Is it all in my head??

I have this theory on "cooking" cables too... I think what is happening might be a combination of oxidation and 'listening experience'.  I have a pair of speaker cables I made that are braided/twisted and when terminated are 9AWG equivalent.  I did an internal bi-wire pair so 6AWG essentially.  When I hooked them up, I was shocked at how bad they sounded compared to the non bi-wire cables i replaced (both made from the same 12-awg stock)... but after listening to them for a week, they started to sound better and more detailed (because 8 wires to each speaker are way better than 4 right?!).  Did they sound better? probably not and i kinda' think they still sound worse now that i think about it.  maybe that too is all in my head...

@kumakahn Exactly...unfortunately for us all, there Are no sources of sound, scientific information that relate to sound characteristics like the ones you mention. That sort of thing, AFAIK, is not to be found in science, per se, but only in marketing...hence we rank ametures are having to try to do it all ourselves, basically.

In the early 2000’s I got into it because of the raging debate at the time of which was better, stranded or solid core...I mean about half the manufacturers were saying that stranded was clearly the "greatest" and the other half said solid core - the only thing I knew for sure is that they couldn’t both be right. That’s when I started experimenting myself, however crudely, it didn’t matter to me, with some orphaned wires I had accumulated. The materials used have an impact. Online I looked at topics like dielectric absorption, which was helpful, but drew no correlation between that and sound quality, per se...only measurements, mostly. But, the experimenting taught me more about what relates to sound quality than anything else. It’s just that I had to carry out repeated experiments over time, a few years in fact for me, in order to be sure I was not inferring something that wasn’t there. Even so, I don’t have any scientific "proof" of what I hear, and yet it’s demonstrable all the same.

@kumakahn: If you can record ~5000 samples of 10sec audio clips approx 50/50 split between cable A and cable B, then we can attempt to train a deep LSTM neural network to distinguish the two.   If the computer can learn a difference, then we can work on reverse engineering what it picked up on.

@kumakahn, 

It seems to me, a few decades back, AudioQuest wrote a "manifesto" about what they considered to be worthwhile goals in wire design. I read it, and for the most part, it is about as good a reference as any manufacturer ever produced in layman's terms. The only problem with it was that, while the principles appeared to be rather valid, AQ itself hardly ever followed its own advice altogether in any single cable design of theirs that I ever saw...they were usually compromised, by their own definition, one way or another. 

...not that AQ had any kind of lock on compromised wire designs.