> Why a year? Why not daily? The vaccinations happen throughout the day.
Because vaccinations take months, perhaps up to a year. In my country it's been at least half a year by now and we're at something like 55-60% of the population.
> Why not per lifetime? You can only get the second dose once.
Lifetime makes it an even worse comparison for natural myocarditis since you can get that in any year of your life. (But if need arises in the future for annual boosters against new strains or something like that, chances are that annual risks will again be the number to look for.)
> I could pick any interval, either customary (hourly, daily, monthly) or completely arbitrary (per 47 seconds) and get any number as a result.
You can pick garbage methodology and get garbage results, agreed. You can do pretty much what you want.
> Most of reported vaccine carditis happens a few days after second dose, so I compared those with the average incidence for a few (5) days.
Which is a complete red herring since how do you know that other kinds of myocarditis don't happen a few days after the initial viral infection as well? Either delay is completely irrelevant since shifting infections of individuals in time does nothing to overall statistics of incidence.