clock menu more-arrow no yes

Filed under:

Optimizing the 2018 batting order

New, 341 comments

Batting orders don’t matter that much but it’s cool when they’re good.

St Louis Cardinals v Pittsburgh Pirates

The 2018 Chicago Cubs are projected to field a lineup of haves and have-nots. The gap between their lineup’s projected best hitter per Steamer by wRC+ (Anthony Rizzo) and their projected worst hitter (Albert Almora) is 49 points. For the Milwaukee Brewers, the gap is 47 points, between Christian Yelich and Manny Pina. For the Cincinnati Reds, there is a 43 point gap between Joey Votto and their next-best projected starter (Eugenio Suarez; the gap between Votto and Billy Hamilton is 77 points).

For better or worse, the St. Louis Cardinals do not have such stark differences between their top hitters and their lesser starters. The lowest projected Cardinals starter by wRC+, Yadier Molina, would be the 4th best projected starter on the Reds—for being the worst of the lot, Molina is fairly competent. But while several Cardinals starters were pleasant surprises in 2017, Marcell Ozuna is the team’s top projected hitter with a relatively modest 129 wRC+.

There are pros and cons to having a balanced lineup rather than one with talent more widely distributed, and if you are somebody who does not put much faith in Mike Matheny’s tactical acumen, a major positive of balance is that there isn’t as much that the manager can do to impact the team one way or another. Aside from, say, batting the pitcher at the top of the lineup (say what you will about Mike Matheny as a manager, but he hasn’t done that), there isn’t that much difference between the optimal lineup and a poorly arranged one.

But an optimized lineup is still preferable to a non-optimized one. If you tell me an investment opportunity that is going to make me a millionaire overnight, I’m going to take it, but if you tell me an investment opportunity that’s going to earn me five dollars, well, I’m going to take that as well! It may not be life-changing but marginal improvement is still improvement. The odds that the Cardinals miss out on the playoffs by such a small margin that it can be chalked up to rearranging the starting lineup is low, but it’s still worth doing just in case.

The most basic principle which a team can practice in constructing a lineup is to put the best hitters in the highest leverage spots in the lineup. The three most important spots in terms of potential impact are 1st, 2nd, and 4th. The leadoff spot is critical for a reason which is so intuitive that it is truly amazing that batting top hitters leadoff hasn’t always been considered as fashionable as putting a speedster there—the leadoff hitter bats the most often. The #2 spot, once the home of bunters and relatively low power hitters, is a spot with nearly as many opportunities as leadoff but with an extra premium on extra-base power. And the #4 spot, while receiving fewer plate appearances than the first two spots (well, the first three spots), comes up in the highest leverage situations.

The three best hitters on the Cardinals by projected wRC+ by both Steamer and ZiPS are Marcell Ozuna, Matt Carpenter, and Tommy Pham (the two formulas agree with Pham at third, though Ozuna gets the Steamer edge while Carpenter gets the ZiPS one). Matt Carpenter fits the new prototype of a leadoff hitter (a prototype which, to Matheny’s credit, was less common when he became a starter in 2013 but, to Matheny’s blame, has become more common by 2018, a season in which Carpenter is expected to bat third).

Craig Edwards wrote on Tuesday about the perception that Carpenter can only bat first, and how said perception is nonsense. I agree, but in the case of the Cardinals, he remains the best option in the very important spot. The projection systems agree that of the three top hitters, Carpenter will have the lowest batting average and the highest on-base percentage thanks to his penchant for drawing walks. Drawing walks is always positive but it is particularly advantageous when other potent hitters are coming up to drive the runner home. The systems also agree that Marcell Ozuna will hit for more power and draw fewer walks than Pham (or Carpenter), making him particularly qualified to bat cleanup. And then Tommy Pham, the best hitter of the trio in 2017, bats second.

Of projected starters, the next two ranked by wRC+ are Dexter Fowler and Jedd Gyorko. The gap between the #3 and #5 spots is very slim, though there is slightly more of a relative advantage to home run hitters batting third (Sky Kalkman explains why here), and Jedd Gyorko has been the bigger home run threat of the two, while Dexter Fowler has been superior at drawing walks. So let’s pencil in Gyorko at three and Fowler at five. The next four spots in the lineup are, intuitively, where to put the rest of the hitters in descending order. The projections agree that it goes Kolten Wong, Paul DeJong, and Yadier Molina. There is an argument for batting the pitcher eighth, a practice Tony LaRussa famously implemented while in St. Louis, but that’s a separate discussion. If you want to put the pitcher eighth, just mentally configure listed #8 hitters as #9. Here’s the lineup we’ve figured through projections. Batting handedness is noted in parentheses as lefty, righty, or switch-hitter.

  1. Matt Carpenter (L)
  2. Tommy Pham (R)
  3. Jedd Gyorko (R)
  4. Marcell Ozuna (R)
  5. Dexter Fowler (S)
  6. Kolten Wong (L)
  7. Paul DeJong (R)
  8. Yadier Molina (R)

That was simple enough. But this also assumes that games are played in a vacuum against some neutral-armed pitcher. This is the optimal lineup if the same lineup were required to be used against right-handed pitchers or left-handed ones, but platoons are a thing. They have not typically existed on a day-in-day-out basis in the Mike Matheny era, so for now, let’s assume that isn’t the case, even though Jose Martinez was definitively better against left-handed pitching than Kolten Wong or Matt Carpenter, and putting Martinez in the lineup instead of one of these batters can be accommodated with some infield shuffling.

But platoon splits could impact the proper arrangement of these players. While the one through eight listed above is mostly in line with a solid platoon against righties (since pitchers are not evenly split between righties and lefties, ability against right-handed pitching is going to disproportionately affect projections), a case could made for flipping Fowler and Gyorko based on handedness—this lineup includes no more than two consecutive hitters with the same handedness.

But against lefties, the game changes a bit. In 2017, the worst hitter among projected 2018 starters against left-handed pitching was Matt Carpenter. The second-worst was Kolten Wong, the lineup’s other lefty bat. If we were to assume this reflects their true talent (we shouldn’t, but that’s beside the point), it stands to logic to put them lower in the batting order, but a left-handed starting pitcher does not mean an entire game of facing lefties. In the National League, in 2017, the average start went 5.52 innings, while the average bullpen went 3.38 innings per game. About 38% of innings were handled by relievers. So in looking at these eight players by 2017 production, how would the lineups compare if optimizing by:

  1. Against right-handed pitching 100% of the time
  2. Against right-handed pitching 62% of the time, using overall wRC+ 38% of the time
  3. Using overall wRC+ 100% of the time
  4. Against left-handed pitching 62% of the time, using overall wRC+ 38% of the time
  5. Against left-handed pitching 100% of the time

Here are the lineups, with some subjectivity based on the general rules assigned earlier:

1. Matt Carpenter, Tommy Pham, Paul DeJong, Marcell Ozuna, Dexter Fowler, Kolten Wong, Jedd Gyorko, Yadier Molina

2. See #1

3. Matt Carpenter, Tommy Pham, Paul DeJong, Marcell Ozuna, Dexter Fowler, Jedd Gyorko, Kolten Wong, Yadier Molina

4. Tommy Pham, Jedd Gyorko, Marcell Ozuna, Paul DeJong, Dexter Fowler, Yadier Molina, Matt Carpenter, Kolten Wong

5. Tommy Pham, Jedd Gyorko, Marcell Ozuna, Paul DeJong, Yadier Molina, Dexter Fowler, Kolten Wong, Matt Carpenter

Even aside from the most obvious flaw in this experiment—that 2017 results are not 100% predictive of 2018 results—the purpose of these calculations is mostly to show how dramatically different lineups can be when optimizing for handedness. The difference between lineups two and four is fairly dramatic, and I would argue that, at least philosophically, these are the correct approaches to assembling a lineup. As great as Matt Carpenter is against righties, perhaps his relative mediocrity should mean he winds up in less important spots against lefties.

Or perhaps we should temper our expectations and just hope Yadier Molina starts more than thirteen games batting sixth or lower in the lineup.