ShiftyRider":3vtdagaw said:
Read the first sentence here if you're still confused...
No confusion. You're arguing that only cases severe enough for hospitalization is the divisor (eg those definitively ID'd with the disease). Then trying to apply that death rate to the entire population. Given testing failures in the US there are relatively few diagnosed with Corona, but not hospitalized; so that is the basic math you are advocating. What if it turns out that there are actually 500,000 cases that have gone undetected because they are so minor. This has been hypothesized for the US by many reputable doctors. Then your stat for deaths % is wildly overstated. You can't treat the divisor as definitive and known when the testing failures are so bad. The divisor is an entire unknown in nearly all countries which is why the death rates are also a complete unknown.
Look at the South Korean data and you will find something like a 0.6% death rate. Why is theirs so much lower than anywhere else? Because they actually know how many had the disease through extensive testing. Then you have factors like heavy smoking culture in some counties, or massive populations of older people ,etc... heavily skewing other underlying factors that will need to be accounted for.
The real stats won't be known for many months and quite probably years (medical types generally move kinda slow).
ShiftyRider":3vtdagaw said:
vanhanbr, somehow what I wrote was confusing. I'm not saying car crash statistics don't exist. I'm saying the mortality percentage of car crashes is a simple formula...
# of car crashes with a fatality / # of car crashes
Do you see the denominator there? It wouldn't include car crashes that went unreported.
You do realize that there are studies and statistical models on unreported car crashes right? Doesn't mean they are on Wikipedia or in the major news outlets. Many of them are secret sauce info within insurance companies, etc... Doesn't mean it doesn't exist or that it isn't taken into account in various places.