Tim,
Thanks for the update. Its not very often that someone complains and then makes a point later to come back and say that things aren't as bad as they use to be. It's easier to pull a Roseanna Danna and just say, "Never mind."
One point of my "Dozen Brian Giles" argument was to point out that as you aggregate over multiple small samples, the sample averages approach population values. At that point, averaging over a dozen 60-game samples of Brian Giles generated pretty close sample stats. You'd think that now averaging over 12 120-game samples, you get even a little closer. I could have picked someone else, say Bagwell, and not as done as well w/ the 60 game samples, but then he'd be much closer with the 120-game samples.
Some of the discussion after my post followed up on issues related to while single small samples don't replicate as well. Random fluctuation is a key factor, but pitching staffs faced is another. It's not uncommon for you to start the season facing 4-5 loaded rotations, and that would depress batting stats even at 60 games. Or, you faced a lot of lefties early on, and (as you see in my leagues) a lot fewer lefties over time. These things affect our players, and they affect real players too. However, we typically don't have sports writers smart enough to help us place fast/slow starts in context.