But to add another remark on number of headphones tested though, one thing 'hard denialists' don't seem to realize is just how hard to prove of a statement they have chosen to defend: you can't just generally claim headphones (or headphone model X) do(es)n't change audibly with break-in until you've tested an absolute s***-ton of them. What if only 10% are improved by break-in? Does it become not-recommendable to new buyers to do a couple of tens of hours of break-in just in case they got one of those units?
Except these are your claims that break-in is possible in small speaker cones, which is a debatable subject and that break-in happens due to mechanical changes in cones, which is impossible to claim since you don't know the parts' repeatability.
The burden on truth is on you to show at least one example and show how it is sure that the difference is due to that, since you're the one claiming. Now, I get where you're getting at with attacking general classifiers and on that front I'd agree (despite that is just nitpicking), but you're hand-waving here on what exactly happens and why. And then...
... it should be easy to show at least one pair which shows important changes, which can be attributed to burn-in, shouldn't it? This one is hard on the same front as you're pointing out - it requires to establish a good comparison basis, thus you cannot reliably assume that a specific model needs break-in if you haven't tested the whole production run or a numerous sample...
Perception changing due to whatever the head thinks is a known phenomenon for anyone who played around (and heard changes) with a disabled equalizer. Degradation of pads and sound changes due to that + variability of on-head positioning are actually something that happens and causes significant changes in sound. Break-in/burn-in of headphones hasn't been proven nor shown in any reliable scientific publications and in experiments by non-scientific sources it is shown that nothing happens. Maybe because it can't?
As for non-recommendability of breaking in: we're talking placebo and esoterics here, which are utterly harmless unless ones stumbles on the idea that break-in with a very loud signal gives better results (being forced sometimes on the net) and burns the crap out of coils. Seems like a strange ritual. Unless one buys a bullcrap pair of headphones solely on the basis that they "will burn-in", which is similar to buying shoes smaller than you need and finding out that they just won't stretch how you expected.
Good discussion nonetheless.