Yesterday, The New York Times had an article by Alan Schwarz that quantified the the average improvement AL starting pitchers experienced when they moved to the NL and the average deprovement NL pitchers experienced when they moved to the AL. From 2000 through 2005, there were 29 AL emigrants and 28 NL ones:
Combined E.R.A.s for the new National Leaguers decreased to 3.94 from 4.79, or 0.85 of a run, while their counterparts' increased to 4.64 from 3.94, a move of 0.70.
Mr Schwarz also applies ERA+ to make the case even stronger. (Afterall, there is a certain cache to making cases more complicated by appling even less intuitive statistics!) Using this stat:
Of the 29 pitchers moving to the N.L. from the A.L., their E.R.A.+ figures increased to 110 (10 percent above league average) from 97 (just below average). [average=100]...Pitchers found moving to the A.L. from the N.L. correspondingly unpleasant -- the E.R.A.+ scores of the 28 pitchers decreased to 100 from 113...
Whatever method fantasy leaguers use, there is statistical support for the common sense adage that AL pitchers benefit from changing leagues while NL ones who change leagues do not. I do wonder what the pre-PED/post-PED splits are, though.