While I won't speak to the level of matching quality, you most certainly can bias the tubes down. They may red plate at what is supposedly 80% of the expected dissipation, but all that means is the tubes are colder than average and require more ( negative ) bias voltage to keep them in their sweet spot.
This is kind of along the lines of Mesa's numbering system. They test their tubes and give them a number. The idea is that if you replace your current tubes with the same number, then you should require no bias adjustment. If you buy a hot set of tubes and you have a fairly cold bias, the tubes will sound horrid. Conversely, if you have a COLD set of tubes, even a modest bias will have them burning up.
A tech's standpoint on biasing is ALWAYS bias the tubes, regardless of the bias topology. Fixed, or cathode biased, you always check and set the bias. Some tubes will happily run at over 100% of dissipation ( max dissipation is a theoretical number really ), while others will not run up to max dissipation happily. The 70% rule, is only a rule because it sets so close to a happy medium, that it is considered safe for 99% of tubes. If you bias to 70% of dissipation, you can generally be assured that the tubes won't go into meltdown, or be so cold that crossover distortion is abhorrent.
So the question is really a matter of, will the tubes bias at all and not have crossover distortion, or do they just not bias to the theoretical number you are shooting for? 25 watts per tube is not necessarily an actual number. Some small percentage of tubes will not quite bias to 25 watts, while others will happily produce above 25 watts happily.
Not all tubes are the same. Some people like colder tubes because they break up sooner. Some like hotter tubes because you get more punch and volume.
