Yesterday in Fish Bites, I posted a link to an interesting quickie study using Baseball Musing’s Lineup Analysis tool of the potential for one of Tony LaRussa’s famously quirky moves: batting the pitcher slot 8th instead of 9th. Here’s a quick look at the results I calculated on the simulator myself.
The current lineup: 4.547 runs/game
The sensible lineup, flip-flopping Bonifacio and Baker: 4.646 runs/game
The LaRussa lineup, flip-flopping Bonifacio and the pitcher: 4.709 runs/game
The difference between the current lineup and the sensible one is worth a modest 16 runs over 162 games, but the difference between the current and LaRussa lineups was calculated at a 26 runs in 162 games.
As you can see, the difference between the current lineup and the LaRussa lineup is somewhat significant. 2.6 wins is nothing to scoff at; it’s the approximate difference between an average player and Dan Uggla last year. But if you take a look at the difference between the sensible lineup and the LaRussa lineup, the run differential is less pronounced, approximately one win less.
Both of these results do come with signifcant caveats, namely that the statistics listed here are for about a third of a season’s plate appearances and by no means stable yet. Sure, one and two wins sound like they matter and they do, given how we evaluate players with the latest run estimators and the closeness of playoff races these days; ask the Mets if they would’ve liked a one- or two-win improvement over Luis Castillo these last few years. But with only a a third of a season’s worth of data and plate appearances, the inserted OBP and SLG values could be of little meaning. Rob Neyer pointed out an Alan Schwarz article in the New York Times earlier in the season about simulators and their ability to project baseball data; you can find Neyer’s post here. Here the study mentioned was a difference between batting Alex Rodriguez cleanup vs. 9th in the Yankees lineup. Over 100 simulated seasons, the average differential was about 40 runs, or four wins. In comparison, our total projects a 1.6 win difference between batting the team’s worst hitter (the pitcher) 8th or 9th in the lineup based on a third of a season’s data. I find that difficult to believe.
Still, the results are significant in telling us something, and that something is not so much that the configuration of a pitcher batting 8th is impressive, but rather that the configuration of batting a player with a .291 OBP second is not. Plate appearances dominate the difference in run potential for lineups, because the more chances efficient and better players get at producing runs, the better those runs get produced. Bonifacio and his .273 wOBA are just not an efficient use of the second most amount of plate appearances for your team. Putting a much more reasonable player available to you at that slot and moving Bonifacio down in the order will provide a similar return.
Finally, this sort of batting order decision clearly takes a backseat to better player evaluation. Using our nifty tool, consider replacing Bonifacio in his current spot in the lineup with a league average player in the National League, sporting a .332 OBP and .404 SLG: 4.778 runs per game, a full 0.231 runs, or 37.4 more runs over the course of the full season. This is probably a bit of an overstatement as well, as Bonifacio was worth about -12.5 runs compared to the average over the first 286 plate appearances, and I can’t imagine him getting anything more than 750 PA’s over the full season batting #2 in the lineup. At 750 PA’s and his current rate of hitting ineptitude, Bonifacio would be worth about 32 runs less than an average player with the bat. Even then, those values trump any lineup changes being made, proving that the best way to fix an ailing offense is usually to replace bad players with decent ones, not shuttling them around in the lineup.
As a final look at the study, and a way to make myself feel smarter, here’s the lineup I designed in my head prior to using the tool, based on what I feel would be the best use of the team’s limited OBP skills.
1) LF Coghlan
2) RF Hermida
3) SS Ramirez
4) 1B Cantu
5) 2B Uggla
6) CF Ross
7) C Baker
8) 3B Bonifacio
This lineup calculated to 4.690 runs per game. Not half bad. I can’t wait to see when Gaby Sanchez comes up and fills in for Bonifacio’s spot. We’ll return and reexamine this later, with better sample sizes for players like Coghlan who currently haven’t played enough yet.