On the other side of it, there's selection bias--most identified cases have visible and notable symptoms, some very severe.

If you have a mild cough or a bit of a cold, you might not think anything of it beyond staying home from work, but that could be COVID-19 (and we might _never_ know that it was).

It is also possible that 2% is an overestimate. Unlike SARS for example, it seems many infected, some estimates of 80%, may not even exhibit outward symptoms, like the high fever of SARS. This of course makes containment very difficult, maybe even impossible, to achieve.

Based on a WHO official's statements they found no evidence that this is happening when they went to China. https://www.statnews.com/2020/02/25/new-data-from-china-butt..., https://twitter.com/HelenBranswell/status/123233479099875328... This would push the current numbers as more accurate than not, as well as projected death rate up for the future as the poster you responded to said.

>Had anyone seen death and recovery by age?

There was a recent (Feb 17) study here with mortality crosstabs (see ‘results’): http://weekly.chinacdc.cn/en/article/id/e53946e2-c6c4-41e9-9...

Case fatality rate (N_deaths/N_confirmed_cases):

[0-9]: NA

[10-39]: 0.2%

[40-49]: 0.4%

[50-59]: 1.3%

[60-69]: 3.6%

[70-79]: 8.0%

[>=80]: 14.8%

Indeed. I've seen that, but not deaths,recoveries data. The above has the potential underestimate problem from case growths.

They include a breakdown by date of onset as well, which isn’t exactly the method you suggest (but I read somewhere that recovery time is ~2 weeks for mild cases and ~3-6 weeks for severe).

Case fatality by date of onset:

[<Dec 31] 14.4% (n=105)

[Jan 1-10] 15.6% (n=653)

[Jan 11-20] 5.7% (n=5417)

[Jan 21-31] 1.9% (n=26468)

[>Feb 1] 0.8% (n=12030)

That does make the overall fatality rate look quite a bit worse than the naive calculation.

From the Diamond Princess we've got a pretty good idea about mortality. It's almost a conspiracy theory: let the disease run its course in a controlled environment to see how bad it is.

Diamond Princess isolated people, no? That is unlike what we'll see in the broader population.

Even apart from undercounting of case #s, we know that the # of severe cases is 5% upper bound which is much more likely to be accurate and the deaths are a fraction of that so I would bet on the number being on the lower end than the higher end. Even a 1-2% mortality rate is fairly substantial tho -- 5-10x as deadly as flu.

> Has anyone seen death and recovery by age?

This calculates Death Rate = (number of deaths / number of cases) which seems like how you're saying not to, but does split by age.

https://www.worldometers.info/coronavirus/coronavirus-age-se...

The case numbers are being severely undercounted though. People aren’t getting tested unless they present severe symptoms.

If you look at the Diamond Princess, where seemingly everybody was exposed to the virus, the actual death rate was about 0.4%, which is within error bars of the flu.

3 more people just died, so it's now 0.8%.

Given how it can take weeks to die after being infected, I doubt they will be the last deaths.

Cruises also skew significantly older than the overall population. It’ll be awhile until we know the final numbers, but what we know so far is reassuring.

This will likely go higher as many cases are still hospitalized with severe complications.

Please define "many cases" and "severe complications". I'm interested.

Good luck finding any official numbers but a UK citizen from the cruise died this morning, so it's not over yet.

https://www.bbc.com/news/uk-51677846

Yes, people have done studies and it shows if your above 50 you have a munch higher chance of dying from the disease.

I’d say though you’re fear mongering posting a comment like this, why ? Maybe you don’t realise ?

> I’d say though you’re fear mongering posting a comment like this, why ?

If you're just stating that without posing a counter argument then I think you have failed to understand what I am saying. And that is precisely why I posted.

1. The methodology being used to arrive at the 2-3% figure is unambiguously vulnerable to bias. It may be the case that other biases (under-reporting mild cases) cancels out the effect, or maybe not.

2. The same methodology when applied to SARS was provably biased, and an alternative seemingly equally valid estimator was accurate for SARS and gives a much higher estimate for the mortality rate of coronavirus.

If, in fact, we are off by mortality estimates by a factor of 4 it needs to be more widely known because the appropriate economic trade-off in handling is different at the two mortality levels. (In particular, if the mortality rate is much higher for working age people then economically disruptive countermeasures are more justified.)

I'm doubtful that any comments on hacker news are likely to result in public panic. I think for the most part the audience here is capable of reasoning about these things and making prudent decisions.

Are you qualified to make these claims and assumptions though, do you study viruses ? You’re a scientist who understand these things well?

Ok it’s fine to speculate, but you’re throwing up some big figures in what’s already a panicked situation and pretty much everything you’re saying is based purely on speculation and anecdotal evidence.

I just find it odd health organisations have already done studies, published numbers yet people all cover the internet (on HN of all places) seem to feel like they want to exaggerate the issue, refute the science and don’t trust the existing numbers. Why ?

I am not, but the paper I was referencing was written by people who are: https://academic.oup.com/aje/article/162/5/479/82647 My interesting the question arose out of my experience with non-epidemiological process estimation.

I spoken with a professional epidemiologist and they also agreed.

If CDC were saying that they considered this case and concluded that d/c is a more reliable estimator, I'd be happy to hear it. Instead, it appears d/c -- which is a good estimator for viruses in steady state-- is being used because it's just the default and the only research I can find which has actually evaluated this question found that d/c underestimates in a rapidly spreading infection.

a doctor friend of mine, familiar with the situation, tells me that 2% is likely an order of magnitude high. That is because in China, only the very ill are getting diagnosed at all. Therefore. the number of deaths is roughly in line, but the number of cases is likely an order of magnitude low.

> That is because in China, only the very ill are getting diagnosed at all.

I don't think this is true, considering China has built new hospitals just for patients who are tested positive but have light symptoms [0].

China also sent 12500 testing kits to Japan [1].

[0]: https://www.zhihu.com/question/369522812

[1]: https://mainichi.jp/english/articles/20200226/p2a/00m/0na/00...

Disclaimer: I'm from China.

In reality the vast majority of people get no symptoms or just basic cold symptoms and will have no idea whatsoever they ever had “COVID-19”.

Both 2% and 7% will end up being massively overstated fear-mongering in a “pandemic” that never really actually qualified as such, and will kill less than 1/100th the number of people who die from Flu this year.

But everyone loves a good disaster story.