Q&A: Alan Hirsch

Throughout much of sabermetric cyberspace, The Beauty of Short Hops: How Chance and Circumstance Confound the Moneyball Approach to Baseball, is being panned, its co-authors, Alan and Sheldon Hirsch, labeled as backwards-thinking ignoramuses [and worse]. Some of the criticism is merited — the book certainly has its flaws — but looking at its content objectively, it is also necessary to ask: Do the authors make some valid points? In this interview, Alan Hirsch defends, and clarifies, several of them.

——

David Laurila: The first chapter of the book is, “Where Moneyball Went Wrong.” Why do you feel that Moneyball — most commonly defined as “identifying undervalued assets in baseball (often through the use of statistical analysis)” — is a failed approach?

Alan Hirsch: Note that in the title of the first chapter Moneyball is italicized. The chapter concerns where Michael Lewis and, to a lesser extent, Billy Beane go wrong. It’s not until subsequent chapters that we discuss sabermetrics more systematically. In the Moneyball chapter, we do not say that identifying undervalued assets through statistical analysis is inherently a failed approach.

The first part –“identifying undervalued assets in baseball” — is beyond reproach. That’s what everyone ought to do, whether your business is baseball or baseball cards. The second part — “the use of statistical analysis” — is the issue, because that’s Michael Lewis’s central claim: the A’s succeeded because of Billy Beane’s unique insights, relative to other GMs, rooted in sabermetrics. Any team that wins with a low payroll, say the Twins, has identified undervalued assets. Nobody has written a ballyhooed book about the Twins because they do not purport to have a new paradigm for identifying these assets. Thus, our analysis of Moneyball revolves around the insights attributed to Beane.

We go point by point to show that Lewis exaggerated — in some respects that’s a charitable characterization — the extent to which Beane benefited from insights rooted in sabermetrics. For now, two examples should suffice. First, there was his belief that statistics alone could identify top prospects. Jeremy Brown and Brant Colamarino are a good place to start if we’re talking about shortcomings in identifying undervalued assets through statistical analysis. A second example is Beane’s alleged use of Voros McCracken’s theory about how to evaluate pitchers.

DL: The crux of McCracken’s theory is that pitchers have no control over what happens once a ball is put into play — defense, park effects and random chance play a significant role in the outcome — thus strikeouts, walks and home runs allowed are better indicators of performance, especially future performance. What makes this a falsehood in your mind?

AH: What’s false is “no control” and, more importantly, the way Lewis milked it to advance his storyline: Beane’s ingenious use of sabermetrics to gain a competitive advantage. He suggested that McCracken’s theory enabled Beane to find undervalued pitchers, such as a journeyman relief pitcher, Chad Bradford, to whom Moneyball devoted 40 pages. We argue in Short Hops that all of this – the very idea that Beane won because of players like Bradford acquired thanks to sabermetric insight like McCracken’s – is dubious. That pitchers have some control over the outcome of batted balls is established if you look at pitchers’ career data, particularly comparing pairs of pitchers who spent most of their careers with the same team (thus controlling to a large degree for ballpark and defense). Pitchers are a complicated package. They succeed as a result of different combinations of skills reflected in walks, strikeouts, home runs, AND weaker-hit balls that generate easy outs.

DL: You mock Sam Hinke, the head of basketball analytics for the Houston Rockets, for saying, “I care a lot more about what ought to have happened than what actually happened.” Given his role — and the fact that Branch Rickey famously said that luck is the residue of design — should process not be his primary concern?

AH: You can try to use statistics to help your team without regarding sports as a social science laboratory. It’s not just Hinkie. In Moneyball, Michael Lewis describes Beane’s perspective as follows: “The game can be reduced to a social science. . . . It is simply a matter of figuring out the odds, and exploiting the laws of probability” because “baseball players follow strikingly predictable patterns.” According to Lewis, Beane hates watching the games. All general managers try to put their team in position to win, but they don’t all share the Beane/Hinkie attitude that all the good stuff happens before the game. For Beane and Hinkie, the game itself is agonizing (not thrilling or beautiful) because the best-laid plans may be thwarted. Other GMs (including some who utilize sabermetrics) love watching the game; they see themselves as fans as well, interested in the “actually happened” as well as the “ought to have happened.” Beane and Hinkie represent an extreme, but one senses things moving in that direction.

DL: In the “Two Cheers for Sabermetrics” chapter, you laud the importance of on-base percentage but also suggest that walks are overrated.

AH: This has a number of components. The walk itself will only be overrated by anyone who believes in the mantra “a walk is as good as a hit.” Obviously that’s only true of singles, and even then only with no one on base. Virtually everyone realizes that, but they may overlook its relevance to the tradeoff between patience and aggressiveness. A team may be unhappy that Vlad Guerrerro or Adrian Beltre doesn’t walk a lot. You may prevail upon them to be more patient, though it may be that few batters can significantly change an entrenched approach. Even then, the overall result may well be negative. For Guerrero and Beltre, a free-swinging style has been effective, and increased patience could easily translate into lost power.

We also make a broader point that the emphasis on OBP has had a limited impact. We credit sabermetrics with emphasizing OBP and deemphasizing batting average. But even as that recognition has become universally accepted, walks and OBP have not increased. We discuss a number of reasons for that, which is really part of a larger phenomenon: the game’s essential continuity. Except when there’s a major new development (such as a livelier ball, smaller strike zone, steroid use), the game resists huge swings and even the impact of the major developments tends to diminish over time. For example, with only transient variation, the batting average across baseball has remained roughly .260 for 90 years.

DL: In the same chapter, you noted that Kevin Youkilis saw his home run totals steadily rise from 2006-2008 while his walks concurrently decreased. You inferred a direct correlation, citing increased aggressiveness. Given that Youkilis saw his home run rate continue to climb in 2009, while his walk rate returned to its earlier levels, does your example hold water? Can a direct correlation be substantiated?

AH: We can infer a likely correlation, because the two-way shift in numbers was so substantial, from 91 walks to 62 and 13 home runs to 29. The point there was the irony that Moneyball cites Youkilis as a quintessential Beane player — underrated because of his ability to draw walks — but he achieved stardom when he (apparently) sacrificed patience for aggressiveness, an approach that allegedly drives Beane crazy. In 2009, he put it all together.

DL: Much of your criticism seems to be centered on issues specific to Michael Lewis’s book, and your views on sabermetrics as a whole are nowhere near as negative. Is that accurate?

AH: It’s certainly accurate that we don’t confuse Moneyball with sabermetrics. At the end of the chapter on Moneyball we make a point of saying that Lewis’s and Beane’s errors need not indict sabermetrics more broadly. I think it’s also accurate to say that we’re more negative about Moneyball — it’s a great read, but as much fiction as non-fiction. As for sabermetrics, we acknowledge that it rescued baseball from a tradition of ignorance and has increased understanding in important areas. But we also offer a hard-hitting critique.

Let me give just one example. UZR and related sophisticated defensive measurements are increasingly used by major league teams. In Short Hops, we offered a detailed critique of UZR, arguing that its usefulness is quite limited at best. If we’re right, that’s obviously important.

DL : In the “What Makes Baseball” chapter, you go beyond Moneyball and sabermetrics and posit, among other things, that, “Baseball is simultaneously easier to follow and more complex than other sports.”

AH: Yes, that chapter was great fun to write. It takes as its starting point a debate between George Will and Donald Kagan, occasioned by Kagan’s review of Will’s Men at Work, about what makes baseball special. In a nutshell, Will sees baseball as a thinking man’s sport – it’s all about preparation and cogitation. His favorite manager is the cerebral Tony LaRussa and his favorite players are guys like Tony Gwynn who spend a million hours watching videotape in search of every little edge. Kagan scoffs. He says what makes baseball great is romantic heroism – Joe D hitting in 56 straight, Babe promising home runs to boys in hospitals and delivering, the stuff that inspires legend and song. And we say they’re both right and both wrong. They’ve each located something compelling about the game, something the other is wrong to dismiss. But neither has located anything unique about baseball: the virtues they describe are present in other sports as well.

We ask just what is distinctive about baseball — apart from the oft-mentioned absence of a clock. And the answer (perhaps I should say, AN answer) is what you allude to and what we call the game’s narrative richness. There’s more going on in baseball than you find in most sports. There’s the four bases (each with an umpire and player nearby), and sometimes you have action at two or even three on a single play. There are the unique ballparks that affect the game substantially. Even fans sometimes become involved in plays, and of course Mother Nature. Wind and sun often play a role. I could go on and on. And yet, with all the cool and bizarre things that happen as a result of the intersection of all these elements, the game is pretty easy to follow. It unfolds one play at a time and the action is spread out. In other sports there’s less variation in what actually happens, yet it’s largely fog of war.

DL: Not all of the feedback you’ve received has been positive. Has this come as a surprise to you, and what does the response to The Beauty of Short Hops tell us about baseball, and baseball fans, in today’s era?

AH: The issue isn’t so much positive or negative response (when you wade in to controversial waters you expect both and we’ve gotten both) but the harshness and incivility of the response in certain quarters. I’m not sure I’d generalize about baseball and baseball fans, but I think this reflects an unfortunate reality about the debate over sabermetrics. It’s evolved into Joe Morgan and a few allies on the one side and sabermetricians on the other, talking past each other or calling each other names. As you know, and as should be obvious to everyone from this interview, Short Hops is not a one-sided screed against sabermetrics. Yes, it’s a hard-hitting critique, but we use arguments and evidence, not invective. And we credit sabermetrics for rescuing baseball from a tradition of ignorance and advancing baseball understanding. But the response from some has been little different from that directed at Morgan.

If I found anything surprising, it’s how willing people are to condemn the book while admitting they haven’t read it. That’s a shame, because there’s ample room for a constructive dialogue about the issues we raise. Are we wrong about UZR? About Win Shares? I’ll bet that some sabermetricians end up agreeing with much of our critique.