PM: Thanks for the clarification, but I understood, I just didn't agree. It's your proposed effect, as well as your analogy about mechanism, that seems out of whack with reality to me. Assuming you're seriously positing this hypothesis as a reason why A/B testing allegedly doesn't work or is misleading, than both cause and effect are fair game for critical examination. If you think I'm taking it too far, it's just that I see the possible extrapolation here -- that your proposed effect implies instantaneous comparison is less reliable than audio memory, and I don't buy that.
For the record, here's how I see this A/B vs. long-term question in its totality: A/B's I think are great for identifying differences, and degrees of change. If performed against a well-known reference they can be a good indicator of relative strengths and flaws (or in the case of bypass testing, absolute strengths and flaws). But that's not always the same as determining which presentation you prefer, and it's never the same as determining whether that preference will ultimately meet your listening needs and expectations. Long-term auditioning I think is necessary (and anyway unavoidable, let's not forget) for determining preferences and ultimate satisfaction. I agree that quick A/B's often don't reveal nearly as much as there is to hear. The solution in my experience is not to throw away A/B's altogether, it's to keep doing them until the finer differences emerge, which they do if you have determination and patience. Once heard, as I said, this method most clearly eludicates differences and degrees of change, and more reliably so than depending on long-term auditioning and audio memory. In practice I prefer to use both methods for their own virtues and not just rely on the latter.
For the record, here's how I see this A/B vs. long-term question in its totality: A/B's I think are great for identifying differences, and degrees of change. If performed against a well-known reference they can be a good indicator of relative strengths and flaws (or in the case of bypass testing, absolute strengths and flaws). But that's not always the same as determining which presentation you prefer, and it's never the same as determining whether that preference will ultimately meet your listening needs and expectations. Long-term auditioning I think is necessary (and anyway unavoidable, let's not forget) for determining preferences and ultimate satisfaction. I agree that quick A/B's often don't reveal nearly as much as there is to hear. The solution in my experience is not to throw away A/B's altogether, it's to keep doing them until the finer differences emerge, which they do if you have determination and patience. Once heard, as I said, this method most clearly eludicates differences and degrees of change, and more reliably so than depending on long-term auditioning and audio memory. In practice I prefer to use both methods for their own virtues and not just rely on the latter.