When that chart turns negative, it means that the ratio of COVID deaths to non-COVID excess is less than one, ie there is more non-COVID excess death than COVID deaths. When there is no non-COVID excess, but still COVID deaths, the chart shoots up; when there’s none of either, it clamps to the black line. This is a log-scale chart of their ratio; that solid black line represents a ratio of 1, ie, equal amounts.
The chart going below the black line represents positive excess death in a ratio that favors non-COVID excess deaths.
Edit: I’ve added a log chart of excess deaths, with the black line at zero, ie, the expected amount.
https://zmichaelgehlke.com/images/shaded-excess-mortality.pn...
> concepts like "non-COVID excess" doesn't even start to make sense
Sure they do: excess deaths which exceed the number of registered COVID deaths in the same period. We can detect that by looking at all cause mortality and subtracting identified COVID mortality.
Negative excess deaths through the pandemics were pretty common and very well explained.
Anyway, such concepts like "non-COVID excess" doesn't even start to make sense. Either you are determining the death causes, or you are doing statistical analysis over unknown causes.