>for me relation of cases to tests and similarly deaths to tests is a metric to check to get an idea if only those already symptomatic are tested (resulting in higher deaths per case) vs. contact tracing to find also asymptomatic carriers. Second is part of the reason for some good death/case stats, but will have less hidden herd immunity, as those "elsewhere" in hiding are already counted.
>
II believe all the data for "cases" is bogus anyway. According to the latest (29 May report) serologic testing for COVID-19 in Florida (123k tests) 4.43 % of those tested have the COVID-19 antibody. I don't claim that this is at all random or reliable testing but if is any way close to the real % it would indicate 960k cases in the state - not even close to the 60k cases identified by testing.
>
>>I'm not saying "anyone's" stats are the correct ones, as different sources have reported different numbers.
>>
>>My only point from a few posts ago is that daily stats (ANY daily stats) are nearly worthless. They can represent an accumulation of test results or death counts going back over several days (or even weeks).
>>
>>To me, here are the "best" (relatively speaking) concepts of metrics for this area, when put together to frame a scenario:
>>
>>- % of increase or decrease in weekly death count
>>- deaths per square mile
>>- population per square mile
>>
>>Here is something absolutely sobering:
>>
>>In the U.S. as a whole, there are .03 deaths per square mile
>>In Manhattan NY, there are 178 deaths per square mile. (Stands to reason, Manhattan is geographically small and yet you have about 1.6 million people)
>>
>>See attached, as of end of last week, here are the top U.S. counties in terms of deaths per square mile