@deonb
You lost me on that one. Care to translate?
Tim
IFR -> Infection Fatality Rate. Of all the people who got infected, how many died.
CFR -> Case Fatality Rate. Of all the people who tested positive, how many died.
CFR is easy to calculate, but not that meaningful. IFR represents the real risk, but to calculate IFR you need to have either perfect testing or great sampling. There are 2 places so far that did that:
Vo, Italy: Tested their entire population. Of everybody who was tested positive, 1.1% died. It's a small sample, but it holds.
South Korea: They tested a sufficient number of people to achieve control with only a test+track strategy. The only way to have achieve without lockdown is to have a very high (near 100%) coverage of tests over cases. Otherwise without the lockdown, death rate would be growing exponentially (which it doesn't over there - it's been stable for 4 weeks). Of everybody who tested positive in South Korea, 1.7% died.
Either way, it points to a Infection Fatality Rate (IFR) of at least 1%. This gives us an easy upper bound of the infection rate - 100 times the death rate. The CFR represents the lower bound. Current CFR is between 2% to 10%, depending on how thoroughly a country tests. So the lowest bound of the infection rate is about 10 times the death rate.
So e.g. if you have 10000 deaths (US today), then you know you have between 100 thousand and 1 million people infected.
EXCEPT, you don't actually know how many are infected today. You know how many were infected 2 weeks ago. (The average time from infection to fatality is 2 weeks). So with 10000 deaths today, we've had 100 thousand to 1 million infections 2 weeks ago. We've had a MEASURED increase of 10x since 2 weeks ago (33000 cases to 330000 cases). So if 3 weeks ago we in actuality had 1 million cases, then today we could have 10 million cases. But there were so many changes that went into those 2 weeks, including partial lockdown measures, increased of number of tests, change of testing criteria etc. that it would likely be lower than that. Either way, it does give us a upper bound of infected today, which is 2.7% of population max.