Coronavirus outbreak

1101 posts
User avatar
From Despair To Where?
Hob Nob Addict
Posts: 9878
Joined: 19 Apr 2004 08:37
Location: Goin' up around the bend.

Re: Coronavirus outbreak

by From Despair To Where? » 08 Apr 2020 15:58

A simple yes or no would have sufficed. So what you're saying is it's a completely unreliable statistic of no analytical worth.


Bye..

Snowball
Hob Nob Super-Addict
Posts: 16838
Joined: 02 Jan 2009 18:35

Re: Coronavirus outbreak

by Snowball » 08 Apr 2020 16:12

The Coronavirus takes approximately 14 days to kill

Keeping it simple for the moment, 14 days ago (24-March) there were 339,752 infections (excluding China) and the deaths 14 days later were (excl China) were 79,831 = 23.5% of Confirmed Cases. It is the cases BACK THEN that create the bulk of the deaths TODAY.

Maybe it would be better to look at say, TEN days ago (28th March).
You get a Cases Since Then Number of 581,643, and for the same number of deaths you get a rate just under 14%


What MATTERS though is how many NEW Cases there are
How many cases have come after the cases which led to 79,831 Dead

0,339,752 Total New Cases Since Day-14 (24 Mar)
0,581,643 Total New Cases Since Day -10 (28 Mar)
1,424,341 Total Cases…………………………… 7th April


If the Day-14 is correct that means that SINCE the Cases which caused yesterday's deaths were recorded a further 1,084,589 Cases have been recorded. (x 3.19).

This more than a million cases will produce more deaths than the total deaths up to today.

If 14 Days is the right number that would mean that in Apr-21st the death toll outside China will be 79,831+254,661 = 327,307 Deaths.

On the 21st April we would reach 327,307 Deaths worldwide ( 7 Apr x 4.19) (Excluding China) - Based on 14 Days Case Confirmation to Death

You think a multiple of 4.19 is silly? In the14 Days from May-24th the deaths multiplied by over 5 times


Based on 10 Days average from Confirmation to Death

On 21st April we should reach 195,586 Deaths (Excluding China - Based on 14 Days Case Confirmation to Death

If Day 10 is a better day to base the algorithm on, that means that SINCE the Cases which caused yesterday's deaths were recorded a further 842,698 Cases have been recorded. (x1.45 Cases).

If 10 Days is the right number that would mean that in Apr-21st the death toll outside China will be 115,755 + 79831 = 195,586 Deaths

You think a multiple of 2.45 is silly? In the14 Days from May-24th deaths virtually doubled.

Hound
Hob Nob Super-Addict
Posts: 13943
Joined: 27 Sep 2016 22:16
Location: Simpleton

Re: Coronavirus outbreak

by Hound » 08 Apr 2020 16:37

If you haven't lost everyone already, you certainly have now

Snowball
Hob Nob Super-Addict
Posts: 16838
Joined: 02 Jan 2009 18:35

Re: Coronavirus outbreak

by Snowball » 08 Apr 2020 17:46

Hound If you haven't lost everyone already, you certainly have now



That won't mean I'm wrong.

Emmer Green Royal
Member
Posts: 240
Joined: 05 May 2006 11:57
Location: St Johann in Tirol, Austria

Re: Coronavirus outbreak

by Emmer Green Royal » 09 Apr 2020 11:17

From the Guardian today:

For the “Covid-19 case cluster study”, scientists from the University of Bonn went back to the town that had the first two fatalities from the virus in Germany and interviewed and tested 1,000 residents. Researchers are also trying to work out exactly how the virus got transmitted at a carnival event in the area on 15 February.

After analysing around half of the tests, the study’s director, Prof Hendrick Streeck, said on Thursday morning that 14% of the population in the area had developed immunity after contracting the coronavirus. Previous estimated had put the infection rate at only around 5%.

Streeck said the fatality case rate of the virus in the area had also turned out to be considerably lower than the currently currently registered for the country as a whole. In Heinsberg, only 0.37% of people who contracted the virus had died.

The latest figure for Germany as a whole, as calculated by Johns Hopkins University, is 1.98%.


SCIAG
Hob Nob Regular
Posts: 4794
Joined: 17 Jun 2008 17:43
Location: Liburd for England

Re: Coronavirus outbreak

by SCIAG » 09 Apr 2020 11:28

Snowball
Hound If you haven't lost everyone already, you certainly have now



That won't mean I'm wrong.

There’s no point being right if nobody can understand what you’re going on about.

User avatar
Wycombe Royal
Hob Nob Addict
Posts: 6148
Joined: 14 Apr 2004 19:31
Location: The posh part of Winnersh

Re: Coronavirus outbreak

by Wycombe Royal » 09 Apr 2020 12:34

SCIAG
Snowball
Hound If you haven't lost everyone already, you certainly have now



That won't mean I'm wrong.

There’s no point being right if nobody can understand what you’re going on about.

That's just an invitation for him to do more analysis of the stats........ :D

Snowball
Hob Nob Super-Addict
Posts: 16838
Joined: 02 Jan 2009 18:35

Re: Coronavirus outbreak

by Snowball » 09 Apr 2020 13:00

SCIAG
Snowball
Hound If you haven't lost everyone already, you certainly have now



That won't mean I'm wrong.

There’s no point being right if nobody can understand what you’re going on about.


True. It's quite simple. You can predict today's cumulative deaths with very high accuracy by looking at the cumulative CASES 10-14 days ago because the death rate starts to average out.

You can take yesterday's resolved cases at 23:59 (417,955) and go back to find when the cumulative cases were 417,955 (or a close approximation). That takes us back 15.1 Days to 21:40 on 24-March. The Cumulative Cases then were 417,895 (only 60 different)

The Death rate for that is 21.14% and for the five days prior

20.72%
21.00%
21.14%
21.33%
21.14%

You can predict today's deaths and cumulative deaths by going back more or less 15 days and multiplying the cases THEN by the overall rate NOW and it comes out very, very close to the actual.

The back-date though, appeared to be 12.5 days and has crept up for the last twelve days. I think that means we have been better at keeping people alive for longer (learning on the job). But I think many just hang on longer. The actual Death Rate for competed cases has slowly but definitely crept up. That would make sense. If we keep people (who eventually die) alive for 15 days when it used to be 12.5, the later days will have a few longer-term patients. Alternatively the pressure in ICU's is increasing and it's getting harder to give maximum care.

There is, BTW, a sign that the death rate for Completed Cases is FALLING now from a maximum of 21.55% to almost exactly 21%. That might seem trivial but 0.55% of the current outstanding cases (1,103,012) is a saving of 6,667 deaths against previous expectations.

The model predicts a cumulative deaths at 23:59 tonight of 94,949+ based on Cumulative Cases at 17:00 25-March of 453,008, which would mean 6,615 deaths today. That's using the base from 15.2 Days ago, not looking at the last couple of Days and marking up.

The last three days deaths have averaged 6,354 deaths per day, so 6,615 looks in the ball park

This is kind of a circular argument because I could get a death rate tonight at 23:59 based on Resolved Cases - Recovered = Deaths. Then, by going back in the data to find a figure that equates to today's resolved cases, it looks like a fudge. Except that the actual number used is the cumulative cases number from 15 days ago. The logic is simple. The cumulative deaths total today is derived ONLY from all the resolved cases. The total still active is just noise. Ignore that million+ they give the deaths in 15 days time (ish)

So, sadly we should get to about 95,000 deaths today BUT the total deaths in 15 Days time will be 305,928

Snowball
Hob Nob Super-Addict
Posts: 16838
Joined: 02 Jan 2009 18:35

Re: Coronavirus outbreak

by Snowball » 09 Apr 2020 13:01

Posting this separately for clarity


That comes from today's cumulative total (so far) of 1,529,640 times 20% which is a rounded figure for the death rate for resolved cases.

In 15 Days, 98% of those 1,529,640 Cumulative Cases will have resolved (got better or died).


ATM the figure is 4:1 for Resolved Cases, so we get the figures for April 24th of 305,928 Cumulative Dead

The UK on that basis will be over 12,000 April 24th (12,146+)


I used a ball-park 20%. The rate is atm 21%

EDIT: Light-Bulb Moment!

I couldn't understand why the death-rate appeared to steadily climb. I just realised that at the start of this data collection (12-Mar) China utterly dominated the totals (First day of records China were 80,000 cases out of a world 137,000 = 58%) and the influence of its numbers diminish every day. (They are now just 5% of world totals. They have been almost static, only increasing by 1,000 cases or so in all this time.

So I removed China from the figures. These are the figures for World-China. They are the Deaths from Completed Cases

Date - - Resolved Better - Dead Death-Rate

25-Mar 053,983 036,044 17,939 33.23%
26-Mar 066,448 045,627 20,821 31.33%
27-Mar 079,261 055,243 24,018 30.30%
28-Mar 091,790 064,182 27,608 30.08%
29-Mar 103,826 073,155 30,671 29.54%
30-Mar 121,359 086,990 34,369 28.32%
31-Mar 138,034 099,202 38,832 28.13%
01-Apr 159,745 116,293 43,452 27.20%
02-Apr 183,947 134,333 49,614 26.97%
03-Apr 206,168 150,664 55,504 26.92%
04-Apr 229,702 168,370 61,332 26.70%
05-Apr 249,546 183,592 65,954 26.43%
06-Apr 271,740 200,500 71,240 26.22%
07-Apr 302,666 224,088 78,578 25.96%
08-Apr 336,955 251,939 85,016 25.23%
09-Apr 345,978 259,614 86,364 24.96%

So Completed Cases are ending up 3:1, 3 recovered, 1 Dead.

That 25% death-rate is, of course, only for confirmed cases.

If the true infection rate is 10 times that than the death rate for the population is 2.5% , twenty-five times seasonal flu.

If the true infection rate is 100 times confirmed cases then the death rate is 0.25, two-and-a-half times seasonal flu.

If the true infection rate is 80x then the population death rate (for those who get it, obvs) is .313%

If it's true that 60% have to be infected for herd immunity, at 80:1 it looks like it would cost 64,000 dead, at 100:1 51,000 dead)

100% 212,193
090% 190,974
080% 152,779
070% 106,945
060% 064,167 (51,000 at 100:1) <<<<<<
050% 032,084


User avatar
From Despair To Where?
Hob Nob Addict
Posts: 9878
Joined: 19 Apr 2004 08:37
Location: Goin' up around the bend.

Re: Coronavirus outbreak

by From Despair To Where? » 09 Apr 2020 14:45

...and that's


Hound
Hob Nob Super-Addict
Posts: 13943
Joined: 27 Sep 2016 22:16
Location: Simpleton

Re: Coronavirus outbreak

by Hound » 09 Apr 2020 15:19

bit of a break from Snowball's stats-athon

Talk that the league is aiming to restart in June and play all the games in a 56 day period (inc play offs). Players would return to training in mid may

Not quite sure how that then feeds into the next season though - maybe a couple of weeks off and then all start again. Esp with the Euros in summer 2021

Mind you, I dont quite know how footy will restart at all if they are going to lockdown every time there is a bunch of new cases and we are strongly advised not to join mass gatherings

URZZZZ
Hob Nob Regular
Posts: 3600
Joined: 20 Apr 2013 18:30

Re: Coronavirus outbreak

by URZZZZ » 09 Apr 2020 15:41

Hound bit of a break from Snowball's stats-athon

Talk that the league is aiming to restart in June and play all the games in a 56 day period (inc play offs). Players would return to training in mid may

Not quite sure how that then feeds into the next season though - maybe a couple of weeks off and then all start again. Esp with the Euros in summer 2021

Mind you, I dont quite know how footy will restart at all if they are going to lockdown every time there is a bunch of new cases and we are strongly advised not to join mass gatherings


There’s a few things wrong with this though:

(A) - When do players get a break? Not talking about the Corona break, an actual one, if they jump into the new season two weeks after the old one

(B) - would make the transfer window/contracts extremely difficult given a two week window

(C) - would be very difficult to set out fixtures for an entire season in a two week window, unless they released them batch by batch but usually people would like an idea of when fixtures are going to be played

Appreciate it’s a difficult situation for everyone involved and there are no right/wrong answers but I would say the players need a good 6 weeks off after the final game to properly recuperate

If that means more midweek games next year/eliminate a cup competition/delay next season too etc etc then so be it, can’t be helped

Would suggest they’ll just play behind closed doors to finish the season off so mass gatherings won’t really be a problem, although that’s just an estimate

Hound
Hob Nob Super-Addict
Posts: 13943
Joined: 27 Sep 2016 22:16
Location: Simpleton

Re: Coronavirus outbreak

by Hound » 09 Apr 2020 15:43

oh yeah. It seems a very high level set of guidelines

I suspect matches this season behind closed doors but streamed over ifollow/Sky

no idea about break in season. Be crap to write September off which is a great footballing month but something has to give

still lots about what happens if a player is diagnosed with it. Does the whole team have to isolate etc?


Snowball
Hob Nob Super-Addict
Posts: 16838
Joined: 02 Jan 2009 18:35

Re: Coronavirus outbreak

by Snowball » 09 Apr 2020 17:57

A few people seem not to understand that the death rate give (say, UK now 11%) is always UNDER the final rate at the end of a pandemic

Here is s a simple way to see what happens.

Imagine a virus which arrives and only infects people for a month. People start to die after 14 Days, and anyone who dies, (exactly 20% of everyone infected) dies exactly 14 days later, In the first 13 days, despite a huge growth in cases there will be no deaths at all. Then, for another 14 or so days the rate will climb. Then the rate will fall and the final rate will be the one I stated at the start... exactly 20%

Exactly 20% of infected people die after exactly 14 days

Death
RATE
01-Apr > 00100 > 000,100
02-Apr > 00150 > 000,250
03-Apr > 00200 > 000,450
04-Apr > 00270 > 000,720
05-Apr > 00350 > 001,070
06-Apr > 00400 > 001,470
07-Apr > 00530 > 002,000
08-Apr > 00670 > 002,670
09-Apr > 00800 > 003,470
10-Apr > 00900 > 004,370
11-Apr > 01000 > 005,370
12-Apr > 01200 > 006,570
13-Apr > 01300 > 007,870
14-Apr > 02000 > 009,870
=======================================================================================
First 14 Days show a death rate of ZERO!

Then, or 14 Days the apparent death rate (crude death rate) keeps climbing. It means nothing
=======================================================================================


15-Apr > 02500 > 012,370 > 0,020 > 00,020 > 00.80% <<<< First time deaths acknowledged
16-Apr > 02500 > 014,870 > 0,030 > 00,050 > 02.00%
17-Apr > 03000 > 017,870 > 0,040 > 00,090 > 03.00%
18-Apr > 03500 > 021,370 > 0,054 > 00,144 > 04.11%
19-Apr > 04000 > 025,370 > 0,070 > 00,214 > 05.35%
20-Apr > 05000 > 030,370 > 0,080 > 00,294 > 05.88%
21-Apr > 06000 > 036,370 > 0,106 > 00,400 > 06.67%
22-Apr > 07000 > 043,370 > 0,134 > 00,534 > 07.63%
23-Apr > 08000 > 051,370 > 0,160 > 00,694 > 08.68%
24-Apr > 09000 > 060,370 > 0,180 > 00,874 > 09.71%
25-Apr > 10000 > 070,370 > 0,200 > 01,074 > 10.74%
26-Apr > 10000 > 080,370 > 0,240 > 01,314 > 13.14%
27-Apr > 10000 > 090,370 > 0,260 > 01,574 > 15.74%
28-Apr > 10000 > 100,370 > 0,400 > 01,974 > 19.74%

29-Apr > 10000 > 110,370 > 0,500 > 02,474 > 24.74%
30-Apr > 10000 > 120,370 > 0,500 > 02,974 > 29.74% <<<

====================================================================================

For a couple of days around the final infection date , the rate is higher than real

INFECTIONS CEASE, DEATHS CONTINUE after 14 DAYS

The apparent death-rate falls steadily until it reaches exactly 20%

=====================================================================================

01-May > 00000 > 000,000 > 0,600 > 03,574 > 24.03%
02-May > 00000 000,000 >> 0,700 > 04,274 > 23.92%
03-May > 00000 > 000,000 > 0,800 > 05,074 > 23.74%
04-May > 00000 > 000,000 > 1,000 > 06,074 > 23.94%
05-May > 00000 > 000,000 > 1,200 > 07,274 > 23.95%
06-May > 00000 > 000,000 > 1,400 > 08,674 > 23.85%
07-May > 00000 > 000,000 > 1,600 > 10,274 > 23.69%
08-May > 00000 > 000,000 > 1,800 > 12,074 > 23.50%
09-May > 00000 > 000,000 > 2,000 > 14,074 > 23.31%
10-May > 00000 > 000,000 > 2,000 > 16,074 > 22.84%
11-May > 00000 > 000,000 > 2,000 > 18,074 > 22.49%
12-May > 00000 > 000,000 > 2,000 > 20,074 > 22.21%
13-May > 00000 > 000,000 > 2,000 > 22,074 > 21.99%
14-May > 00000 > 000,000 > 2,000 > 24,074 > 21.81%

15-May > 00000 > 000,000 > 0,000 > 24,074 > 20.00% <<<<<<<<<<<<<<<<

Snowball
Hob Nob Super-Addict
Posts: 16838
Joined: 02 Jan 2009 18:35

Re: Coronavirus outbreak

by Snowball » 09 Apr 2020 18:08

At tonight's Govt briefing Dominic Rabb gave these stats

65,077 Cases
16,784 Hospitalised
07,978 Died in Hospital
01,500 Critical
(this figure not yet given for today, so approximate based on yesterday)

This shows that a staggering 47.53% of hospital cases are dying.

When the critical cases and those lingering are added, that must surely go over 50%

12.26% of Confirmed Cases are shown to be dying. This is a crude rate, much under-played) and will rise to around 25% of CC's

Emmer Green Royal
Member
Posts: 240
Joined: 05 May 2006 11:57
Location: St Johann in Tirol, Austria

Re: Coronavirus outbreak

by Emmer Green Royal » 09 Apr 2020 21:17

Snowball
Exactly 20% of infected people die after exactly 14 days



That is TOTAL BOLLOCKS as you have NO IDEA how many people are infected.

Snowball
Hob Nob Super-Addict
Posts: 16838
Joined: 02 Jan 2009 18:35

Re: Coronavirus outbreak

by Snowball » 10 Apr 2020 01:07

Snowball


You can predict today's deaths and cumulative deaths by going back more or less 15 days and multiplying the cases THEN by the overall rate NOW and it comes out very, very close to the actual.


The model predicts a cumulative deaths at 23:59 tonight of 94,949+ based on Cumulative Cases at 17:00 25-March of 453,008, which would mean 6,615 deaths today. That's using the base from 15.2 Days ago, not looking at the last couple of Days and marking up.



The prediction was actually 97,764 and the actual was 95,657

==================================================

The TV people are going on about the UK today showing a tiny fall from 888 to 881, but Wales (41 today) because it changed the way it returned its numbers went from 24 to 12 hours. Pro-rata they would have returned 82 and the UK total would have been 923.
Last edited by Snowball on 10 Apr 2020 01:16, edited 1 time in total.

Snowball
Hob Nob Super-Addict
Posts: 16838
Joined: 02 Jan 2009 18:35

Re: Coronavirus outbreak

by Snowball » 10 Apr 2020 01:08

Emmer Green Royal
Snowball
Exactly 20% of infected people die after exactly 14 days



That is TOTAL BOLLOCKS as you have NO IDEA how many people are infected.



It was an EXAMPLE for Chrissakes, NOT THIS pandemic. Dear God!

See here? Imagine a virus which arrives and only infects people for a month.

People start to die after 14 Days, and anyone who dies, (exactly 20% of everyone infected) dies exactly 14 days later, In the first 13 days, despite a huge growth in cases there will be no deaths at all. Then, for another 14 or so days the rate will climb. Then the rate will fall and the final rate will be the one I stated at the start... exactly 20%

User avatar
bcubed
Hob Nob Addict
Posts: 5195
Joined: 30 Oct 2004 18:16
Location: Would do better with a stick of rhubarb

Re: Coronavirus outbreak

by bcubed » 10 Apr 2020 09:26

Emmer Green Royal From the Guardian today:

For the “Covid-19 case cluster study”, scientists from the University of Bonn went back to the town that had the first two fatalities from the virus in Germany and interviewed and tested 1,000 residents. Researchers are also trying to work out exactly how the virus got transmitted at a carnival event in the area on 15 February.

After analysing around half of the tests, the study’s director, Prof Hendrick Streeck, said on Thursday morning that 14% of the population in the area had developed immunity after contracting the coronavirus. Previous estimated had put the infection rate at only around 5%.

Streeck said the fatality case rate of the virus in the area had also turned out to be considerably lower than the currently currently registered for the country as a whole. In Heinsberg, only 0.37% of people who contracted the virus had died.

The latest figure for Germany as a whole, as calculated by Johns Hopkins University, is 1.98%.


That's interesting
And I understood it as well

Hound
Hob Nob Super-Addict
Posts: 13943
Joined: 27 Sep 2016 22:16
Location: Simpleton

Re: Coronavirus outbreak

by Hound » 10 Apr 2020 09:29

bcubed
Emmer Green Royal From the Guardian today:

For the “Covid-19 case cluster study”, scientists from the University of Bonn went back to the town that had the first two fatalities from the virus in Germany and interviewed and tested 1,000 residents. Researchers are also trying to work out exactly how the virus got transmitted at a carnival event in the area on 15 February.

After analysing around half of the tests, the study’s director, Prof Hendrick Streeck, said on Thursday morning that 14% of the population in the area had developed immunity after contracting the coronavirus. Previous estimated had put the infection rate at only around 5%.

Streeck said the fatality case rate of the virus in the area had also turned out to be considerably lower than the currently currently registered for the country as a whole. In Heinsberg, only 0.37% of people who contracted the virus had died.

The latest figure for Germany as a whole, as calculated by Johns Hopkins University, is 1.98%.


That's interesting
And I understood it as well


It is. And the work going on there is very important to find a way out of where we are

1101 posts

Who is online

Users browsing this forum: Basildon, Za Vas and 55 guests

It is currently 27 May 2020 09:51