrong on the facts. That’s strange for an academic who has no evidence to back him up.

    The editors of the journal that published this should be more than a little embarrassed … the slightest bit of peer review from someone who knew something about sabermetrics would have caught this stuff. There is really no excuse, in my opinion.

  6. Mark says:

    On the “clutch” hitting vs. “clutch” pitching argument: Pitching is an action and hitting is a reaction, so a pitcher can absolutely change his approach, throwing more strikes and reducing walks and even throwing harder or sharper reducing power, while a hitter will most likely have a similar reaction no matter what the situation. A hitter who buckles under the pressure of RISPs probably will be weeded out before reaching the majors. Any fluctuation in RISP is probably random and will even out close to a hitter’s career skills the more the at-bats pile up.

  7. Rory (mr. man) says:

    I read Zimbalist’s defense; it’s now linked on BTF. Here’s my comment from there (BTW, why doesn’t Biz of Baseball have a comments section?)

    I’m not impressed. Though I know Zimbalist is not exclusively a sports economist, he seems to be the leading academic in what is an awfully small field. For him to claim to be an expert on baseball economics without an intimate familiarity with James’, and especially Pappas’ and Silver’s work (he seems to be totally unaware of the existence of their work) is ignorant.

    The ‘Scully Model’ that pervades academic baseball economics is outdated; it was ahead of its time in 1974 but has since been passed by leaps and bounds by Pappas’ MW/MP work and now Silver’s MORP. By all means, sabermetrics shouldn’t be construed to be the be-all-end-all of player evaluation (who wants to place odds that Zimbalist hasn’t even read beer and tacos?), but it’s miles ahead of the Scully model, and though no one would suggest you run a baseball team solely by the numbers, no one would ever suggest operating without them.

    Zimbalist (or some other sports economist) could make a career out of putting the sabernomic theory to the test at a rigorous, academic level. To be unaware of its importance is unacceptable for someone who calls himself an expert in the field.

  8. joe arthur says:

    You are partly correct about the citation of Thorn and Palmer. You expected Zimbalist to cite the Hidden Game of Baseball (1984) in which OPS is defined. In Total Baseball (1st edition 1989) Thorn and Palmer presented this value as “PRO.” Zimbalist clearly used this book as his source, not Hidden Game, and he does cite it in a footnote to a paragraph in his appendix A in which PROD is discussed. However the reference to (Thorn and Palmer 1989) is not specifically as his source for “PROD”, but instead in relation to linear weights. I have not reread Zimbalist’s book and there may be some more general allusion to Total Baseball as his source for baseball statistics somewhere.

    I don’t care for Zimbalist’s unqualified assertion that he did cite the book either (his citation is like a neighborhood play at 2nd base), and I understand that you were in kind of a footnote war with Zimbalist because he implied you made an undocumented assertion about clutch hitting, when you had documented it, but there’s a difference here. Your footnote charge is more serious. I don’t think that you should have made what amounts to a charge of plagiarism based on recollection alone.

  9. ZimbalistFactChecker says:

    I checked Baseball and Billions for Thorn and Palmer (1984) and found:

    –neither author listed in the index (unless Zimbalist counts Arnold Palmer or Jim Palmer as one of the authors) :-)
    –Thorn and Palmer, Total Baseball, 1989 (not 1984) listed in the references section
    –no reference to Thorn and Palmer (any pub, any year) in Chapter 4 on player performance and salaries
    –Thorn and Palmer (again 1989, not 1984) cited in endnote 5 of Appendix A. In this note Zimbalist says he uses PROD (aka OPS) instead of a weighting concept proposed by Thorn and Palmer (1989).

  10. joe arthur says:

    ZimbalistFactChecker -
    all correct but I think misleading without context.
    JC says Zimbalist says he cited the 1984 book, but Zimbalist does not say that. Zimbalist says he cited the 1989 book. Which he did, though not specifically acknowledged as the source for PROD itself. Zimbalist never claims to have cited Hidden Game of Baseball (which would be Thorn and Palmer 1984). JC asserts that he should have cited this particular book, though he identifies it by year only and not name. The basis of this assertion is not clear – there are two possibilities.
    1) JC believes Zimbalist took the concept from this particular book without attribution (plagiarism). [The concept does reappear in Total Baseball [1989] though, and the player data for PRO in Zimbalist’s table in fact probably came from the 2nd edition of Total Baseball (1991), since it is data for the 1989 season.]
    2) he believes that Zimbalist implies his independent invention of the measure without properly acknowledging earlier work.

    In either case, the prior question is whether the concept had become sufficiently public domain to require citation in the first place. You aren’t ordinarily expected to footnote the original inventors of OBP or ERA, for example. OPS (PRO) caught on quickly, in newspapers and other books, popularized in part by its appearance in the sabermetric-savvy encyclopedia Total Baseball(1st edition 1989), also by Thorn and Palmer. Clear citation of that book would be sufficient, and that (1989 book) is what Zimbalist has claimed in his reply. Also JC is on record from last year as arguing that it is not necessary in academic work to cite non-peer reviewed, non-academic work.

  11. joe arthur says:

    oh and Zimbalist chapter 4 (p.90) does direct the reader to appendix A for further discussion of player performance measures. So the reference within the appendix is the only place you might “require” a reference.

  12. ZimbalistFactChecker says:

    The argument that PROD/OPS is in the public domain is reasonable to me. That, however, was not Zimbalist’s response.

    JC said Zimbalist failed to cite TP84: “These estimates also measure hitter value using a measure known as “PROD”. PROD is the sum of OBP and SLG—yes, that is OPS—yet Thorn and Palmer (1984) are not credited.”

    Zimbalist responded that he had cited TP (but gave no year–joe arthur is therefore incorrect to say that “Zimbalist says he cited the 1989 book”). Here is the relevant paragraph from Zimbalist’s response”

    In any event, I am simply not persuaded by some of the specific arguments that often come out of that literature. As part of Bradbury’s rebuttal, he asserts that I used PROD back in my 1992 book, Baseball and Billions, and that I did not cite Thorn and Palmer. That’s incorrect.

    I then pointed out that Zimbalist had cited a different work by TP and, more importantly, that he had cited TP89 not as a source for PROD/OPS but as a source for an alternate (linear weights) measure of player performance rather than the PROD/OPS measure that Zimbalist used in his book.

    There is nothing “misleading without context” in this sequence. If anything, the misleading parts are Zimbalist citing TP but not specifying which year/book and joe arthur following up by claiming that Zimbalist had cited a specific work (which, as noted above, he did not do).

  13. joe arthur says:

    ZFC – principally what I meant was misleading about your post is that talking about The Hidden Game of Baseball (1984) is nothing more than a red herring. My own phrasing was misleading, you’re right. I blended together his recent online reply with what Zimbalist actually did in his 1992 book. I should have said that “Z says he cited Thorn and Palmer, but the Thorn and Palmer book he was talking about was the 1989 book.” They are talking past each other about what the book is, not surprising since JC did not give a full citation, with title, in his initial response. I thought I indicated that they were talking past each other clearly enough in my post #8 before yours, and I criticized Zimbalist for implying that he had cited it (as a source for PROD) when he really cited it for something else.

    I give Zimbalist a complete pass on the confusion about the book. He claims to have cited “Thorn and Palmer”, and he did cite it(albeit not exactly on point to JC’s charge) and it does indeed contain the relevant material.

    Nor do I believe JC is right to criticize him for failing to cite the 1984 book anyway. Now that I have looked at the Hidden Game itself (p.51), I see that Thorn and Palmer in fact credit the OPS formula to Dick Cramer in 1972; his name for it was Batter Run Average (BRA).

  14. joe arthur says:

    my earlier reply to #12 seems not to have appeared – I don’t know if it’s still awaiting moderation from many hours ago or if I didn’t properly submit it.

    New information first: Hidden Game of Baseball [= Thorn and Palmer 1984] (p51) gives credit to Dick Cramer in 1972 for the formula which we now call OPS. [And a closely related version of the formula was constructed even earlier, in 1964, by Earnshaw Cook]. So there is no fair reason to say that Zimbalist should have cited this particular book in the first place in connection with the statistic “PROD.”

    In my followup post (#10), I will agree with ZimbalistFactChecker that my sentence “Zimbalist says he cited the 1989 book.” is misleading, because it sounds as though I am saying Zimbalist in his reply said 1989 expressly, which he did not do.

    I don’t intend to quibble any further with ZimbalistFactChecker on this; we do agree on the facts about the citation dispute, but disagree very much on interpretation, mine being that there is no basis to be using Zimbalist’s failure to cite the Hidden Game of Baseball as evidence to criticize him. The context which I said ZFC#9 lacks is available in my posts 8,10 and 11. The key point here is that I have read both Thorn and Palmer 1984 and the relevant portion of Thorn and Palmer 1989 and can say with confidence that the relevant material is in both books, and only referred to as PRO in the 1989 book, and that if a citation for PROD were necessary at all, citation of the 1989 book alone would suffice perfectly. Because of the similarity in the references, it is understandable that Zimbalist in his reply could fail to understand that JC was really talking about a different book. I don’t care for the social science style of bibliographic citation; it leads too easily to honest confusion even if you attend to the year but don’t recognize it – [different work I don't know? earlier edition of same book? typo in the citation?]

  15. JC says:

    Thanks to Joe and ZimbalistFactChecker. (Joe, comment 12 was stuck in the spam filter.)

    In an effort to be conciliatory, I am going to strike the statement and refer people to the comments. You are free to form your own opinions. I am also going to close down the comments, because the Mitchell Report is keeping me very busy. I stand by my response. If you have any further comments, you may direct them to my personal e-mail.

    Here is the endnote in question from Baseball and Billions in which Zimbalist explains his use of PROD.

    5. The measures of performance used in this exercise are, of course, imperfect. Home runs enter slugging percentage a four times more valuable than singles. A single and a walk enter equally. Stolen bases and caught stealing do not figure in our measurement at all. Kevin Ryan (1991) has suggested using a purer measurement based on Thorn and Palmer (1989). They ran computer simulations of over 100,000 baseball games and estimated the run values of each type of play. For instance, a single is estimated to be worth .47 runs, a walk .33 runs, a double .78 runs, a triple 1.09 runs, a homer 1.40 runs. We used the “run value” of performance measure and came up with very similar estimates. The fit of our regressions was not improved.

    At issue is not the actual source. Thorn and Palmer (1989) or Cramer (1972) would have have been acceptable citations for justifying the use of PROD. However, the statement above references this as the source for an alternate measure that was not employed, not PROD. And I do not believe this necessarily rises to a charge of plagiarism. In regard to the standards Zimbalist used to criticize my justifications of statements in my book, my treatment passes well above the threshold set to justify criticism. That was my main point in making the statement. But again, I am striking the statement, and any future readers of this post will be directed to the comments for an explanation.

    Also, I never said that academics need not cite non-academic work. Here is my response to what I believe Joe is referring to.