>> that is the natural "desire" of a virus driven by evolutionary necessity. killing the host is severely maladaptive and selected strongly against.

>>you really only get that sort of inverted gradient when you have non sterilizing vaccines with meaningful partial efficacy.

My understanding of Marek's (disclaimer that people are not chickens and this discussion is not necessarily analogous but the principles of viral evolution are still the same) is that it is a scary example of OAS precisely because vaccinated birds live LONG ENOUGH to go to market, not to a ripe old age.

You're right that it's a fundamental principle that viruses (metaphorically speaking) don't "want" to kill their hosts. But they don't "care" if the host dies AFTER serving as an incubator and transmitter. That's long enough.

My understanding of OAS is that it's so dangerous precisely BECAUSE if a virus develops mutations that give it phenomenal transmissibility, the window within which it's strategically bad for the host to die can be a lot shorter, because we don't need it for very long.

Which means that deadliness ISN'T being selected out along normal patterns.