Oakland’s Platoon Advantage

Remember when the Texas Rangers were running away with everything and were on course to be the the Greatest Team in the History of Whatever? Well, this afternoon the As might just take the 2012 American League West title from them. Certainly the biggest factor in the Oakland’s success this year was the incredibly boring Moneyball movie (if only they had used the alternate script), but there are many other storylines to follow, from injured pitchers to rookie pitchers to Yoenis Cespedes‘ video to San Jose. Bu for all the wondering about what New Inefficiency the As may or may not have found, one thing worth investigating is how the As have maximized their runs scoring by turning back the clock to a strategy more common fifty years ago than it is now: platoons. How much of a difference has successful platooning made for Oakland?

The As have employed a number of platoons during the season. The the sake of simplicity and brevity, I want to focus on two of them: Brandon Moss and Chris Carter at (mostly) first base, and the Seth Smith and Jonny Gomes at left field and designated hitter.

It should be said that none of those players has been strictly platooned, and there are times when both parties in the platoon have played at the same time. There are defensive differences between the players. Moreover, the number of plate appearances for everyone on the team will change if we project worse offensive performance, which will effect contextual run-scoring. With those qualifications noted, I think simply going with a simple linear-weights-based approach (as implemented via wOBA) should be close enough for our purposes.

The temptation is simply to plug in the players we pick as the hypothetical full-timers numbers into the vacated plate appearances. However, it is not that simple — there is a difference between observed performance and true talent. We can’t simply plug in, say, Brandon Moss’s numbers against southpaws this season and assume that is what he would have done if he had gotten Chris Carter’s plate appearances. Moreover, just as estimates of true talent (i.e., projections) need to be used in an estimate like this for overall performance, we also have to account for how each individual player’s platoon split needs to be estimated given his past performance. For the overall projections of true talent, I will use Oliver’s updated forecasts for 2012 true talent wOBA, and will then apply this basic method for estimating hitter platoon skill.

Rather than going through all the possible combinations, let’s hypothetically assume that Oakland instead gives the full-time jobs to the two players that Oliver sees as having the best overall offensive talent for each slot: Moss (.337 projected wOBA versus Carter’s .323) and Smith (.334 just edges out Gomes’ .333). You are probably noticing just how much all of those players have outperformed even these updated projections, something I will briefly touch on at the end of the post.

As of this writing, Chris Carter has received 109 plate appearances versus left-handed pitchers this season, and has had a .384 wOBA against them. Moss has actually been okay versus left-handed pitchers this season with a .332 wOBA, but in just 60 plate appearances. Using the methods mentioned above, Moss’s true talent wOBA versus left-handed pitching is a .314 wOBA. If he had received Carter’s 109 plate appearances versus lefties, the difference compared to Carter’s actual performance would have been about six runs according to linear weights.

Neither Carter nor Moss’s observed or projected splits really compare to those of Smith and Gomes, however. Jonny Gomes got 196 plate appearances versus left-handed pitching this season, and destroyed them for a .416 wOBA. Seth Smith has always struggled versus lefties (.257 career wOBA against), and even after regression, his true talent wOBA versus southpaw pitchers projects at .293. Over 195 plate appearances, a .293 wOBA would be expected to be about 19 runs (!) worse than Gomes’ .416.

Putting both (un)platoons together, we estimate that Oakland would have scored about 25 runs less if they had not used Gomes and Carter versus left-handed pitching in the way they did. In this run environment, 25 runs are closer to being worth three wins than two. I don’t know, do you think that would have made a difference in this divisional race?

As I mentioned above, we should acknowledge that, even being platooned, Smith aside, Carter, Moss, and Gomes all hit better than most would have projected them to hit this year. But that is the case with just about every team that outperforms expectations. Pretty much every team has some players that outperform expectations, and some that under-perform them. Teams that win usually have done better on the positive side of the ledger. Yes, these platoons have worked better than one might have expected. Still, even if we just compare how the platoons would be projected to do in the playing time given rather than how the projected non-platoons did versus how things went, I still have the As gaining about eleven runs from them — just over a win. At the moment, yeah, I think that would have mattered.

The As may not have planned it just this way, and they clearly had some good fortune this season. But doing things like taking four part-time players and getting and additional one to three wins out of them is a smart way of maximizing returns. What exactly is the residue of design, again?

My thanks to David Wiers for his assistance with certain Oakland-ey details.