I've used about 8 or so eh 12ay7 ... they measured very well ... around 6 of 8 were very usable in first-stage gain applications. The other two were somewhat lower emmission.
Matching across stages seemed to be reasonably consistent, but selecting for this would also eliminate prolly 2 in 8 again.
Signal-hum margin (in a high-gain application) was very good, about the same as the eh6922 (different type pinout, lower mu and rp) and just a little less than the jj5751 (same pinout, higher mu and rp).
Harmonic distortion spectrum was nearly all h2, with low and sparse h3+ and very similar to the eh6922 and jj5751. Most of my 12AX7, in contrast, have much 'busier' harmonics h3, and greater.
....
So, in my experience, eh12ay7 is a very good performer .. I use it alot in place of the ax7, where it can tame some of the 'over the top' gain and attendant increased thd, particularly in the higher registers.
IF I need that extra bit of gain, I use a jj5751 - it retains that nice, sparse higher harmonic distortion distribution but with a dB or two extra level.
...
I have a few ge 6072 nos tubes, which were reasonable pricey ...
They did perform very well too, but surely didn't 'eclipse' the modern eh version .... possible one could convince onself ( as I have!) that it sounds a bit more 'refined' despite no real advantage in the basic measures ;D
So, for some 'signature pieces' there can be a good case for fine nos, priced accordingly, but for most everything else, well selected current production can measure a decent bit better!
So, big fan of current production eh 12ay7 and jj 5751 in a high gain app.