clock menu more-arrow no yes

Filed under:

Continue to think like a GM. Some used to play.

New, 3 comments

Matt Corbett wrote a fantastic piece that urged baseball fans to think more like a player and less like a general manager. However, these perspectives may not be so opposing.

Zack Greinke, seen here daydreaming about FIP
Zack Greinke, seen here daydreaming about FIP
Mark J. Rebilas-USA TODAY Sports

It has been nearly 50 years since Star Trek first appeared on television. More specifically, it's been 49 years since the automatic sliding doors on board the USS Enterprise NCC-1701 or the touch-button communicators that the crew used made their debut. Star Trek is frequently credited with leading to the invention of many things that are now enjoyed by contemporary society. Some even give the writers of the show near-prophetic credit for foretelling technological advancements. But, of course, it could also be true that the future inventors of automatic sliding doors happened to grow up watching Star Trek and therefore had their ideas shaped by the popular science fiction show.

Over at FanGraphs, Matt Corbett wrote a fantastic article in which the thesis -- and forgive my paraphrasing -- was: 'sabermetrics encourages people to think like management, which creates discourse that doesn't help the actual playing of baseball.' Corbett views this as a problem because "[g]etting better at playing involves countless mechanical repetitions with the goal of honing one’s neurology to the point at which certain tasks no longer require conscious attention to perform." These are difficult points with which to disagree.

In pursuit of that thesis, Corbett addresses a few specific areas of sabermetrics. First, the idea that some sabermetric phenomena held to be "real", like pitch framing, were already perceived as valuable before sabermetrics came into the fold. Second, because sabermetrics deals with selection bias and uncontrollable environments leading to gratuitous assumptions, true science isn't really possible in the field. Third, sabermetrics should stop focusing on analyzing outcomes and start analyzing true causal mechanisms (actual skill). Don't worry - I'll come back to Star Trek.

Assigning credit

A sizable portion of the article discusses advancements in pitch framing. Corbett does a pretty good job of showing first, how little sabermetrics cared about pitch framing, second, how sabermetrics started caring about pitch framing, and last, how sabermetrics demanded credit for the popularity of pitch framing. While paraphrasing in this way may seem out of context, Corbett addresses the fact that he uses Baseball Prospectus as a "rough proxy," so it isn't entirely unfair. Corbett re-assigns credit where he thinks it is due, to the "grizzled scout, [the] career baseball guy, [the] former-player color announcer who knew in [their] bones and always insisted that a catcher's receiving ability was crucial." Corbett also asserts that sabermetrics was essentially wrong for belittling scouts for their beliefs about catcher defense and then flipping about catcher defense and never apologizing for it. So why did I start this piece by mentioning Star Trek?

Let's follow one linear narrative for a moment. Gene Roddenberry and his team of writers all sit in a room imagining the future. One -- perhaps even Roddenberry himself -- suggests that there will be automatic sliding doors. They all agree on the idea with such vigor that they not only implement it in the show, they choose not to address the doors' presence in the show. The characters -- who are played by actual 1960s humans -- are to act like automatic sliding doors are the norm (which they aren't) and never to discuss them. Of course, Lee Hewitt and Dee Horton had already invented the automatic sliding door by the time Star Trek first aired, but they weren't made popular yet. As a matter of revisionist history, Star Trek fans -- 'Trekkies' -- declare their show the true inventor of pitch framing automatic sliding doors.

Of course, you understand now that Hewitt and Horton are the grizzled scouts who had their idea "stolen". That makes Roddenberry et al. Baseball Prospectus, who were arguably the first to show the world the true implementation of a great idea. It's not really Roddenberry's fault that Star Trek takes credit for the automatic sliding door, but you wouldn't really blame him for failing to admit that he didn't invent it. After all, he may have made it popular.

In the same way, the "grizzled scout" would have bet presumably large sums of money that pitch framing did exist and was worth something. Sabermetrics, still young and finding its place in the baseball world, bet against the grizzled scout. That was until Mike Fast -- a 'sabermetrician' for the purposes of Corbett's essay and this response -- decided to find out if pitch framing was worth anything. However, when Fast actually found there was a value to pitch framing, he didn't claim it as a victory for sabermetrics (as Corbett's essay implies); Fast claimed it for Fast, rightfully so since he demonstrated the value numerically. What's more, Fast's research hinged heavily on PITCHf/x data which, again, was made possible by a person -- who may or may not self-identify as a sabermetrician but still contributed (and still contributes) to actual statistical analysis of baseball -- who asked the question 'can I know this?'

Sabermetrics as a science

That question brings up another point. I don't know Mike Fast at all, so I don't know if he identifies as a sabermetrician per se. However, based on his work and his seeming quest to know things through observation, that does make Fast a scientist. The same can be said of Beyond the Box Score alum Dan Turkenkopf who also worked to quantify pitch framing using PITCHf/x.

What isn't actually clear is why this even matters. To me, there is no denying that sabermetrics is a science. To Corbett, sabermetrics apparently isn't a science. To you, however, it shouldn't matter what anybody defines -- or doesn't define -- as science. While Corbett makes good points on how some self-identified sabermetricians carry out their research in unscientific ways, that doesn't make the entire field pseudo-science. A pseudo-scientific method would be to use circular and self-proving logic. A bit like that "grizzled scout" said in Statement B of Corbett's article: "He is letting his hands drift too far away from his body, so pitchers are busting him inside, and he’s popping up what he isn’t whiffing." Identifying what's wrong isn't the problem with Statement B; the implicit solution -- don't let your hands drift too far away from your body -- is the problem because it affirms the consequent.

Sabermetrics follows observation, estimation, quantification, and trial-and-error like many other sciences. Nothing exemplifies this better than Fast's ability to take on a topic that sabermetrics (personified) had already ruled didn't exist. When Fast realized he had proved the intelligentsia wrong, he published those findings, which were welcomed by the intelligentsia. Those are good scientific practices.

Going back to assigning credit, the point of science is to determine if an observed phenomenon is real or not. The "grizzled scout" -- really a group of scouts reaching a consensus -- believed the phenomenon to be a real thing; however, without Fast's scientific examination, it wasn't really proven. Can catchers repeat the skill? Can they get better at it? These are questions that sabermetrics can help answer. Pitch framing is an example of a phenomenon held to be true upon further examination; there are examples where using science disproved previously held truths. In reality, everyone is working toward the same goal - understanding how the game really works. Not only does assigning credit not really matter, but trying to discover truth is at the heart of science.

Parallel Universes

Lastly, Corbett addresses this problem through a very specific lens. He self-identifies as a former player -- though he also begins with the caveat that he's "[l]ike many baseball fans." In this opening, Corbett appeals to both authority and flattery, though not necessarily fallaciously. He does so for this payoff: "Sabermetrics provided essentially no help in making me a better baseball player." This is perhaps the boldest claim I have ever read. To say this is quite the meta-analysis of one's upbringing. He would not only have to understand an upbringing in which sabermetrics was not involved, but also know beyond any doubt that he would have been at least as good at baseball in that parallel universe. This point is raised because Corbett believes the playing of baseball is "at heart an intuitive [task]." Anybody who refutes this remark would have to possess a near-impossible amount of evidence. However, the ideas of sabermetrics come directly from intuition. It was Bill James' intuition that suggested on-base percentage could actually win more baseball games than previously thought. A baseball player (Joey Votto for instance) can benefit from knowing that fact. It was Voros McCracken's intuition that suggested not allowing home runs or walks is the key to a pitcher's success. A baseball player (Zack Greinke for instance) can benefit from knowing that fact.

Sabermetrics wasn't originally conceived to make players play better. It was a product of fans trying to understand the game. It gets implemented at the front office and management levels because that is where it is most easily applied at the moment. In Michael Lewis' Moneyball, there is even a chapter describing the Oakland Athletics management telling its farm teams to start taking more walks. That means coaches were being forced to understand the value of on-base percentage, not just general managers. However, what this created was discourse on approaches at the plate. Only players stand at the plate.

In addition to the on-base example above, which led to players being more selective at the plate and has caught fire since Moneyball, the rise of the defensive shift is another example. Opposing teams used to shift against just David Ortiz. The Pirates, after bringing aboard Dan Fox, a Baseball Prospectus alum, integrated the defensive shift into their entire minor league training program. While it took some time for players (both fielders and pitchers) to buy in, the introduction of sabermetrics into fielding created discourse that helped improve play on the field. Players can think more deeply about plate appearances: "Based on the count, the hitter and his tendencies, and what my pitcher is trying to do, I have a good idea of where to stand to cover the space where the ball will most likely go." Players have undoubtedly thought about that for as long as they have been playing, but players (and coaches) now have the available data to refute or prove those thoughts. Scientific experiments usually try to generalize findings to a larger population, but there is a lot of baseball analysis where N=1. Russell Carleton at Baseball Prospectus has written about this before. The goal of any examination where N=1, the examination of one player, is to improve play on the field. The example of shifting for one player in a specific way is N=1 type stuff, but the methodology can apply widely.

General managers and coaches are not a dichotomy; they work in symbiosis. Furthermore, Billy Beane is a former player. This isn't a player vs. management discussion anymore. It's a player and management discussion.

Corbett's essay is a superbly written discourse on how some self-described sabermetricians fail to recognize that some of their own opinions are also based in fallacy. His final realization -- that sabermetrics should help create greatness instead of simply measuring it -- is quite true. It's just that sabermetrics is still finding its place in the baseball universe, constantly forced to defend itself. Sabermetrics began with trying to measure the right outcomes (pointing out the inaccuracies of traditional stats like wins, RBIs, and ERA can be considered science, right?), but the natural evolution of sabermetrics as better tools and technology become available will be to evaluate the right process.

To Corbett though, I appeal. You seem to have an impressive grasp on sabermetrics and such a knack for dismissing its value. If you believe sabermetrics can be better, and it can be, don't just tell sabermetrics personified to do it - help sabermetrics through its next stages.

. . .

Michael Bradburn is a Featured Writer for Beyond the Box Score. You can follow him on Twitter at @mwbii or reach him at michaelwbii@gmail.com.