C’mon, Amir. As you know, Pass Labs amps have been very well received by audiophiles over decades now. They clearly have sound signatures that are pleasing to the ears of many listeners, and the suggestion that a meaningful percentage of those reactions would likely change if only those listeners were to A/B their amps with those that measure with less distortion is dubious, at best. And the same could be said of high-quality tube amplifiers.
To be clear, I don’t doubt that some listeners would arrive at conclusions that would be at odds with their long-standing, stated preferences. But given the vast weight of the feedback from audiophiles who apparently prefer amps which measure with some distortion in the audible frequencies, it is, in my view, highly improbable that their choices are primarily due to marketing-related biases.
I think ASR is—to a degree—caught in its own hermetic logic here.
One the one hand, Amir does ignore Toole’s advice that speakers should be listened to comparatively for evaluation to be meaningful. Even something as simple as setting up a curtain and turntable and enlisting helpers from his cohort of followers is dismissed. Instead he does measurements first, preps some PEQ filters based on them and on his experience, then compares "blind" original and EQ’d playback through the device under test (just one speaker). Nothing wrong with most of this (except, I would listen and take notes first—as another Kippel user, Erin of Erin’s Audio Corner does—that argument has run a few times on ASR). But I can AB between EQ and original and hear differences (most can) and that doesn’t substitute for comparative testing of (as Toole recommends) three or four speakers. Apart from the magnitude argument, it’s unclear how one type of listening is kosher and another isn’t, but that’s a longer discussion and requires more than ventilation of talking points.
But when it comes to amplifiers, ASR rarely listens, with the general justification that magnitude of differences are too small to differentiate (with some exceptions). As you can see in this thread all listening (that isn’t done by ASR with partial protocols) is routinely dismissed as sighted bias, expectation bias, focus bias etc (you too can play bias Whac-A-Mole). This despite the fact that double-blind ABX isn’t logistically straightforward. And dismissing well-defined subjective listening protocols like those published by B&K or B&O, or the expert listening we use daily to produce music. The escape clause is of course the null test, but that's also a longer conversation.
In cases where people do make a good-faith attempt to try DBT, it can also degenerate into a shambolic series of gotchas. Before my time at ASR I nonetheless read a long and winding thread where enthusiast/reviewer GoldenSound tried to respond to a challenge Amir made to compare DACs. Amir managed to misunderstand straightforward logical concepts like one vs many wrt comparison criteria, using that to reset his conditions halfway and eventually doxxed and banned his antagonist, and possibly some other posters in that conversation. I modified my appreciation of ASR ethics based on that thread. We are all flawed human beings though, so I enjoy ASR for what it offers and tolerate the weaknesses.