Yes, that’s part of what’s being talked about, in connection with various ideas being batted around that may explain some of the specifics of the BtB rankings that ring less than true.

My apologies for jumping into the middle of the discussion you’re having with hairball here with vague references to stuff that’s simultaneously being discussed elsewhere.

The same translation is applied to all estimated “runs”, regardless of the actual disparity between leagues in the area those runs measure: offense, pitching, or defense. If the disparity in talent between leagues isn’t roughly equivalent in each area, the uniform translation results in distortion of “true value”. The above posits that the relative ranking of the Royals and Giants would have been affected by such a distortion.

Irrelevant to strength of schedule issues. Not irrelevent to an assessment of the ranking system.

To translate performance across leagues, the system applies a correction for an effect measured in the aggregate to specific entities and not to the aggregate. The results of this kind of application are bound to lead to silliness in individual cases. Example: the Giants’ ballpark suppresses home runs by lefties in the aggregate by x%, so perhaps if I apply a correction to Bonds’ home run numbers at home I should expect to get a truer picture of his performance. But that’s not what such a translation yields: in fact there’s no evidence that the park affected Bonds specifically to anywhere near the same degree as the league overall. The translated numbers for Bonds are bunk.

So, my question is: why should I assume that applying a uniform translation to all teams, regardless of construction, when normalizing performance across leagues? How do I know that some types of performances will be suppressed or amplified differently from others?

I believe a team’s areas of strength relative to the league’s areas of disparity would matter. I doubt that there are a sufficient number of interleague games to measure any subtleties of effect, however, so I’m not expecting any light to be shed on the question.

Anyway, a lot of words on what’s essentially a trifle.

I could be convinced of that. Those 2006 studies showed that the disparity changed dramatically between the early part of the decade and, say, 2005, so it would not be inconsistent with those findings for it to have changed again — not just in magnitude but in composition as well — since then.