I believe I gave you my answer.
Do it on Stevens system. And if you want to BT it, then follow the basic (and accepted) ABX methodology.
None of this " AB,,XXXXXXXXXX" get all the x's right or you lose bullsh!t.
ah well, in that case it was MY misunderstanding, sorry.
you might be right in that stevens system is the most capable of showing these differences. And indeed he quite clearly does believe there are definite differences that can be heard on his system.
But if we go further, is there a reason (for example) that you did not say 'do the test on my system'? After all, (presumably) you claim to be able to hear these differences too?
See, as you nominated steves system as 'the one', is it not reasonable to extrapolate that many many other systems (including all of the ones on this forum) will not 'as easily' show these differences? Further, if we accept that, then for the everyday enthusiast who's systems are NOT of the standard of stevens it would in fact be true that those caps won't make a lick of difference?
What use then to them? Surely the funds spent on a cap upgrade that will not make a difference in their system could be better spent elsewhere?
Indeed, would it not be useful data to know what standard of system would benefit? That alone becomes an interesting question.