Your mental model is mostly meaningless. There's no such thing as a generic death rate for a disease, because the death rate is dependent on numerous factors. You cannot compare the death rate between countries with different demographic profiles and ask "which is correct". How old are they? How fat are they? The idea that 5% of a person's social network is going to die is not accurate because social networks aren't evenly distributed. Most people don't have many contacts over 80 unless they're over 80 themselves. Most people will see very few deaths to people they're connected to. Old people will see a lot. And it's sad, but most of the people dying are already above life expectancy and close to death from alternative causes.

Hospitals being overwhelmed is an exaggerated problem. It's not happening in many places. Most hospitals only need to provide a small amount of care and can actually scale this up rather quickly. You hear about people being triaged in Italy, but you need to remember that the people being triaged are likely to die regardless of whether or not they're admitted to the hospital. Averages are not your friend here. If 50% of ICU patients survive, that does not mean that everyone going into the ICU has a 50% chance of surviving. If we get to a point that triaging is needed (I think dubious in major US cities), that does not mean fatality rates will skyrocket.

Social distancing has been pretty darn effective, and it's probably a good idea to cautiously continue. But we also need to be pragmatic.

> Hospitals being overwhelmed is an exaggerated problem. It's not happening in many places.

...because people are introducing restrictions to keep the number of cases down.

> Your mental model is mostly meaningless. ...Social distancing has been pretty darn effective, and it's probably a good idea to cautiously continue. But we also need to be pragmatic.

Ok, but at least mine I told you what the exact numbers are and where they came from. What are your numbers, and where did they come from?

You can't not have a model: you can either have an explicit one that you can reason about and question, or you'll have an implicit one you can't reason about or question.

You've waved your hand about why 5% death rate for a "fast burn" scenario is probably not exactly the precise number. I never said it was; but I think there's a pretty good chance it's in that ballpark.

You, on the other hand, haven't said what you think the death rate for a "fast burn" will be.

It is IRRESPONSIBLE to recommend a course of action without thinking through the consequences. Just wishing that 5% of people won't die isn't going to prevent those people from dying.

I'd like a 0.001% fatality rate in a "fast burn" scenario as much as anybody. If you can convince me that's a good number, I'll go around and try to convince others. Until that time, I'm going to continue going with the best guess I have, and a guess which lots of experts seem to be making as well.

> ...because people are introducing restrictions to keep the number of cases down

I mean, yes, I do agree that social distancing is a good idea and that we should continue to do so. I acknowledged that much. But my real point here is that hospital capacity can scale up significantly higher than you might assume. Average level of care goes down for each person, but that's mostly fine. We can't treat people for this disease, so all the hospitals are doing is trying to keep people alive long enough to get over it themselves. A half hearted benefit of this is that since there isn't much you can do, you don't have to do much to help people as best as possible.

> Ok, but at least mine I told you what the exact numbers are and where they came from. What are your numbers, and where did they come from?

When I say your mental model is meaningless, I'm referring to "which 20-40 people do you want to die so that the rest can go back to work". You're thinking in averages, and that's not the case. The CFR of any population is going to be determined, foremost by the age distribution. I don't know if this is still true, but a while ago the CFR in italy was lower in every age group relative to China but worse overall due to age group differences and simpsons paradox. The concept of death rate is truly meaningless. I'm not saying you're wrong, I'm saying its a pointless discussion. But sure, range of 0.5%-7% for a randomly chosen locale. Less than 1% average for those under 60 years old across the board.

For the record, I don't think rushing to herd immunity makes sense either, but there is a middleground.

> It is IRRESPONSIBLE to recommend a course of action without thinking through the consequences. Just wishing that 5% of people won't die isn't going to prevent those people from dying.

I think its pretty naive to say it's irresponsible to not have explicit threats on the disease but hand wave explicit threats on the economic consequences. I think it's better to let a 90 year old person die than to throw a 20 year old into poverty personally. Globalism is complicated. Depressions can cause famines, the rise of violent authoritarianism, and just general despair even if your particular society holds together. Poverty sucks, and it disproportionately hurts people who are already poor or in poor countries.

Do you have a link for this 500,000 tests per day? Every article I am reading says they only started ramping up testing in the past few weeks.

> And the total number of swab tests done by 4 April was well over 1.3 million.

https://www.bbc.com/news/health-52234061

The population of Germany is 83 million. I find it hard to believe that they've tested more than 5 million or so people in 10 days.

So I did mangle my numbers somewhat, thanks for checking. As of April 5th they'd done 1.3 million tests, found 95k cases, and had 1.4k deaths [1]. Presumably there are at least some people tested multiple times; so those 1.3 million tests don't represent 1.3 million people.

The only way to get the "die with the best medical care" number below 0.1% is if you have a million untested people who are sick.

If I was a proper statistician, with those numbers I could probably come up with models or estimates of how many possibly untested people there might be in Germany. My instinct is that number can't be millions, because at least some of those millions would be sick enough to go to the hospital, and then all of their connections would be tested and be found out.

But even then, what we really need is the "would have died without medical care" number, which is what the number would look like if we did a "fast burn" and let everyone get sick at once. For that number to be less than 1% (which would still be pretty dire), we need closer to 2 million unknown cases. I just don't see how 2 million positive cases can hide from the kind of testing Germany is doing. The same goes for Korea.

[1] https://www.who.int/docs/default-source/coronaviruse/situati...