Can you link me to a study which does a reasonable job of attempting to come up with a "Covid deaths prevented by vaccines number" while taking into account the effects of increased seroprevalence, decreased virulence of the virus itself, and the "no more dry tinder" effect?