clock menu more-arrow no yes mobile

Filed under:

The Yankees have a nearly optimal Opening Day lineup

They could make a few changes, but this lineup is about as good as it gets.

MLB: Spring Training-Boston Red Sox at New York Yankees Butch Dill-USA TODAY Sports

Modern sabermetrics has been around for over a decade now, and one the earliest advances of the movement was understanding how lineups worked, and how to make them better. While traditional orthodoxy had your speedy player first, best hitter third, and slugger fourth, the actual numbers bear different results.

In short, a manager wants their best hitters to be first, second, and fourth in the lineup, namely with the goal of giving your best hitters the most plate appearances. The biggest misconception is that speed above all is most important at the first spot; in fact, it is on-base percentage. It remains correct that power is most important at the fourth spot, and then your players should be ranked by wOBA for spots two, three, five, and beyond.

What this yields for an optimal Yankees lineup is pretty surprising. To figure this out, I plugged in the Yankees’ Opening Day lineup into’s lineup markov tool, which strings together runs per game based on a player’s wOBA and slugging percentage; for this exercise, I used their projected numbers from Steamer.

The Opening Day lineup is projected to score 4.749 runs per game, which we know isn’t totally accurate, but it gives us an idea of what their talent is like. It obviously ignores speed, opposing pitcher match ups, park, and other external factors, but we don’t have access to that aggregate information, so this is the best we can use to get a rule-of-thumb analysis.

The good news is that this lineup is nearly optimal. Based on lineups ranging from 4.584 runs per game and 4.78 runs per game, this lineup sits at 4.749. That means that the difference between the current lineup and the ideal lineup replicated over 162 games only has a difference of about five runs. That’s basically rounding error.

But what does look interesting, something that I didn’t even think of, was what this optimal lineup actually looks like. What I mean is that Matt Holliday is hitting first. If we follow the rule that on-base percentage is the most important stat for the first slot, then Holliday might be best suited. It’s radical, and it flies in the face of even current orthodoxy, but you can see how breaking that orthodoxy might net the Yankees a few spare runs. That’s pretty important when you’re sitting on the edge of contention.

Unless, of course, Greg Bird or Gary Sanchez are on base more. Based on how this team is “supposed” to act, and there’s a lot of variance this year in particular, Holliday should be hitting first in the lineup, as the player most likely to get on base. I’m not going to fault Joe Girardi for not doing this, because it likely draws more scrutiny than he would like, but it’s something to think about.

The other piece is that, for the most part, the rest is constructed correctly. Aaron Judge could be a little higher, but we still don’t know what his true capabilities are. Brett Gardner could be a little lower, but his speed and decent on-base abilities mean it’s basically a wash. And Jacoby Ellsbury, who should not have been hitting second to begin with, rightfully slides down, as the numbers says he should (they say sixth, and he’s fifth).

Of course this is all subject to change. Didi Gregorius is likely to return, and I’m sure there will be injuries, acquisitions, and constant changes, both in personnel and analysis of what a player’s true talent is. That’s even more so the case in a year where some players, while exciting, are relative unknowns at the major league level. For now, though, I commend Girardi for the lineup he’s put together, and while I could make my own adjustments, I’ll just be happy to watch baseball this afternoon. I doubt I’m going to miss the three-hundredth's of a run.