Coming off his remarkable 2001 season, Barry Bonds had attained a reputation for dominating right-handed pitching while being a mere mortal against lefties. Looking at wOBA numbers from 2000 and 2001, he averaged .538 against RHP and .445 against LHP. (We should note that wOBA isn't as useful when evaluating Bonds, given that he receives a large number of semi-intentional walks that get scored as unintentional.) Compared with the average platoon split of 25 points of wOBA for left-handed hitters, Bonds seemed unusually susceptible to left-handed pitching. National League rivals responded the obvious way—by loading up on lefties, especially those with previous success against Bonds.

The result? Predictably, the fraction of Bonds' plate appearances against left-handed pitchers increased. More importantly, how did this strategy work? Unfortunately for those rivals, not very well. In 2002 and 2003, Bonds averaged .579 against lefties and .523 against righties.

Interestingly, from 2000 through 2003, Bonds' wOBA vs. right-handed pitching was .531, compared with .508 against left-handed pitching—a platoon split virtually identical to the average for lefties. So which is he: the righty masher from 2000 and 2001, the lefty masher from 2002 and 2003, or a typical left-handed batter? Let's expand this question to batters as a whole. Do all left-handed (or right-handed) batters share the same platoon splits, or is there some spread in “platoon talent?” If the latter, how can a manager estimate a batter's “true” platoon split in order to take advantage of it? And of course, what about pitchers?

As usual, our primary concern is the effect of randomness when dealing with small data samples. Left-handed batters face left-handed pitchers in about one-quarter of their plate appearances, meaning that for a regular starter, we're dealing with around 150 plate appearances per season. That may seem like a lot, yet in 150 PA, a typical player's wOBA will be pushed up or down by 40 points due to randomness alone. As usual, we'll deal with this by looking for predictive power—whether or not an extreme platoon split is a sign of a non-average platoon split in the future. We'll start by examining left-handed hitters, and looking for seasons with extreme platoon splits. Of players with at least 400 plate appearances against right-handed pitching and 100 against left-handed pitching in consecutive seasons, here are the ten lefties who fared the best against southpaw opponents, and how they fared the following season.