Go Back  Bike Forums > The Lounge > Coronavirus/COVID-19
Reload this Page >

Fatality rates in CA

Notices
Coronavirus/COVID-19 Discussion of the novel coronavirus

Fatality rates in CA

Old 06-21-20, 07:19 PM
  #76  
79pmooney
A Roadie Forever
 
79pmooney's Avatar
 
Join Date: Oct 2014
Location: Portland, OR
Posts: 8,398

Bikes: (2) ti TiCycles, 2007 w/ triple and 2011 fixed, 1979 Peter Mooney, ~1983 Trek 420 now fixed and ~1973 Raleigh Carlton Competition gravel grinder

Mentioned: 99 Post(s)
Tagged: 0 Thread(s)
Quoted: 2279 Post(s)
Likes: 0
Liked 696 Times in 483 Posts
Originally Posted by 3alarmer View Post
...I only took the minimum to graduate in my field, which was not especially recognized as anything other than "Survey of Statistical methods for Dummies."
However, I do have a lot of practical experience with people spinning numbers, because I worked in a couple gubb'mint jobs. In the SSA, people literally live and die over numbers. Careers are made and lost.

The reason I ended up in California in the first place was because the people in the office I transferred to here were better at spinning numbers than the office I left in TN.


So it's not like I don't understand the principles of bending the truth using numbers and questionable interpretations of them.
Engineer here. A saying I had posted for years. "Facts are brittle. Truth is elastic." And yes, people bend it all the time.
79pmooney is offline  
Old 06-21-20, 11:01 PM
  #77  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
Originally Posted by 3alarmer View Post
...I only took the minimum to graduate in my field, which was not especially recognized as anything other than "Survey of Statistical methods for Dummies."
Yeah, I've taught that class :-)

However, I do have a lot of practical experience with people spinning numbers, because I worked in a couple gubb'mint jobs....So it's not like I don't understand the principles of bending the truth using numbers and questionable interpretations of them.
A very important skill not enough citizens have. Did you know 78.3% of all statistics are made up on the spot? There's likes, damn lies, and statistics.
RubeRad is offline  
Likes For RubeRad:
Old 06-21-20, 11:10 PM
  #78  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
Originally Posted by 3alarmer View Post
...the other data, of course, would be increasing hospital admissions, and the 12 other states where this is happening. All those people in all those states are currently free to hop in the RV with the kids and granny, drive over to the interstate, and follow the yellow brick road straight to California. So I'm getting a tingle in my fallacy sense. California, as much as we might wish it, is not a separate island state, protected by a large wall. This is in addition to my general unease over your bolded statement.
I'm not trying to make any prognistications, just looking at the (slightly) downward slope of the deaths-per-day graph from the recent past.



...so what's your position on testing, tracing, and isolating positive cases ? Horse has left the barn ? Too little, too late ? Are you still going with isolate the vulnerable and let the rest of us get on with our lives ? I recall that was at one time your position on this, and I find it somewhat offensive. Not that I'm easily offended, mind you. Once you've been "hot", it's pretty hard to take offense at much of anything.
Funny, I don't recall stating a position like that, or holding it privately, or any position in particular. If I did, it was just wondering out loud, and I've forgotten it. I've heard science-types say there needs to be testing early and often, that's the way to safely allow the public to be public and quickly identify and contain outbreaks, and that sounded right to me.

However, the radical difference in fatality rates does suggest that we should at least consider 'isolate the vulnerable, and let the rest get on with our lives'. But 'get on' may mean or require lots of testing of the general public, I don't know.

This is an interesting read about pooled screening (note NOT 'pooled testing' in which every individual receives results from their pool's test -- that approach has regulatory problems) and that also makes a lot of sense to me. Basically the way it works is, every student, professor, staff on a college campus submits a sample once or twice a week, they're tested in pools of 16. Pooled tests that turn up positive, individual's samples are re-tested (you don't even make them resubmit, just hold back half the original samples). So you basically cut your testing effort/workload/cost/time by a factor of 16.
RubeRad is offline  
Old 06-21-20, 11:19 PM
  #79  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
Originally Posted by wgscott View Post
I'm still not at all confident I am doing this right, but if you look at the last 50 days for both graphs 1 and 3, there is a roughly a 2.1-fold increase in cases, but a 2.4-fold increase in testing. I falsely thought 3 was much flatter than 1, but it is about the same (possibly slightly steeper slope).
Hmm. It's hard to tell exactly, since the graphs are so jaggedy. From the May tic-mark to the present, the cases trendline looks to me to be a little less than 2x actually (1800-3400?). For the testing, I guess it looks like 30,000 at the beginning of may to ... uh, I'm not sure what to call now, 65,000? 70,000? That most recent value of 80,000 may be just an anomalous spike.

I wonder if at some point the fatality rate falls because the virus picks off easy targets early on, and later the remaining pool of the population is tilted towards less-at-risk demographics?
RubeRad is offline  
Likes For RubeRad:
Old 06-21-20, 11:22 PM
  #80  
3alarmer
Friendship is Magic
 
3alarmer's Avatar
 
Join Date: Nov 2010
Location: The Big Tomato
Posts: 18,756

Bikes: old ones

Mentioned: 258 Post(s)
Tagged: 0 Thread(s)
Quoted: 17977 Post(s)
Liked 1,653 Times in 1,208 Posts
Originally Posted by RubeRad View Post



Funny, I don't recall stating a position like that, or holding it privately, or any position in particular. If I did, it was just wondering out loud, and I've forgotten it.
...it was in that other forum that I can no longer access, and even if I could, I'd be too lazy to look for it. But I recall being quite surprised by it at the time. If you recall, there was much noise and consternation about the shut down economy, and IIRC, it was in that context. Ironically, the longer this goes on, the more convinced I become that this is probably how the whole thing will play out in the United States. "Teh vulnerable" (whoever they are, and however we figure out how to determine that) are going to end up sucking hind tit in this pandemic.

Everyone will be filled with thoughts and prayers for their premature demise, but hey, they were vulnerable, and should have known they were and taken precautions. The rest of us need to get back to work. That sort of thing. It's pretty easily envisioned. Vulnerability is not something we especially cherish, and it elicits little sympathy in 21st Century America. Things are tough all over.
3alarmer is offline  
Old 06-21-20, 11:27 PM
  #81  
3alarmer
Friendship is Magic
 
3alarmer's Avatar
 
Join Date: Nov 2010
Location: The Big Tomato
Posts: 18,756

Bikes: old ones

Mentioned: 258 Post(s)
Tagged: 0 Thread(s)
Quoted: 17977 Post(s)
Liked 1,653 Times in 1,208 Posts
Originally Posted by RubeRad View Post

I wonder if at some point the fatality rate falls because the virus picks off easy targets early on, and later the remaining pool of the population is tilted towards less-at-risk demographics?
...given the random nature of exposure at this early stage in the pandemic, I'm not seeing that as a rational odds quote. There's still a huge population that (in theory at least) remains unexposed at this point in time. All those early speculative studies about how maybe large numbers have already been exposed or had the virus and remained asymptomatic or recovered nicely have been discredited (from what I read).
3alarmer is offline  
Likes For 3alarmer:
Old 06-21-20, 11:38 PM
  #82  
Hondo Gravel
Viking Berserker
 
Hondo Gravel's Avatar
 
Join Date: Jan 2018
Location: Hondo,Texas
Posts: 1,813

Bikes: Too many Motobecanes

Mentioned: 3 Post(s)
Tagged: 0 Thread(s)
Quoted: 1250 Post(s)
Liked 1,103 Times in 712 Posts
I see too many I don’t give a F people around where I live. Luckily I’m in a rural area so staying isolated is easy. HS strength and condition camps all started then were canceled when people involved started getting infected. The kids will go back to school in mid August then there will be a spike and school will shut down again. Football is a contact sport obviously and the band and all that goes with school stuffed together like sardines. I guess we will see what happens. I put on my KISS mask when I’m in a store or enclosed establishments. The KISS mask makes me a cool middle age rocker that stills rocks at least I think so.
Hondo Gravel is offline  
Old 06-22-20, 07:53 AM
  #83  
3alarmer
Friendship is Magic
 
3alarmer's Avatar
 
Join Date: Nov 2010
Location: The Big Tomato
Posts: 18,756

Bikes: old ones

Mentioned: 258 Post(s)
Tagged: 0 Thread(s)
Quoted: 17977 Post(s)
Liked 1,653 Times in 1,208 Posts
.
...there was this one summer at band camp.
3alarmer is offline  
Old 06-22-20, 08:47 AM
  #84  
wphamilton
Senior Member
 
wphamilton's Avatar
 
Join Date: Apr 2011
Location: Alpharetta, GA
Posts: 15,006

Bikes: Nashbar Road

Mentioned: 70 Post(s)
Tagged: 0 Thread(s)
Quoted: 2792 Post(s)
Liked 175 Times in 123 Posts
I have to take a step back to put these numbers in context, before attempting a statistical analysis.
  • Initially reported fatality rates in the US were much higher than globally. The most likely explanation is that confirmed infections were underreported. "Testing" controversy bears this out.
  • In April California began a state-wide expansion of testing facilities and opened up access to testing
  • Immediately the labs were overwhelmed and backlogged, 2 weeks worth of backlogged tests at their maximum capacity.
  • California has been, and is, expanding tests daily.
None of this is conclusive but I submit that it is likely that a broader cross section of people are being tested, as opposed to those having serious symptoms, resulting in a lower apparent fatality rate.
wphamilton is offline  
Old 06-22-20, 09:54 AM
  #85  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
Originally Posted by wphamilton View Post
None of this is conclusive but I submit that it is likely that a broader cross section of people are being tested, as opposed to those having serious symptoms, resulting in a lower apparent fatality rate.
I'd rephrase that 'apparent' in the other direction: initially because of lack of testing cases of infection were underreported, which caused an apparent fatality rate which was higher than truth. Now that we're doing more testing and doing a better job at approaching the infection rate, we are observing the fatality rate dropping closer to truth.
RubeRad is offline  
Old 06-22-20, 10:13 AM
  #86  
wgscott
Occam's Rotor
 
wgscott's Avatar
 
Join Date: Jul 2013
Posts: 6,162
Mentioned: 61 Post(s)
Tagged: 0 Thread(s)
Quoted: 2366 Post(s)
Liked 1,241 Times in 652 Posts
Originally Posted by RubeRad View Post
Hmm. It's hard to tell exactly, since the graphs are so jaggedy. From the May tic-mark to the present, the cases trendline looks to me to be a little less than 2x actually (1800-3400?). For the testing, I guess it looks like 30,000 at the beginning of may to ... uh, I'm not sure what to call now, 65,000? 70,000? That most recent value of 80,000 may be just an anomalous spike.

I wonder if at some point the fatality rate falls because the virus picks off easy targets early on, and later the remaining pool of the population is tilted towards less-at-risk demographics?
I used a robust-fit with a graphing program called Prism. I don't know exactly how it works (it is commercial and closed source) but essentially it gives a max liklihood fit and rejects outliers), and more importantly, I scraped the numbers from the graphs rather than use the primary data, so the error bars are too large.
wgscott is offline  
Old 06-22-20, 10:14 AM
  #87  
wgscott
Occam's Rotor
 
wgscott's Avatar
 
Join Date: Jul 2013
Posts: 6,162
Mentioned: 61 Post(s)
Tagged: 0 Thread(s)
Quoted: 2366 Post(s)
Liked 1,241 Times in 652 Posts
I wonder if it turns out infected people can get reinfected, whether the lingering damage from the first round of infection might increase the pool of at-risk individuals. (Maybe that would explain seemingly healthy young people dropping dead.)
wgscott is offline  
Old 06-22-20, 10:28 AM
  #88  
79pmooney
A Roadie Forever
 
79pmooney's Avatar
 
Join Date: Oct 2014
Location: Portland, OR
Posts: 8,398

Bikes: (2) ti TiCycles, 2007 w/ triple and 2011 fixed, 1979 Peter Mooney, ~1983 Trek 420 now fixed and ~1973 Raleigh Carlton Competition gravel grinder

Mentioned: 99 Post(s)
Tagged: 0 Thread(s)
Quoted: 2279 Post(s)
Likes: 0
Liked 696 Times in 483 Posts
Originally Posted by 3alarmer View Post
...it was in that other forum that I can no longer access, and even if I could, I'd be too lazy to look for it. But I recall being quite surprised by it at the time. If you recall, there was much noise and consternation about the shut down economy, and IIRC, it was in that context. Ironically, the longer this goes on, the more convinced I become that this is probably how the whole thing will play out in the United States. "Teh vulnerable" (whoever they are, and however we figure out how to determine that) are going to end up sucking hind tit in this pandemic.

Everyone will be filled with thoughts and prayers for their premature demise, but hey, they were vulnerable, and should have known they were and taken precautions. The rest of us need to get back to work. That sort of thing. It's pretty easily envisioned. Vulnerability is not something we especially cherish, and it elicits little sympathy in 21st Century America. Things are tough all over.
Other cultures honor their elders. We had elders who have served in the military, seen combat and may well have lifelong health issues from what they saw. (Agent Orange for example.) To some cultures, heros. Here those elders are expendable. Just the vulnerables. Eliminating them brings us closer to herd immunity.

My grandfather telling me what he witnessed in WWI was true gift to me. Shared wisdom. Well worth sucking it up and living a life in a manner where others may also live. Now I keep thinking of the early phrase in the "Declaration of Independence" I have hanging in my living room; "life, liberty and pursuit of happiness". In that order. Life first. Others' lives before my liberty and pursuit of happiness. To me, stepping up and being a man means honoring that. And that means I will not have full liberties until my living situation changes. I have a rent paying housemate who will not live if I being this COVID into the house. I fully expect to have to live a life of limits as long as he lives here.

Ben
79pmooney is offline  
Old 06-22-20, 10:29 AM
  #89  
3alarmer
Friendship is Magic
 
3alarmer's Avatar
 
Join Date: Nov 2010
Location: The Big Tomato
Posts: 18,756

Bikes: old ones

Mentioned: 258 Post(s)
Tagged: 0 Thread(s)
Quoted: 17977 Post(s)
Liked 1,653 Times in 1,208 Posts
Originally Posted by RubeRad View Post
I'd rephrase that 'apparent' in the other direction: initially because of lack of testing cases of infection were underreported, which caused an apparent fatality rate which was higher than truth. Now that we're doing more testing and doing a better job at approaching the infection rate, we are observing the fatality rate dropping closer to truth.
...I don't think it's arguable that a true fatality rate for a novel infectious pandemic disease can be stated accurately except in retrospect. Thus I would never do that. I do find it distressing when people (anyone, yourself included) imply that this apparent reduction is somehow a good thing. This is how all epidemiology works....fatality rates are best viewed in the rear view mirror, with the real data available.

Let's just state that fatality rates for this disease are highly dependent on circumstances, (access to health care, first versus third world, etc.) They have yet to be written out, and will vary. We can say with some degree of certainty, that the death numbers (the people we bury that are in excess of the normal annual curve adjusted for season) are pretty far up into the stratosphere in the USA. If we compare them to other countries where they've done a better and more aggressive job of management according to standard principles of epidemiology, it's embarrassing.

I thought maybe I was incapable of shame at my age, but I'm pretty close as an American when we talk about Covid 19.
3alarmer is offline  
Old 06-22-20, 10:51 AM
  #90  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
Originally Posted by 3alarmer View Post
...I don't think it's arguable that a true fatality rate for a novel infectious pandemic disease can be stated accurately except in retrospect. Thus I would never do that. I do find it distressing when people (anyone, yourself included) imply that this apparent reduction is somehow a good thing.
How is getting closer to the truth not a good thing?

Let's just state that fatality rates for this disease are highly dependent on circumstances, (access to health care, first versus third world, etc.)
Hundo P

We can say with some degree of certainty, that the death numbers (the people we bury that are in excess of the normal annual curve adjusted for season) are pretty far up into the stratosphere in the USA. If we compare them to other countries where they've done a better and more aggressive job of management according to standard principles of epidemiology, it's embarrassing.
Absolutely. Obviously we're doing it wrong
https://www.worldometers.info/corona...y/south-korea/
https://www.worldometers.info/corona...country/japan/
https://www.worldometers.info/corona...y/new-zealand/
https://www.worldometers.info/corona...country/india/
https://www.worldometers.info/corona...try/indonesia/
https://www.worldometers.info/corona...ntry/pakistan/
https://www.worldometers.info/corona...ry/bangladesh/
https://www.worldometers.info/corona...y/philippines/

All of those countries have radically fewer cases/deaths per capita, and most of them (I would guess all but NZ) have much higher population density than the US.

Sort this table by Population (click on right column header), ignore China because they're probably lying their asses off, and look at the distribution of Cases/Deaths per 1M pop of all those high-population countries. Embarrassing is putting it lightly
RubeRad is offline  
Old 06-22-20, 11:01 AM
  #91  
wgscott
Occam's Rotor
 
wgscott's Avatar
 
Join Date: Jul 2013
Posts: 6,162
Mentioned: 61 Post(s)
Tagged: 0 Thread(s)
Quoted: 2366 Post(s)
Liked 1,241 Times in 652 Posts
Originally Posted by RubeRad View Post
ignore China because they're probably lying their asses off,
I don't know a lot about statistics and error analysis, but I am pretty sure that unless you have compelling evidence to reject data, this kind of thing simply introduces expectation bias.
wgscott is offline  
Old 06-22-20, 11:39 AM
  #92  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
https://www.worldometers.info/corona...country/china/

Literally 0 new cases/deaths since Apr 20, plus China having a long history of 1984 Ministry of Truth-style manipulation? I call shenanigans.
RubeRad is offline  
Old 06-22-20, 01:00 PM
  #93  
wgscott
Occam's Rotor
 
wgscott's Avatar
 
Join Date: Jul 2013
Posts: 6,162
Mentioned: 61 Post(s)
Tagged: 0 Thread(s)
Quoted: 2366 Post(s)
Liked 1,241 Times in 652 Posts
Originally Posted by RubeRad View Post
https://www.worldometers.info/corona...country/china/

Literally 0 new cases/deaths since Apr 20, plus China having a long history of 1984 Ministry of Truth-style manipulation? I call shenanigans.
I agree it is a police state.

That doesn't mean what they are reporting now (since April 20) is inaccurate.

Taiwan offers a helpful control, as it is a different government.

https://en.wikipedia.org/wiki/2020_c...reak_in_Taiwan
wgscott is offline  
Old 06-22-20, 01:40 PM
  #94  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
https://www.worldometers.info/corona...ountry/taiwan/

Taiwan reports a trickle of daily new cases,which is entirely credible (and admirable for a crowded nation of 23M). For China, source of the outbreak, pop ~1.4B to have literally 0 new deaths or even cases for 2 solid months, is beyond incredible.
RubeRad is offline  
Old 06-22-20, 02:48 PM
  #95  
skookum
cyclotourist
 
Join Date: Jul 2002
Location: calgary, canada
Posts: 1,164
Mentioned: 2 Post(s)
Tagged: 0 Thread(s)
Quoted: 314 Post(s)
Liked 115 Times in 70 Posts
Beijing reported a cluster of new cases last week.
On Thursday, Beijing reported 21 new cases of Covid-19, down slightly from the 31 reported on Wednesday. There were two additional cases in Hebei province that were also linked to the Beijing cluster.

The city on Thursday ordered all hotels be shut down, as well as restaurants in high-risk areas. Officials said Beijing was not under lockdown but urged all residents not to travel or gather unnecessarily, and pledged to ensure continued food supply.
https://www.msn.com/en-gb/news/world...al/ar-BB15OVN3


Taiwan has controlled the virus well, partly because they have been extremely skeptical of information coming out of China and suspended flights from Wuhan in January.

China did a good job of locking down and suppressing the virus, but I would not believe their numbers.
skookum is offline  
Likes For skookum:
Old 06-22-20, 03:32 PM
  #96  
3alarmer
Friendship is Magic
 
3alarmer's Avatar
 
Join Date: Nov 2010
Location: The Big Tomato
Posts: 18,756

Bikes: old ones

Mentioned: 258 Post(s)
Tagged: 0 Thread(s)
Quoted: 17977 Post(s)
Liked 1,653 Times in 1,208 Posts
Originally Posted by RubeRad View Post
How is getting closer to the truth not a good thing?
...I think my point was that , calculation of mortality/fatality rates for a novel disease, while probably a valuable exercise in dictating which ones are gonna be really bad, the really bad ship has sailed.


Originally Posted by RubeRad View Post
Absolutely. Obviously we're doing it wrong
https://www.worldometers.info/corona...y/south-korea/
https://www.worldometers.info/corona...country/japan/
https://www.worldometers.info/corona...y/new-zealand/
https://www.worldometers.info/corona...country/india/
https://www.worldometers.info/corona...try/indonesia/
https://www.worldometers.info/corona...ntry/pakistan/
https://www.worldometers.info/corona...ry/bangladesh/
https://www.worldometers.info/corona...y/philippines/

All of those countries have radically fewer cases/deaths per capita, and most of them (I would guess all but NZ) have much higher population density than the US.

Sort this table by Population (click on right column header), ignore China because they're probably lying their asses off, and look at the distribution of Cases/Deaths per 1M pop of all those high-population countries. Embarrassing is putting it lightly
...there was an interesting thing on the radio while I was driving down to Marin County today. Some public health policy wonk or another from the state of California mentioned that, while it is true we are doing a lot more testing, none of it can be construed as "random testing". The people getting tested (because there's still a shortage of materials and backup at the labs), are people either symptomatic or people exposed to somone with the disease as a contact. So any surmise based on the principles of random testing in a population to determine exposures is based on a fallacious assumption.

I'll just let that sit here for a while.

Last edited by 3alarmer; 06-22-20 at 03:37 PM.
3alarmer is offline  
Likes For 3alarmer:
Old 06-22-20, 05:19 PM
  #97  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
Originally Posted by skookum View Post
Beijing reported a cluster of new cases last week.

https://www.msn.com/en-gb/news/world...al/ar-BB15OVN3


Taiwan has controlled the virus well, partly because they have been extremely skeptical of information coming out of China and suspended flights from Wuhan in January.

China did a good job of locking down and suppressing the virus, but I would not believe their numbers.
OK, so if you go to this page of China's statistics
https://www.worldometers.info/corona...country/china/

And find the Daily New Cases graph, and hover the mouse very carefully, you can get a popup for Jun 14 or so reflecting that cluster.

AND if you check the box to turn on either the moving 3-day or 7-day average, then you can hover anywhere on the curve and see that, in the blank-looking space since Apr 21, there are daily totals of like 2 or 3, which occupies no pixels at the scale of that graph.

So I was wrong, China is not reporting 'literally 0' cases, they're reporting practically 0. The Daily Deaths graph does actually report Literally 0 deaths since Apr 21.

I still don't believe 'em.
RubeRad is offline  
Old 06-22-20, 05:38 PM
  #98  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
Originally Posted by wgscott View Post
I used a robust-fit with a graphing program called Prism. I don't know exactly how it works (it is commercial and closed source) but essentially it gives a max liklihood fit and rejects outliers), and more importantly, I scraped the numbers from the graphs rather than use the primary data, so the error bars are too large.
That sounds pretty good. If you really want to dig, the LA Times maintains their raw data on github:
https://github.com/datadesk/california-coronavirus-data
RubeRad is offline  
Likes For RubeRad:
Old 06-22-20, 06:03 PM
  #99  
RubeRad
Keepin it Wheel
Thread Starter
 
RubeRad's Avatar
 
Join Date: Aug 2011
Location: San Diego
Posts: 9,217

Bikes: Surly CrossCheck, Moto Fantom29 ProSL hardtail

Mentioned: 0 Post(s)
Tagged: 0 Thread(s)
Quoted: 20 Post(s)
Liked 579 Times in 432 Posts
Originally Posted by 3alarmer View Post
while it is true we are doing a lot more testing, none of it can be construed as "random testing". The people getting tested (because there's still a shortage of materials and backup at the labs), are people either symptomatic or people exposed to somone with the disease as a contact. So any surmise based on the principles of random testing in a population to determine exposures is based on a fallacious assumption.
Yes, this means that cases are significantly undercounted, because they don't include asymptomatic infections and probably some smaller amount of people who suspect their symptoms are covid19, but because they are mild, they ride them out in isolation and don't report them.

Back to my OP, as it is much harder to have an 'asymptomatic death', or a death that goes unnoticed/unreported completely, the death statistics are much less subject to underreporting. Overreporting deaths would involve either counting somebody dead from covid19 who actually died from something else, or counting somebody dead who isn't actually dead. Both of those are quite unlikely.

So the fatality rate is: significantly accurate number of deaths, divided by significantly underreported number of infections. Thus the fatality rate is always an overestimate. More testing will bring the apparent/measured/observed fatality rate down closer to the truth, but as noted the only way to get a true fatality rate (with statistically rigorous confidence/error bars) is widespread truly random sampling of the population, regardless of symptoms or self-selection for testing.

The best window into this are special cases where we are able to test complete, controlled populations, like in prisons, and on cruise ships.

The Diamond Princess had 3711 total passengers+crew, 712 of them tested infected (331 without symptoms), and 13 died.

That's a perfectly known fatality rate for the population on that ship of 13/712 = 1.826%, That fatality rate would be different than the overall fatality rate for the general public, based on how different the demographics are on the cruise ship. Unless this was some special kind of booze cruise, this demographic probably tilts old. So if you're too old/sick to go on cruises, you're probably looking at a higher risk than 1.826%. If you're younger/stronger than the average cruise ship customer, you're probably looking at a lower risk.

In Lompoc, over 1000 prisoners tested positive. Searching Google, it seems that 4 of those prisoners died, the 4th death reported Jun 1. Since it's been 3 weeks since then, I presume the deaths are done and all other infected prisoners recovered. For that contained case, 4/1000 is 0.4%. Prisoner demographics are maybe younger(?) and stronger(?) than the general public, so I would expect the general public overall fatality rate to be larger than 0.4%.
RubeRad is offline  
Old 06-22-20, 06:25 PM
  #100  
wgscott
Occam's Rotor
 
wgscott's Avatar
 
Join Date: Jul 2013
Posts: 6,162
Mentioned: 61 Post(s)
Tagged: 0 Thread(s)
Quoted: 2366 Post(s)
Liked 1,241 Times in 652 Posts
I do worry that the emphasis is not always on the most important statistics.

The MERS-CoV fatality rate is about 30%. However, SARS-CoV-2 has proven to be far more lethal (whether that fatality rate turns out to be 5%, or 10 times smaller).

Also, is there any reason to trust the Bureau of Prisons more than you do the Chinese Government?
wgscott is offline  

Thread Tools
Search this Thread

Contact Us - Archive - Advertising - Cookie Policy - Privacy Statement - Terms of Service - Do Not Sell My Personal Information

Copyright 2018 MH Sub I, LLC dba Internet Brands. All rights reserved. Use of this site indicates your consent to the Terms of Use.