Although I have not done a general survey of reactions to Joe Girardi‘s decision to have Robinson Cano hit second in eight of the Yankees’ first 11 games, I can imagine many saber-friendly fans are excited to see the player who is mostly likely the Yankees’ best hitter in the second spot. Ever since the findings The Book’s chapter on lineup optimization became popular among baseball bloggers, complaints about managers “wasting” the second spot in the lineup (where the best hitter, or at least one of the best three, according to The Book) have increased.
I do not know whether Joe Girardi is putting Cano second because of sabermetric insights or simply because with Derek Jeter and Mark Teixeira out, he is at a loss regarding how to split up the lefties in his lineup otherwise. The latter suggestion is what Wallace Matthews believes. Matthews is not a fan of Cano hitting second. After all, a hitter with Cano’s average and power simply can’t hit second, right? Matthews:
Robinson Cano is not a No. 2 hitter, not in any way, shape or form, and not on anyone’s lineup card in baseball. Except, of course, for Joe Girardi’s…. This adjustment, however, does not add up. In addition to all the home runs the Yankees lost to free agency, the injuries to Curtis Granderson and Mark Teixeira took two bats that accounted for 67 home runs last season out of the lineup. So instead of installing Cano, who hit 33 HRs last year, in the middle of the lineup where he belongs, Girardi moves him up.
My intention is not to pick on Matthews, who makes some other interesting batting order suggestions and notes that Girardi ends up batting two lefties together, anyway, with Cano coming right behind Brett Gardner at the top of the order. Matthews’ comments quoted simply serve to illustrate a certain mindset about what sort of hitters should or should not hit second. I will take it for granted that it is just fine to have the team’s best hitter hit second. What I want to examine is whether, in this specific case, it makes sense if the Yankees are going to hit Cano right behind another left-handed hitter.
Before getting to the Yankees and Cano, I probably need to briefly justify a post about batting order. After all, batting order does not matter all that much. Studies have found that the difference between a typical and optimized batting order is worth between just five and 15 runs a year. Moreover, one really needs to do something with models and simulations to get the best batting order, and I am not going to get into that specifically here.
Still, I have toyed with the idea of something like a “Number Two Watch” for a while as an occasional series, and I want to give it a try here with a situation that interests me for a particular reason. As for the issue of how much batting order batters, I have two responses: 1) Batting order matters about as much as other stuff we worry about in baseball. 2) If we are only supposed to write about important things, why would we blog about baseball at all? On to the issue at hand
It is hard to say how long the Gardner-Cano one-two punch at the top of the Yankees’ batting order will last. The Yankees utilized it from their third game up until yesterday’s game, when Vernon Wells (~!) hit second and Cano moved down to third. I assume this was at least in part due to the Orioles sending left-handed starter to the mound, so the right-handed Wells went into the game. But even if it ends the Cano hitting second experiment, there is still something worth talking about.
One difficulty when talking about even just one batting order spot is that one cannot consider just one hitter and his traits. The notion of a player being “ideal” for a particular batting order slot tends to cover up the fact that where a hitter is best slotted depends on the specific skills of the other hitters in the lineup. Putting all the likely combinations of base/out situations for various possible batting order combinations is not something one humans can realistically do, which is why simulations and models are important. This is all a roundabout way explaining why I am going to limit the discussion to one specific point: granted that the Yankees have decided to bat the left-handed hitting Gardner first, is it still okay to have another lefty, Cano, second?
Matthews himself notes that Girardi ended up with Cano and Gardner hitting next to each other because he was trying to split up his lefties, but in particular combinations of his currently available hitters, it was impossible to do completely. So Girardi opted to put two lefties up top. The problem with putting two lefties together, even against a right-handed starter, is that in late, high-leverage situations, the opponent might be able to better leverage a left-handed reliever against the team.
That is a concern, and, again, what we really need is a very sophisticated model or simulation (better than any publicly available ones that I know about) to help us work out all the possibilities. We can still think things through on an abstract level, though.
As noted above, for the sake of simplicity let’s just assume Gardner is at leadoff at least for now. Gardner himself has a pretty small observed platoon split, but using his current ZiPS Rest-of-Season projections and regressing his platoon numbers, he projects as a .330 wOBA hitter versus righties and a .305 wOBA hitter versus lefties.
As for Cano, he had a mammoth platoon split last year, sporting an incredible .461 wOBA versus righties and a backup cathcer-esque .290 wOBA versus southpaws. Prior to 2012, though, he actually had a smaller-than-average platoon split for a left-handed hitter. Cano’s current ZiPS projection is for a .377 wOBA. Applying the same method used above for estimating his platoon skills and applying it to that projection, I have him as a great .394 wOBA hitter versus righties, and a still-good .355 wOBA hitter versus lefties.
Obviously, with these two batting next to each other, an opposing manager is going to try to leverage his left-handed specialist against them in crucial situations down the stretch. But keep in mind that with their current lineup against right-handed starters, the Yankees are going to have to have two lefties batting next to each other somewhere. And Cano’s .355 expected wOBA versus lefties is still very good — even if Gardner projects to be pretty bad against lefties, Cano is not, even if he is not nearly as good against righties.
To repeat perhaps one too many times — one really needs to go through all the options for the batting order, and I cannot do that here. One could, for example, suggest that Kevin Youkilis, who projects as a .385 wOBA hitter versus lefties given the methods above, should hit second. Maybe, but note two things. First, as Matthews noticed, the Yankees would still have to have two lefties somewhere in the lineup, and it is impossible to plan for which set of lefties will come up in a crucial situation late in a game.
Of more interest to me is a second point: while some game situations are more crucial, this sort of thinking emphasizes those over the majority of the plate appearances in which the Gardner/Cano pair would likely face against the starter. One might argue that Youkilis or another righty should hit second versus a lefty starter. Leaving that aside, versus a right-handed starter, Cano is much more effective. In those games the opposition will have only one, maybe two chance to bring in a lefty specialist against them. And not all of those games will have crucial situations — some of them will be blowouts one way or the other. It might happen, it might not. However, one can be pretty sure they are going to get at least three shots against the righty starter, unless he gets knocked out of the game, in which case a crucial situation is pretty unlikely.
In other words, why emphasize the chance of a one high leverage situation over the relative likelihood of several series of plate appearances versus a right-handed starter that both hitters would face? My point is not that I know for sure. (I admit to at least partly violating my own warnings by not simulating or modelling the probabilities myself.) It may indeed be better to hit Youkilis second. But, assuming Gardner is going to lead off, it is not obvious to me that he should. Like other general batting order “rules,” splitting up lefties is a good idea, but like the rest of those rules, it is not necessarily a hard and fast one. As I have wondered elsewhere, in particular cases, it might be the case where the (very slightly) increased chance of scoring more runs by being willing to stack hitters of the same handedness versus a starter might outweigh the benefits of betting on facing a certain kind of relief specialist.
Per game, and even per season, we are talking about just a few runs. But when we talk about “little things” making a difference, this is one I think it at least worth considering.