Back in February, with Chinese data.
The quickest correction is this: The "active cases" are to be read as "tests with positive results."
Some of those are asymptomatic, probably not many. Others have sniffles. Most are not admitted to the hospital--and therefore it'll be hard to clear those cases.
The positive tests are a fraction of those actually infected. Hospital admits are a much larger fraction of serious cases.
Hospitalization information is harder to come by. Consider https://covidtracking.com/data/state/new-york . Since we can't get data for the country, let's see how it works for a state. NY has decent data reporting, if you get past their changing what the numbers mean mid-process.
12,192 deaths (when I'm writing, 12:10 pm CDST on 4/17/20), 23,887 recovered. Sounds horrible. One third die. Knowing there are 17,735 still hospitalized ... Don't know if that helps or hurts.
Number of positive tests: 222,284. Since the test has about 30% false negative rate, the number of positives just for those tested is more like 317,000. That number won't be cleared because nobody knows to check up on them. But since deaths are more likely to be counted than actual infections, that's still not the right number for measuring mortality, esp. now that NY has taken to recording COVID deaths based on posthumous clinical diagnoses and not just the PCR test. You can imagine the hypothetical situation where there were only 20 tests for the last month but 200 deaths from COVID, meaning the mortality rate is > 100%.
But now it's 12,192 dead out of 222,284 reported, and we're down to 5%. Or more likely 12,192 dead out of 317,000 or 3.8%--but you know, some of the dead weren't tested when alive, so they're not part of even that posited 317,000. And because of the sampling bias, "all the dead" but only "part of the infected", that's just an upper bound.
Pitch in demographics and you see that most of the dead are 70 years old and above. Go below 60 and the death rate drops. Take those under 18. If 5% of adults are infected, it's been the same for minors. But only a handful of deaths have been recorded for that group--their mortality rate isn't zero, but it's damned close to it. Meanwhile, for seniors, it's higher. One report says for confirmed cases ("positives') it's over 4% for those 70-80 and close to 8% for those over 80. Look at the US age pyramid and you get some idea how to use those numbers. The same source says 1.38 for confirmed cases as 0.66% for all cases (but that last number is still a crap shoot). Again, no single mortality number will work for the population as a whole--the average masks the risk and serves to scare some unnecessarily and make some think the risk is lower than it is.
New York's unlikely to have 40% of its population infected, so the testing bias is towards those most likely to be infected, which means displaying symptoms. It would be nice if that were so, of course. The more infected for that number dead, the lower the mortality rate and the closer we are to herd immunity.
= new reply since forum marked as read
