Replies

Well that was just replied straight from my pings page, didn't notice y'all had a conversation still going :p

You never give the raw numbers on “resolved” cases. You just keep saying 29% dead.
China started posting the infected and dead numbers four weeks ago. The death toll was 2.5% on their first post. The China death toll has hovered consistently around 2% since that first post.

Well apparently you only looked at calenel's history, cause I definitely have several posts with raw numbers. Some even show step-by-step math! And China's 'official' death toll has never been that low, it started close to 50%. (To be expected, as recoveries take the entire duration of the sickness, while the deaths come quicker.) As of a couple days ago we were at 29%, and right now they show 21%. (I'll give you all the raw numbers below.)

And, Wiki is tracking all the statistics in real time - they leave the mortality rate box empty on all their charts.

Wikipedia is not the best source for real data on anything, but let's take this rabbit hole. Some random guy on wikipedia (incorrectly) says the rate is 2%, with citation 15/16. (Even though you said wikipedia doesn't state the mortality rate.) If we go to footnote 16 ("Wuhan Coronavirus Death Rate". Worldometer. Retrieved 5 February 2020.), we'll find a whole article about this, that SAYS EXACTLY WHAT WE'RE SAYING.

- In other words, current deaths belong to a total case figure of the past, not to the current case figure in which the outcome (recovery or death) of a proportion (the most recent cases) hasn't yet been determined.
- If we use the flawed formula (deaths / cases) we get: 813 / 37,552 = 2.2% CFR (flawed formula).
- An alternative method, which has the advantage of not having to estimate a variable, and that is mentioned in the American Journal of Epidemiology study cited previously as a simple method that nevertheless could work reasonably well if the hazards of death and recovery at any time t measured from admission to the hospital, conditional on an event occurring at time t, are proportional, would be to use the formula: CFR = deaths / (deaths + recovered) which, with the latest data available, would be equal to: 910 / (910 + 3,324) = 21% CFR (worldwide)
- Finally, we shall remember that while the 2003 SARS epidemic was still ongoing, the World Health Organization (WHO) reported a fatality rate of 4% (or as low as 3%), whereas the final case fatality rate ended up being 9.6%.

So the alternative method (what me and calenel are using) isn't fully accurate, as that would require the deaths to be spread evenly through the course of disease. Instead, they spike around I think week 2, and then fall off (so more recoveries) as you move into weeks 3-4+. So this formula will start with a higher death rate as there's no recoveries, then as more cases resolve, it'll fall somewhat until it reaches close to the actual rate.

Just for fun, I looked up the USA mortality rate for the Spanish Flu epidemic. The CDC says about 2.5% with a national infection rate of about 30 million.
You need to take your 20% mortality prediction to a hospice!

Why compare this to the flu? It's not the flu, it won't have similar infection or death rates except entirely by coincidence. Also, you do realize there's several types of flu? Here's a short list, with their approximate mortality rates. Many of the ones not listed are 0-2%, or are non-communicable to humans (swine/avian/equine flus mostly). Note different strains within these might be different as well.

A/H2N2: ~.1%
A/H3N2: some strains .1%, others up to 20%
A/H1N1: up to 20%
A/H3N8: varies, 6-35%.
A/H7N9: ~30%
A/H5N1: ~60%
Flu Type C barely does anything to you.
Flu type D is not even transmissible to humans

And here's your numbers, direct from the Chinese site. https://ncov.dxy.cn/ncovh5/view/pneumonia?scene=2. In parentheses is John Hopkins's worldwide totals. https://gisanddata.maps.arcgis.com/apps/opsdashboard/index.html#/bda7594740fd40299423467b48e9ecf6. They may not fully line up, as I have no idea how recently updated the Chinese stats are, so there may be a bit of overlap/gap.

Total Confirmed: 40,235 (40,540)
Total Suspected: 23,589 (no stat)
Severe Cases: . . 6484 (no stat)
Deaths: . . . . . . . 909 (910)
Recovered: . . . . 3328 (3383)

Mortality Rate: 909 / (909 + 3383) = 21.18%