I was watching this too, and was curious if they were really worth that. . .
We have all asked the same question about various high priced tubes in the beginning. So let me offer my opinion/experience.
As recent as two years ago, if you were seriously exploring the best sound available from tubes, you were on track to spending at least a couple thousand dollars, usually more. But there were no alternative approaches to the buy'em and try'em method. You became familiar with the legend and lore, purchased the tubes, and then exchanged notes with others.
With regards to the value of that legend and lore, note that many here have found the inexpensive 2C51 group at least the equal of, sometimes surpassing, the expensive 'legendary' tubes. The rest is commonsense. The most highly regarded tubes in this thread vary in price from a few dollars to several hundred, but the differences in SQ are subtle, not extreme. None stand head and shoulders above the others. So are you missing something important if you only explore the less expensive tubes? My answer would be No. (You will miss the disappointment of spending $400 on a pair of tubes only to find you prefer another $4 pair to be superior.)
About the PWs specifically, factory specs are not available, but here's is what we do know about the PWs and the tubes that followed immediately afterwards:
In both versions the cathode/grid/anode amplify the signal, and these components, materials and tolerances determine the SQ. As with all tubes.
With the PWs, after assembly, the glass tube is heated and pressed against the top mica plate, which is smaller in diameter than the glass tube inside diameter.
With the tubes that followed immediately after, the mica plates are slightly larger in diameter than the inside diameter of the glass tube, no heating and pressing required.
Neither of these glass/mica arrangements affect the performance of the cathode/grid/anode directly.
So the important question would be, did the cathode/grid/anode assembly change in some significant way that affects the SQ in some significant way?
Possible... But there is no consensus or objective evidence.