This came up in the game thread yesterday, and unfortunately I missed some of the more interesting statistical comments before I had to leave.
The subject was Brandon Inge, and his fantastic 48 PAs against Cliff Lee having a predictive value going forward. I knew I had encountered a study before, but could only find a pretty crappy example at the time. With a bit more research, I was able to find the original source: Tom Tango's The Book. The whole book is available online, and I've linked to some relevant bits available here. The chapter is a quick read, so I'd recommend just scrolling there if you're interested in the study and methodology. The conclusion?
The Book Says:
Knowing a player will face a particular opponent, and given the choice between that player's 1,500 PA over the past three years against the rest of the league, or twenty-five PA against that particular opponent, look at the 1,500 PA.