I think the best argument for the validity of ABX is a test done some 20+ years ago (I remember the article but unfortunately don't have the reference).

A bunch of golden ears participated in an amplifier comparison and, no surprise, null results. They couldn't distinguish their ass from their elbows, with one exception.

The exception was J Gordon Holt - he hit like 19 out of 20 or something reasonably similar. That told me something - that a guy who was never accused of being full of shit as far as I know showed he could walk the walk, unlike his contemporaries who just talk the talk.

Now he's a crafty old bastard, and he may have found a reflection in a window someplace that allowed him to see the signal input lights on the amp front plates or something like that. But assuming it was on the up-and-up I would claim that it's a clear demonstration that ABX does not obsecure the details as claimed by its detractors - you just have to be as good as you claim to be when somebody is watching (as opposed to all the crazy claims they make in reviews of stuff they hear when nobody is watching).