1539 replies to this topic

[gamesguru]

You're assuming that we even know all the variants in circulation.  That's rather important when you're extrapolating genomic shifts back to a starting point.  Or that we actually know the error rate when these viruses copy.  I'm going to go out on a limb here and suggest that we don't.  That's what we've seen in the past.  Subsequent analysis will modify that November estimate. By the way, do you know the error bars/confidence interval on that estimate?  Since you didn't know that was where that November start date came from, I'm guessing not.

 

Time and time again I see this.  Models that agree with someone's conclusion are to be treated as gospel.  Models that don't are junk.

 

Where specifically have we seen it in the past?  Flu phylogenetic estimates often predate as well, so the opposite trend is also common in modern tests.  We sometimes push the epidemic forward rather than back.

 

The estimates are not based on phylogeneitc estimates, rather epidemiological predictions.  I read the study, yes?  Where are you getting this, I'm reading otherwise??

These are the simple facts that follow assuming each person becomes infective at 7 days and infects on average—prior to mitigation—3 people.

Methods: We developed a simple disease-transmission model in which the 2019-nCoV epidemic was modeled as a branching process starting in mid-November 2019, with a serial interval of 7 days (time between cases) and a basic reproduction number (R₀) of 2.3 (new cases from each old case), based on available data and assuming no intervention (Figure 1). The epidemic start date aligned our modeled case counts to point estimates from international case exportation data (4). The model estimated plausible values of the effective reproduction number (R_e; reproduction number in the presence of control efforts) after implementation of a quarantine in Wuhan and surrounding areas of China on 24 January 2020 (3) (Figure 1).

 

 

But you were touting the predictions of the Oxford model at one point (must I really go back and quote the posts?).

 

If you knew these were flawed "18th century models" why were you using them? Surely not because they supported your position.

 

Come on, let's at least be intellectually honest here.  I have admitted that my estimates of when this started are my opinion based on history. 

Please, quote it.  Where did I say the US would follow the original IHME model?

[gamesguru]

A overview from February from Italian researchers,

Early phylogenetic estimate of the effective reproduction number of SARS‐CoV‐2 Abstract

To reconstruct the evolutionary dynamics of the 2019 novel‐coronavirus recently causing an outbreak in Wuhan, China, 52 SARS‐CoV‐2 genomes available on 4 February 2020 at Global Initiative on Sharing All Influenza Data were analyzed. The two models used to estimate the reproduction number (coalescent‐based exponential growth and a birth‐death skyline method) indicated an estimated mean evolutionary rate of 7.8 × 10⁻⁴ subs/site/year (range, 1.1 × 10⁻⁴‐15 × 10⁻⁴) and a mean tMRCA of the tree root of 73 days. The estimated R value was 2.6 (range, 2.1‐5.1), and increased from 0.8 to 2.4 in December 2019. The estimated mean doubling time of the epidemic was between 3.6 and 4.1 days. This study proves the usefulness of phylogeny in supporting the surveillance of emerging new infections even as the epidemic is growing.

Highlights

• The aim of this study was to investigate the temporal origin, rate of viral evolution and population dynamics of SARS‐CoV‐2.

• The Bayesian approach used indicated a most probable origin of the epidemic between October and November 2019.

• The estimated effective reproductive number increased from 0.8 to 2.4 in December 2019 when the mean doubling time was about 4 days.

• This study proves the usefulness of phylogeny in supporting the surveillance of emerging new infections.

[Daniel Cooper]

Please, quote it.  Where did I say the US would follow the original IHME model?

 
From the covid policy thread:
 

Yes well i doubt the US wants to maintain that even for a few years.

The question was how far they are willing to go to save 2 million lives. We'll see.

There are other threads where you state that in the US "2 million lives" are at stake.

I know where that figure of "2 million" came from. Do you?
 

[gamesguru]

 I know where that figure of "2 million" came from. Do you?

 

Pretty much came from the idea that it would kill 1% if it infected everyone.  But what exactly are you trying to argue here, that I referenced a 200 yr old model and therefore no models should be trusted anymore?

[Daniel Cooper]

Pretty much came from the idea that it would kill 1% if it infected everyone.  But what exactly are you trying to argue here, that I referenced a 200 yr old model and therefore no models should be trusted anymore?

 

 

Not at all.  Models are the only thing we have to predict the future.  But we need to understand that a model predicting 2 million deaths isn't the same as there actually being 2 million deaths.  It's a prediction. The outcome of the model is built into the assumptions, which happened to be wrong in that case.  Models are nothing more than humans taking their best guess at how a system works, coding that into software, and running it to produce the output.  The validity of the output is entirely dependent on the quality of the guesses that went into the front end.

 

My point is that you aren't picking your models based on your assessment of the validity of the assumptions those models are built on. In fact neither you nor most other people are generally aware of those assumptions because it is often the case that the details of the model are not even disclosed in academic papers.  But rather you are picking models based on whether they tend to support or refute the conclusion you're trying to make at any given moment.  That's a poor basis for model selection.

[gamesguru]

My point is that you aren't picking your models based on your assessment of the validity of the assumptions those models are built on.

 

Ok well that's slightly judgmental then.  It really assumes I haven't examined the limitations of the IHME model vs. the U of T model and I haven't come to a reasonable conclusion of my own.

 

Again, let's try to be as critical of the models informing your views as the ones informing mine.  Oh wait, you haven't produced any models yet, or any data points that would definitively suggest the U of T model is off.. you've just negatively criticized the idea as a whole

[Daniel Cooper]

No, I've told you several times now that my estimate of an earlier date for the arrival of covid-19 in the human population is based on understanding the history of prior pandemics and some deductive analysis of the timeline that China has given us.

 

None the less, my opinion is looking more correct than the initial official story that covid arrived on the scene around the 1st week in December.

 

But, I freely admit, it's just my opinion.  There is no computer model that I can point to to back me up.

 

 

[gamesguru]

No, I've told you several times now that my estimate of an earlier date for the arrival of covid-19 in the human population is based on understanding the history of prior pandemics and some deductive analysis of the timeline that China has given us.

 

So what is the prior pandemics and timeline of China's data doing to inform you of something so different from the widely accepted models?

[Daniel Cooper]

So what is the prior pandemics and timeline of China's data doing to inform you of something so different from the widely accepted models?

 

Well, it's very common in pandemics in general to see the estimate of when the first cases arrived being pushed back in time.  As we've discussed before, this has happened in the case of HIV (estimates now go back to the end of the 19th century for when SIV crossed over to humans) and in the case of the 1918 pandemic where you'll see initial human infection dates from 1915 to the very early 1910s and even earlier.  And so it is for most other new human pandemics.

 

As far a China goes, my reasoning goes something like this:

 

1.) Q: China's offical timeline is Patient 0 gets sick around December 10th, arrives in a Wuhan hospital on December 16th, the Chinese medical community figures out they have a novel and dangerous virus after the 16th and informs the WHO on the 31st (2 weeks).  Is this plausible?

 

A: No, this is not plausible.  It would very likely take longer to determine they had a novel virus and it would almost certainly take their political leadership longer than 2 weeks to inform the WHO given the likely damage to the Chinese economy.  The political leadership would wait until they were certain that the virus had already escaped China and that it would be certain to be traced back there.  My conclusion is that China was certain that the virus had already escaped China sometime before December 31st when they informed the WHO.

 

2.) Q: If they lied about the timeline, why?

 

A: Clearly the truth would put China in a bad light.

 

3.) Q: Then what are the likely possibilities for the truth?

 

A:

a.) They were aware of the virus much earlier such that they would be accused of sitting on the information for an unreasonable amount of time before informing the world.

b.) They indeed found they had a novel virus very quickly because they had reason to believe that a virus under study had escaped their lab in Wuhan and were on the lookout for cases showing up in hospital.

 

I chose to believe that a.) was more likely, but b.) could be as well.  Note that if b.) I am not suggesting this was an engineered virus or some sort of bioweapon. I don't believe that.

 

4.) Q: If a.) is true how long would they have sat on the information in order to fear that the true timeline would put them in a bad light?

 

A: I don't think anyone would have faulted the Chinese for taking a month to 6 weeks to figure out that they had a new virus on their hands. This is complicated time consuming stuff and you generally don't go looking for new viruses until you have a number of unexplained patient deaths. Therefore they would have had to have known earlier than the first of November.  Therefore, my guess is that the Chinese knew they had a new virus running around no later than the end of October. I'd guess mid-October just because. It certainly could have been earlier.

 

You asked, and this was my reasoning. Now, I'm certain that someone will label this as "ill informed", but I'm pretty sure I'm well informed about my own thought process.   But of course you are free to disagree with it.

[gamesguru]

What about France too are they lying?  Allegedly no deaths there before Feb 25, and no cases prior to Dec 27.

 

Based on a serial interval of 7 days and an R0 = 3, we see the French have the same curve as Wuhan just delayed 6 weeks.  So they would reach the 10,000 case criticality around mid-March.. which is exactly what the worldometer histogram confirms.  In order to get the initial infection pushed back in France before November, some befuddling of the figures would be necessary.

[Daniel Cooper]

France's first known case being December 27th is entirely consistent with my proposed timeline.

 

And no one but the Chinese government knows what Wuhan's curve looks like.  I certainly have no confidence that they've given us the entire story.  Do you?

 

 

Edited by Daniel Cooper, 11 June 2020 - 11:39 PM.

[gamesguru]

Thank you.

 

At some point someone suggested the epidemic began in Wuhan in August or September.  But it doesn't seem anyone is maintaining that anymore.  The idea it started this early and spread so fast is pure farce.  It is wholly inconsistent with timelines seen, for example, in France and the San Francisco Bay area.

[Evawatson]

After relaxing restrictions and easing lockdown, Iran is beginning to see a second wave of coronavirus infections.

 

 

OHH.  It is Scary. we have to strictly follow the rule wear mask,  social distance, Stay Home. Use hand sanitizer, Rest Rapid Test Kit for Coronavirus is doing its own work and still helping the doctors.  We have to break the chain guys. our doctors can take rest from tireless job.   the year 2020 is a really cruel year.  Nobody will want to remember this year.  and I strictly boycott china. I hate.

Edited by Evawatson, 12 June 2020 - 03:50 AM.

[Kalliste]

 

 

However, the usual antibody tests look in the blood. There, the scientists found clearly detectable amounts of IgG only in the severely ill, which also occurred at an early stage. “The current antibody tests do not cover all cases by a long way,” says study director Boyman. Therefore, it can be assumed that five times more people than are detected in broad-based antibody tests have already had contact with the new coronavirus. This is because less than one fifth of all infected people fall seriously ill and consequently have clearly detectable antibodies in their blood.

https://swprs.org/co...-of-infections/

[Daniel Cooper]

Thank you.

 

At some point someone suggested the epidemic began in Wuhan in August or September.  But it doesn't seem anyone is maintaining that anymore.  The idea it started this early and spread so fast is pure farce.  It is wholly inconsistent with timelines seen, for example, in France and the San Francisco Bay area.

 

My timeline is my opinion on when the Chinese political leadership knew they had a new virus circulating around.  I don't think they could possibly have known later than mid October and feel confident that when they informed the WHO they were already certain that the virus had escaped China.  They would have known this through contact tracing that turned up exposed persons that had left the country and exposed Chinese citizens that had come into contact with Westerners who has subsequently left the country.  The Chinese leadership would not have informed the WHO until they were certain that covid was outside of China and it was likely to be traced back to Wuhan.  I feel reasonably confident about this because the timeline they have released is pure fantasy and there has to be a reason for that.

 

Now, when did the virus enter the human population?  My opinion - sometime prior to mid October. August - September is plausible and dates even earlier.  I just don't know but earlier would not surprise me, but it could have entered the human population as late as early October.  

 

The history for these sort of things is for these estimates of when a new virus entered the human population to be pushed backwards as more information is developed.  But, it's possible that they'll find it crossed over late September/early October.  If I were a betting man I would not bet on it, but it's certainly possible. 

 

It will be interesting to watch this develop as we learn more about the virus.

[gamesguru]

Now, when did the virus enter the human population?  My opinion - sometime prior to mid October. August - September is plausible and dates even earlier.

 

Have to disagree that this is impossible based on what we know, and exceedingly improbable based on what we might.  With a latent period < 7 days and an R0 = 3 (prior to public awareness and in the height of flu season), it is impossible.  If COVID had seeded in Wuhan prior to mid November, things would have exploded into a million cases by January at that rate.  The detection occurred at the decided threshold of ~50k cases, which is basically a few months after doctor's started to ring the bell.  Kind of how it gripped France and NYC in late December, and those effects were undeniably felt by March.  The idea of the "unmitigated case" doesn't exist because the doctors become aware of sick people in the area, word gets out, and people start acting at least somewhat cautious.

 

I'm also operating under the assumption that it reached other major nations no more than 4-6 weeks after genesis in Wuhan.  So if it started in mid November, it landed in France and USA by January.  If it started all the way back in August, it had to land in France by mid-September and we just have no evidence anyone was sick in France prior to Dec 27, and the timeline for the phylogenetic tree supports this view.  If it started in August, we would have seem clear epicenters erupting round the world in October.  This disproves the original assumption by contradiction, and places the burden of proof on the camp believing things started so early.

 

I would need to see strong reasoning against this concrete logic of latent period, R0, and exponential curves exploding if we push things back to August.  Til then it's just rampant speculation held together on a whim.

Edited by gamesguru, 12 June 2020 - 10:40 PM.

[pamojja]

With a latent period < 7 days and an R0 = 3 (prior to public awareness and in the height of flu season), it is impossible.

 

I would need to see strong reasoning against this concrete logic of latent period, R0, and exponential curves exploding if we push things back to August.

 

Sure, if one persistently ignores that from now about 200 countries with infections for months there exists also 30 countries with exactly zero death, 40 countries with up to 10, about 6 countries up to 100 deaths, 40 countries up to 1000, and the vast majority with a much much milder death-toll than from regular influenza and pneumonia seasons.

 

And only takes into consideration the 19 odd countries countries, where it indeed came with a death-toll as with any very strong influenza season.

[gamesguru]

Sure, if one persistently ignores that from now about 200 countries with infections for months there exists also 30 countries with exactly zero death, 40 countries with up to 10, about 6 countries up to 100 deaths, 40 countries up to 1000, and the vast majority with a much much milder death-toll than from regular influenza and pneumonia seasons.

 

And only takes into consideration the 19 odd countries countries, where it indeed came with a death-toll as with any very strong influenza season.

 

No coincidence most of those countries are the ones failing to take it seriously.  New Zealand and Croatia have less than 20 active cases but out of humility are refusing to call the pandemic over, stating that premature celebrations don't help us put the real problem behind us.

 

The countries with new outbreaks are being driven by stubborn mindsets of people without masks and indulging in their liberties despite the very consistent threat of the virus to continue its spread.

[pamojja]

No coincidence most of those countries are the ones failing to take it seriously.  New Zealand and Croatia have less than 20 active cases but out of humility are refusing to call the pandemic over, stating that premature celebrations don't help us put the real problem behind us.

 
Again as usual, you're only cherry-picking 2-3 countries which confirm your bias. Without any relevance to your initial argument (ie: With a latent period < 7 days and an R0 = 3)

 

All the 30 countries with zero death which might take it serious - or not - with in total more than 2200 infected: Uganda, Vietnam, Mongolia, Faeroe Islands, Gibraltar, Cambodia, French Polynesia, Bhutan, Macao, Eritrea, Namibia, Saint Vincent, Timor-Leste, Grenada, New Caledonia, Laos, Saint Lucia, Dominica, Fiji, Saint Kitts Nevis, Falkland Islands, Greenland, Vatican City, Seychelles, Papua New Guinea, Caribbean Netherlands, Saint Barth, Lesotho, Anguilla, Saint Pierre Miquelon
 
All 48 countries with up to to 10 deaths (with above 21,000 infected, 220 deaths total): Iceland, Jamaica, Mauritius, Zambia, Madagascar, Montenegro, Jordan, Bermuda, Malta, Tinidad Tobago, Maldives, Taiwan, Barbados, Central Africa, Cabo Verde, Angola, Syria, Myanmar, Libya, Benin, US Virgin Islands, Malawi, Hongkong, Zimbabwe, Monaco, Antigua Barbuda, Palestine, Saint Martin, Suriname, Aruba, Eswatini, Brunei, Belize, French Guiana, Comoros, Mozambique, Rwanda, Cayman Islands, Curaçao, Gambia, Botswana, Lichtenstein, Turks Caicos, Burundi, British Virgin Islands, Monserrat, Reunion, Western Sahara
 
All 62 countries with up to 100 deaths (with above 280,000 infected, about 2,450 deaths total): Kenya, Oman, Somalia, Cuba, Mauritania, Lithuania, Kazakhstan, El Salvador, Chad, Qatar, Estonia, Niger, Haiti, Senegal, Thailand, Nicaragua, Ethiopia, Burkina Faso, Andorra, Sierra Leone, Tajikistan, Tunisia, Ghana, Channel Islands, Côte d'Ivoire, San Marino, Djibouti, Bahrain, Albania, Lebanon, Kosovo, Liberia, Slovakia, Mayotte, Latvia, Kyrgyzstan, South Sudan, Guinea, Singapore, Isle of Man, Congo, Gabon, Venezuela, Uruguay, New Zealand, Tanzania, Uzbekistan, Cyprus, Nepal, Sint Maarten, Guinea Bissau, Georgia, Martinique, Guadeloupe, Togo, Guyana, Costa Rica, Equatorial Guinea, Sao Tome Principe, Paraguay, Sri Lanka, Bahamas
 
All 44 countries with up to 1000 deaths (with above 680,000 infected, about 15,600 deaths): Saudi Arabia, Japan, Ukraine, Argentina, Algeria, Austria, Denmark, Dominican Republic, Hungary, Bolivia, Afghanistan, Iraq, Sudan, Panama, Nigeria,, Moldova, Guatemala, Czechia, Finland, Honduras, Belarus, Israel, Kuwait, UAE, South Korea, Armenia, Serbia, Norway, Morocco, Cameroon, Greece, Mazedomia North, Bulgaria, Bosnia Herzegovina, Yemen, Puerto Rico, Malaysia, Azerbaijan, Luxembourg, Slovenia, Croatia, Congo DR, Mali, Australia

All 19 countries, affecting a mere 12% of the world population, which indeed have a unusual death-toll compared to the 2017 average influenza/pneumonia mortality: USA, Brazil, UK, Italy, France, Spain, Belgium, Canada, Sweden, Turkey, Ecuador, Switzerland, Ireland, Andorra, Channel Islands, San Marino, Isle of Man, Sint Maarten, Monserrat

 

 

I would need to see strong reasoning against this concrete logic of latent period, R0, and exponential curves exploding if we push things back to August.  Til then it's just rampant speculation held together on a whim.

 

The usual behavior of this virus in 88% of the world population isn't just rampant speculation or on a whim. It is rather the rule. Therefore all it needs for this virus to be around a couple of months or years earlier without being recognized, is its default behavior.


 

 

[gamesguru]

OK well it doesn't prove anything looking at it that far away.  Andorra for example has a population of only 77,000 so ranking it based on total deaths in a group with Thailand makes no sense.  In fact all of the countries listed are either very small (population < 100,000), or they are showing an excellent response.  Less than half of people are asymptomatic, and at least 30-40% of people with symptoms are being diagnosed in most places.

[pamojja]

Irrelevant to your utterly speculative claim not seen in the vast majority of the world-population - ie: With a latent period < 7 days and an R0 = 3

[gamesguru]

Wrong it is your twisted narrative which is utterly irrelevant to the experience felt in the majority of ill-prepared nations compared against prepared ones.  Among less proactive nations, practically only Australia and New Zealand stand out as nations not implementing strict mask and social distancing rules while still having things well under control.  If you know a nation with more than 10 million people that enforced absolutely zero social distancing and has less than 50 deaths per million, please point me in that direction.  The usual sophistry of quoting 10 articles and writing essay posts does nothing but delay the inevitable defeat.

[pamojja]

Again, completely irrelevant to your fantasy claims of a fixed RO. Or any articles - but the real hart numbers available to everyone.

 

But right away Malawi comes to mind, where a court overruled the implementation by its president. With a population 21 million. 5 deaths within 1 1/2 months.
 

Edited by pamojja, 14 June 2020 - 01:29 PM.

[gamesguru]

Never claimed a fixed R0, it varied in Germany according to the lockdown.  In early stages Wuhan, it was 3.  In Australia seems to be closer to 1, perhaps due to seasonal factors.

 

Malawi is one country in an equatorial climate, with poor testing capacity and probably more deaths than reported.  The idea they didn't social distance is also an exaggeration, as many large events were canceled and the people at large were gripped by fear.  Likely the climate and lack of testing also explain their numbers.

 

India has also practically no deaths, but there to there is an explanation, they were seeded late (being foes of the Chinese) and at a time late in the flu season, they also implement aggressive lockdown in response to early outbreaks in less mobile nation, but despite all this have poor testing capacity and likely nowhere near the true numbers.

[pamojja]

You clearly claimed an R0 = 3. To counter Daniel*s argument, that an earlier outbreak than the officially admitted practically being impossible:

 

Have to disagree that this is impossible based on what we know, and exceedingly improbable based on what we might.  With a latent period < 7 days and an R0 = 3 (prior to public awareness and in the height of flu season), it is impossible.  If COVID had seeded in Wuhan prior to mid November, things would have exploded into a million cases by January at that rate.  The detection occurred at the decided threshold of ~50k cases, which is basically a few months after doctor's started to ring the bell.  Kind of how it gripped France and NYC in late December, and those effects were undeniably felt by March.  The idea of the "unmitigated case" doesn't exist because the doctors become aware of sick people in the area, word gets out, and people start acting at least somewhat cautious.

 

I'm also operating under the assumption that it reached other major nations no more than 4-6 weeks after genesis in Wuhan.  So if it started in mid November, it landed in France and USA by January.  If it started all the way back in August, it had to land in France by mid-September and we just have no evidence anyone was sick in France prior to Dec 27, and the timeline for the phylogenetic tree supports this view.  If it started in August, we would have seem clear epicenters erupting round the world in October.  This disproves the original assumption by contradiction, and places the burden of proof on the camp believing things started so early.

 

I would need to see strong reasoning against this concrete logic of latent period, R0, and exponential curves exploding if we push things back to August.  Til then it's just rampant speculation held together on a whim.

 

Without it, you're really without any counterargument. It could have been in Wuhan without causing many deaths since many month, just as it does in the majority of the world population (88%).

Edited by pamojja, 14 June 2020 - 01:47 PM.

[gamesguru]

Wrong, I claimed it was R0 at the start of the epidemic when Wuhan had no idea (and even the citizens) that a new disease was spreading.  Such conditions of non-distancing are ideal for the virus spread.  Quickly people caught on, and in other areas, for example a graph from Germany showing the time-dependent nature of the R0 value.

 

 

As such under typical conditions (as existed in Wuhan prior to December) it is widely assumed to have an R0 value of at least 2.5 with a latent period not exceeding 7 days, and as such mandating a genesis of no earlier than late October - mid November.

[pamojja]

As such under typical conditions (as existed in Wuhan prior to December) it is widely assumed to have an R0 value of at least 2.5 with a latent period not exceeding 7 days, and as such mandating a genesis of no earlier than late October - mid November.

 

Your 'typical condition' is a ideological construct very rare in real life. Even for a very minority of countries it is very short-lived. The remaining countries, adding up to 88% of the world population, didn't experience.

Edited by pamojja, 14 June 2020 - 02:11 PM.

[gamesguru]

Your 'typical condition' is a ideological construct very rare in real life. Even for a very minority of countries it is very short-lived. The remaining countries, adding up to 88% of the world population, didn't experience.

 

Regrets to consider other factors: population density, seasonality and climate, culture attitudes, and lack of testing all must be considered.

 

Several places are ideal storms.. US has bad cultural attitude combined with ideal climate and good population density.  Similar of Brazil.

[pamojja]

US has bad cultural attitude combined with ideal climate and good population density.  Similar of Brazil.

 

US by now is at 2.39 times the average incidence of influenzia/pneumonia mortality. While Brazil at 1.35 times til now still within completely normal fluctuations.
 

[gamesguru]

Brazil also have practically no deaths til April, e.g. seeded later.  They also have poor testing, and are a Southern Hemisphere nation so maybe are just rearing up for flu season.  Granted that, the 1.35x factor seems reasonable and may become worse.