His reasoning doesn't matter. Pitcher vs hitter matchups are an extremely small sample with almost no predictive ability. In the Book they took cases where a batter had at least 27 before PA and at least 9 after PA. Of the 30 matchups where a hitter dominated the pitcher, the after PA were roughly split into 10 hitter continued to dominate, 7 hitter matched career stats, 13 hitter was dominated by the pitcher. On the flip side, They looked at the 30 matchups most dominated by the pitcher. The after came out as 8 mauthups dominated by the pitcher, 10 dominated by the batter, and 12 neutral. I am not sure if you could come up with a more random after/before spread. Even if you have 2 guys playing in the same division, they are not likely to face each other more than 24 times in any given year.(18 head to head games, guessing 6 series of 3 games, starting pitcher starts against the opposit team every series, 4 at bats per game, all guesses on the high side as it is unlikely a starting pitcher starts in every head to head series) Assuming that Fangraphs has the point where stats stabalize correct, Link 50 PA: Swing % 100 PA: Contact Rate 150 PA: Strikeout Rate, Line Drive Rate, Pitches/PA 200 PA: Walk Rate, Groundball Rate, GB/FB 250 PA: Flyball Rate 300 PA: Home Run Rate, HR/FB 500 PA: OBP, SLG, OPS, 1B Rate, Popup Rate 550 PA: ISO [/pre]We need over 2 years worth of data before anything becomes reliable. By the time we would get enough head to head matchups to mean anything, the hitter and batter are so far removed from many of the PA for the data to mean anything. What a hitter did 4+ years ago isn't likely relavent to what he is as a hitter now.