October 11, 2013
Why Sabermetricians Don't Talk About Chemistry
On Thursday’s episode of ESPN’s Baseball Tonight podcast, host Buster Olney, while discussing the Cardinals’ NLDS victory over the Pirates with ESPN’s Pedro Gomez, made a comment that sabermetricians do not often discuss matters of team chemistry or clubhouse culture. (Well, maybe once in a while…) Olney then proceeded to talk about how he believed that one reason for the Cardinals’ success, both within the NLDS and more broadly over the past few years, has to do with the culture that the club has worked to cultivate. Olney cited, among other things, that Cardinals catcher Yadier Molina, himself worthy of some legitimate MVP support this year, is also one of the hardest workers on the team. Olney pointed out that Cardinals management (Tony LaRussa and, later, Mike Matheny) has gone out of its way to specifically ask its star players to set an example for the rest of the team. He reasoned that other players on the team see this sort of commitment from Molina and are inspired to commit themselves to similarly hard work, and pointed out that it’s rare for sabermetricians give much credence to this as a reason that some teams win while others fall by the wayside.
Buster is right that chemistry and culture are not common topics, for the semi-obvious reason that most of what’s referred to as chemistry takes place behind closed doors and isn’t measurable even if we wanted to measure it. There’s probably also a healthy suspicion that some of the common wisdom about chemistry is post hoc, ecological fallacy-driven narrative building that would fail an undergrad research methodology course. The Cardinals won and Molina is a good player and a hard worker. Therefore, Molina’s work ethic must have been the difference between the teams. Science just doesn’t work like that. (That’s also a straw-man argument, but one that I’ve jumped to plenty of times. How unscientific of me!)
It’s often assumed that the problem is that because the matter doesn’t lend itself to being quantified, sabermetrics dismisses the idea of culture affecting performance altogether. There have been times when I’ve been guilty of that, but the real reason that I’m shy to speak definitively on matters of chemistry in baseball is that it’s just so hard to do good science on the matter. In that case, the best thing to say is “I don’t know” and be done with it. Maybe “I don’t know” has been interpreted as “I assert that it has no effect at all.” I would suggest that the former is a much better answer than the latter.
But maybe there’s a little bit of room for Molina’s work ethic after all. There is research in the field of public health that looks at what leads to people adopting attitudes that lead to healthy behaviors (e.g., quitting smoking, eating more healthfully). Baseball teams certainly have an incentive to promote “willingness to work hard at baseball.” There’s research that shows that in other areas of life, a big predictor of whether or not someone will adopt a positive attitude toward a health behavior is whether someone whom they respect also endorses it. The effect for this is much greater than receiving information about the benefits of the behavior. If you want to get someone to wear sunscreen, don’t give them them a pamphlet showing the benefits of sunscreen. Let them have a conversation about sunscreen with someone whom they respect.
In light of this research, public health initiatives within groups—for example, anti-bullying initiatives at schools—have begun to focus on working with “opinion leaders” in the school. The idea is that if you spend your time convincing those few people, they will have a larger effect over the entire culture of the school than will trying to convince everyone. There’s good, solid science behind the idea. Molina would be a good candidate to be an opinion leader. He is good at baseball, and men tend to value competence (along with personal warmth) as markers of someone whom they should emulate. (I have no idea whether Yadier Molina is a nice guy…I’ll give him the benefit of the doubt and assume that he is.) With that in mind, the LaRussa-Matheny strategy makes complete sense. Tony LaRussa seems to have identified Molina as an opinion leader and tried to use that to his advantage. The rest of that diagram kinda draws itself. It’s not foolproof and it’s a diffuse, rather than direct, effect, but that’s how groups work.
This is not a direct test of whether Molina’s work ethic really does rub off on the Cardinals and yes, we will have a hard time putting a run value on it, but it’s a very reasonable hypothesis. Given our current data limitations, indirect evidence is probably the best evidence that we’re going to get. It doesn’t mean that other teams don’t also have these opinion leaders. In fact, I’d wager that most teams at least try this strategy, whether formally or informally. It’s not the one and only reason that the Cardinals beat the Pirates in the NLDS (Adam Wainwright’s curveball may have had something to do with it.) The Pirates might have had a similar leader. Maybe most teams do and having that leader only allows a team to keep pace with all the others. That doesn’t make it unimportant. And given that this sort of strategy has almost no cost and that any effect is likely to be positive, teams might as well try it.
If, as sabermetricians, we are attempting to apply rigorous scientific method to the study of baseball, I suppose that instead of ignoring chemistry, we can still do acceptable science through analogous reasoning. There are problems with that (not all analogies work!) and we do need to keep a healthy skepticism, but I’d argue that it’s time that we stop ignoring the “soft factors” like culture and chemistry and start doing the best we can with what’s available. And talk about it once in a while. If there’s one thing that sabermetrics is guilty of, it’s idolizing the third decimal place. Instead, let’s get more comfortable starting off sentences with “I can’t be totally sure about this, but it’s reasonable to think that…”