“More infected means lower death ratio, yes?”
If you assume that all have passed the threshold to active status and then resolved. Who knows how they are counting? But guaranteed most of those cases have not yet resolved, and if Brazil’s HCS collapses, you get a huge spike in deaths, both direct and indirect. But in general, more infected means more dead in direct proportion. Again, there is no vast hidden herd of already infected that somehow survived and gained immunity without being detected.
In a situation where the testing and HCS are saturated - like Brazil - you can ballpark the case count by assuming that only the serious and critical cases are being officially reported and that there are another 4x mild and asymptomatic, and then if there is community spread (usually these things go together) another 3 cohorts (based on R0 of virus) of the same size incubating. Altogether you get 20x the reported cases as your upper limit for infected - with a lot, 75%, of the cases TBD as far as severity goes.
That’s in a saturated situation. If testing is keeping up or catching up and you have contact tracing going on, and people practice aggressive social distancing, that reduces the multiplier, possibly by a large amount. But overall, 2% or more of the cases will be fatal. Barring a miracle cure.
“Strategies to reduce COVID-19 transmission at the Smithfield Foods Sioux Falls Pork Plan”
- see https://assets.documentcloud.org/documents/6876342/Smithfield-CDC-report.pdf