Add new comment

You can not say that something is a "significant sample" if you don't know what the spread in the population is in regards to that skill. Some stats require a relatively small sample to be meaningful because there is large spread in the measured skill, like groundball-rates. Other stats require a massive sample (almost a whole career) to tease anything out because there is almost no difference between players, like with clutchness. In the case of right-hander platoon splits, it is much closer to the latter. Read The Book to understand what I am talking about, that's where I got the reqression number. For lefty batters the regression is about half that of righties (meaning their is a larger variation among them) and for switch-hitters it's even less, with about 600 PA's needed to regress half-way (which makes sense because they are almost like two different hitters).
The reality is that there is very little variation amongst right-handed hitters (at least as far as major leaguers are concerned) when it comes to their platoon split. So if you see a guy like Guti who has hit lefties extremely well but poorly against righties, then he is almost certainly a better batter against righties than his numbers indicate (and worse against lefties than his numbers suggest). Therefore, platooning him with another right-hander like Wells wouldn't do you any good, especially since Franklin is a better defender.