July 16, 2012
Resident Fantasy Genius
To Platoon or Not to Platoon?
On Thursday, reader “jimcal” asked me in the comments section of my article to give my thoughts on platooning players in fantasy baseball. While platooning is a bit of a complicated subject, I’ll do my best to tackle it all in one article today. When considering platooning, there are two main concepts that the discussion can be distilled down to: sample size and opportunity cost.
What most people don’t realize is that very few players truly need to be platooned. We tend to look at a player’s performance versus same-handed pitching either for the current year or even over a three-year period when making such decisions, but this isn’t nearly enough data to make a reasonable guess as to whether the player is best used in a platoon (absent scouting data that supports his performance, which makes this a more complicated decision).
In The Book: Playing the Percentages in Baseball, Tom Tango, Mitchel Lichtman, and Andy Dolphin found that “a righty who has 2,200 plate appearances against left-handed pitchers will be regressed exactly halfway toward the league average... For a lefty, the number is about 1,000.” This number is the same as the “stabilization point” that many tend to associate with the usual array of stats that we use to analyze players (i.e. strikeouts “stabilize” more quickly than BABIP). (I do need to note that I think the term “stabilization” is misleading, but it’s the word most often associated with the concept. I’ve written in length on the subject here and here, and Russell Carleton tackled it again today.)
What these numbers tell us is that, if we were trying to estimate a righty’s platoon skill, it would take roughly 10 seasons just to account for half of the variation in the skill. To put that in perspective, I found that a pitcher’s BABIP—a stat that most would never take at face value, realizing that it is riddled with “luck”—takes roughly eight seasons to reach the same level of “stabilization.” Yes, you read that correctly: a righty’s platoon skill is more unstable than a pitcher’s BABIP! In other words, we would be better off trusting a pitcher’s single-year BABIP than we would a hitter’s single-year platoon data.
A great example of this is Curtis Granderson, whom many fantasy owners feared would struggle going into the 2011 season if he was overexposed to lefties. From 2005 to 2010, Granderson posted a terrible .621 OPS against lefties in 858 plate appearances. Now that he’s posted rates of .944 and .856 in 2011 and 2012, respectively, the media and fantasy players alike are clamoring to declare that he’s solved his “issues.” The problem with this analysis is that Granderson may not have truly had any issues to begin with. With fewer than 900 plate appearances against lefties in his career prior to 2010, we hadn’t even reached the point at which we his past performance was a stronger indicator of future performance than mere league average. Yes, he did work on his swing with Yankees hitting coach Kevin Long, which has been credited with much of his success over the past year and a half (not just versus lefties but overall), but the point still stands: don’t read too much into platoon splits absent a good reason.
Once this issue is behind us, we deal with one that’s more specific to the fantasy game: opportunity cost. Once we find players who are worthy of being platooned, either because of our own analysis or because the player’s manager plays him only against certain pitchers, we need to consider the implications of using up a bench spot to milk some extra value out of one active spot. Opportunity cost, in this case, boils down to asking ourselves the question, “If I don’t use this bench spot on the second half of this platoon, what will I use it on?” This is a question I can’t really answer for you; it’s going to depend heavily on a large number of league- and team-specific factors. How many bench spots do you have? How valuable are those bench spots? How deep is the league? What options are available on waivers that could fill the spot instead? How deep into the season is it? Would you be better served stashing a top prospect who might be recalled in a few weeks? Would you be better off stashing hot-hitting Player X and gambling that his breakout is for real? How much more valuable does my active roster spot become by using this specific platoon arrangement?