I have posted this extract from a professional tube tester in this Forum before, and is worth posting again, as it ties in nicely with some of your observations. This relates specifically to 5998’s / 421a’s.Tube testers are a bit of a hobby horse for me, so those who have heard this rant before will need to excuse me.
<rant>The thing about (most) tube testers is they are designed to test if tubes are bad, not if they are good or matched. The engineers at Hickok, HP and etc. had an almost impossible job: to pick a single operating point from the entire phase space of currents and voltages to test to determine if a tube needs to be replaced. And by in large they did an amazing job solving that problem, but IMO, the relevance to audio is limited. Yes, if you pay for a "NOS" tube and it tests weak there is a problem. And I always test tubes (and adapters) for shorts before putting them in an amp. But in general the more tubes I test the less correlation I find between test results and audio quality. Weak tubes can sound great, and last a long time in an audio application. Or put another way, your amp and your ears is really the best way to test tubes for audio.
And then there is "matching," which is a whole 'nother kettle of fish. Is the operating point you test the tubes at the same as the operating point in your amp? Probably not, so just because tubes match on your tester doesn't mean they are matched, and visa versa. There are fancy testers that can test tubes at an arbitrary operating point, or even better over a range of operating points (curve tracers), and those are truly useful tests. But they are even more expensive than a "plain" Gm tester. Some tube dealers will include the curves with tubes you buy, and you can buy a lot legitimately tested and matched tubes for what a high end tester costs these days. </rant>
All that said, if you're really interested in tubes, a tester is a useful (and fun) thing to have, and as far as hobbies go it is way better than cooking meth, as a tell my wife.
“The PC/TC figures are pretty much a standard that people use to buy / sell valves to give an idea of their ‘goodness’, but in reality the 421A / 5998s are very high power triodes used mainly in Military applications back in the day, so to be honest for the purposes of matching the valves for your preamp, the PC score is really irrelevant, whether a valve can drive 90mA or 40mA is of no consequence when it is in circuit driving 5/6mA, they will pretty much all work until they become noisy or completely exhausted. A more useful parameter to match against is the TC (transconductance, loosely described as the gain of the valve), so use that as a basis for matching, but again don’t worry unduly, because I think you would be hard pushed to hear the difference between a valve with a TC of 9 or 14, but always good practice to get as close as possible. The main criteria with your amp I would suggest is how noisy they are, as that directly affects the listening experience, old valves can become noisy and microphonic, as you know.”