clutch

Play around with a model yourself: https://www.josstam.com/post/lockdown-blues

There goes that lone example from the “do nothing” crowd. But what do you expect from people who lack any sense of objectivity and scientific literacy? Often wrong, never in doubt. I'd revisit the deaths/mil number in 16 months and it should be similar across most of the countries. The area under the flattened and peaked curves are about the same...

A bit of cherry-picking. In the same thread: Nostromo26 "Why did you stop plotting deaths on the 20th? Daily deaths haven't gone down since then. Here's a chart updated through 4/25." But regardless, this model predicts that the number of deaths would be close to 0 by July... I guess we will start counting deaths like China at that point.

This is a noteworthy data point that I didn't keep up on--I thought it was always 6, not 14. Thanks for the update, Spekulatius . Well it’s a small dataset and shouldn’t be taken for face value it rather as a ballpark idea one what the distribution may look like. Once we get more data coming in, we might revise. Softly the ~1% IFR rate is a reasonable ballpark I tighten 0.5% or 15-, but it’s unlikely to be 0.1% or 5% at this point. None of this is set in stone. Doctors will get better at fighting this as we get a better understanding of how the disease works and how to prevent the downward spiral that seems to affect some patients. The crew was substantial and none of them died. Regardless, you can estimate the IFR by the age groups in the ship, and if you just consider the passenger population (which would reflect the sub-population the median age of 69) -- you get an IFR of 2.5% (14/567 infected). If it's 2.5% IFR for the median age of 69 in the general population, that's really good news, considering the ratio of deaths between such a population group vs. its complement. Based on the data so far, it seems the death rate for 50 years or older is about 25-30x higher than the 49 or younger on average. (source: https://www.worldometers.info/coronavirus/coronavirus-age-sex-demographics/) So if you assume 2.5% IFR for the 50 years or order, the IFR for younger would be about 0.083-0.1%! That is a very, very low number. More of these data indicate that we should focus on protecting the elderly. (I know this wasn't your original point; just thought it was useful exercise)

I love the confidence! Even though you are so consistently proven wrong, you still post with gusto! What is the Infection Fatality Rate of the flu? Nobody actually believes it is 0.67%, do they? This comprehensive review shows ~10 deaths per 100,000 H1N1infections: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3809029/ So comparing your bogus numbers for CV to these bogus numbers for H1N1, CV is 67 times more deadly than the flu! I love the stupidity! the 21% positive antibody test results on NYCers (3000 person sample) is the best DATUM we have on creating the correct denominator of the mortality per infection rate. what is your problem, Larkin? https://www.statnews.com/2018/09/26/cdc-us-flu-deaths-winter/ I agree that this is a meaningful study. It’s large enough and reasonably random. There might be a bias in just choosing people outside vs at home but it is likely not a strong one. 0.67% morbidity is ~7x deadlier than the flu though. The flu kills between 10-50k annually and infects ~30M (roughly ) so thats in the 0.1% ballpark. In addition, it’s much more infective. On thing I overlooked when looking at the IFR rate is that death cases have a long tail. They typically occur many weeks after the infection and display of symptoms Example of this was the Diamond Princess when only 6 death were reported first, but subsequently ended up with 14 dead. While that is a small sample size, I think the likely conclusion is that the IFR rate is higher than the ~0.7% rate calculated probably by as much than a factor 2 It’s all highly uncertain at this point and ballpark estimates, but better than nothing. https://en.wikipedia.org/wiki/2020_coronavirus_pandemic_on_Diamond_Princess ...and normalize with respect to the age. The median age of the passengers was 69. I believe all of the deaths were passengers. So, if you assume IFR of ~1.4% for that old-age population group, the IFR for the entire population would be much, much lower.

You guys are really fucked in US if this is the typical attitude of a non Trump supporter. Start a civil war, already.

You cannot project what has been observed in those countries. The deaths/million figure will change over time. If we have no vaccine, that number will become similar across the developed world over time, perhaps normalized with respect to the population age distribution. Maybe some countries have better treatments and that will help. However, those countries with low cases/deaths will eventually catch up, because it seems impossible to irradicate this virus via containment.

Okay, this is the last time I will correct you for a few weeks. The experts prediction of both mild/asymptomatic cases and IFR are accurate. Maybe it is your lack of expertise that is the issue (for example not knowing the difference between IFR and CFR)? The original model presented by the White House estimated 1.5M-2.2M deaths, if there was no mitigation. Guess how many deaths there would be assuming your 0.67% "mortality rate" and an attack rate of 70% (rough estimate of infection rate needed for herd immunity)? 328M * 0.7 * .0067 = 1.5M Seems pretty accurate to me. Maybe the experts know more about this pandemic than Cerzeca? Something doesn't seem right though, so help me understand. The same White House model projected that 100,000 and 240,000 would die with the mitigation measures in place. Are they speculating that there will be a vaccine available before the US reaches the attack rate? Another possible explanation -- If you flatten the curve, you might be able to avoid some deaths due to the healthcare system overflow, but I expect that the area under the curve would still be similar (compared to when you have a spike). So are they saying that they would save an order of magnitude of lives by keeping the health care system below the threshold? Part of my problem with these projections is that there is no transparency or explanation whatsoever... The reason for the discrepancy is that in the white house models and what we see is that in those models they assumed a 50% compliance with the shelter in place/whatever you may call them orders. What was observed is that the compliance rate is around 90%. So big miss on the assumption there. You basically have the numbers coming below the model because the quarantine is working way better because the people are behaving much, much better than assumed. They are taking it more seriously than the government assumed and that is making the quarantine much more effective. No, that doesn't address my question. I'm not questioning the discrepancy between the projection and actual data here (although that is also important). I'm curious how the experts came to project a much lower number of deaths with the mitigation in place. KCLarkin suggested that the projection of 1.5M-2.2M deaths was based on: 328M * 0.7 * .0067 = 1.5M assuming that people die until the US reaches the herd immunity, without any mitigation measures. But how is the number reduced to 100,000-240,000 with the mitigation measures? Does this mean we do not reach the herd immunity in this case? Then, why would the death number stop at 100,000-240,000? Are they banking on a potential cure or vaccine? If so, have they indicated this at any time? There are other factors that are modeled in mitigation other than compliance such as reasonable ramp up of testing and tracing to bring the overall R0 down during and after the shelter in place is slowly removed. Clearly experts were wrong on how dysfunctional the policy response will be even after 2 months of this. Those measures would slow down the infection rate, not the total number of deaths. Unless you get a vaccine or cure, the number of deaths would simply get stretched out over time. The area under the flattened curve can still the same as a sharp one. So the projections must have taken account some scenario that the virus would be mostly irradicated. Or not? None of the journalists ever asked this question?? Maybe the projections are time bound? What is the time period then? I think a lot of people should pay attention to other than Faux news and do a proper DD into what is actually being done and reported. These model (one by white house and other) have been very well reported. And they get updated as new data comes in - http://covid19.healthdata.org/united-states-of-america http://www.healthdata.org/covid/data-downloads https://www.imperial.ac.uk/media/imperial-college/medicine/sph/ide/gida-fellowships/Imperial-College-COVID19-NPI-modelling-16-03-2020.pdf Well, I live in Canada and don't watch Fox. And not sure what this question has anything to do with it. The first two resources you share show data and not any of the assumptions/parameters used in the models. So they don't answer my question. I did find an answer in the paper you shared: "The major challenge of suppression is that this type of intensive intervention package – or something equivalently effective at reducing transmission – will need to be maintained until a vaccine becomes available (potentially 18 months or more) – given that we predict that transmission will quickly rebound if interventions are relaxed." If this is the assumption behind the government's projections, does the public know? You can't expect people to look up and read these scientific papers. Also, say we keep on with this intervention approach for 18 months -- do people realize what the total death numbers would be like IF there was no vaccine? In the end, it will be pretty much similar to the number as the case without the intervention (minus the number of deaths due to healthcare system overload). That is because the number of people who get infected and die will still be the same, just stretched over a much longer period -- the area under the two curves are about the same. All this while the entire economy is shut down. So really, this intervention approach is in there to buy us time. Hoping that a vaccine is developed. It's a reasonable approach given that we could potentially save more lives this way than just letting everyone gets infected within a short period of time. But I never hear any government official or press communicates this important assumption... which leads me to believe that most of the public do not know what the "flatten the curve" strategy really entails.

Okay, this is the last time I will correct you for a few weeks. The experts prediction of both mild/asymptomatic cases and IFR are accurate. Maybe it is your lack of expertise that is the issue (for example not knowing the difference between IFR and CFR)? The original model presented by the White House estimated 1.5M-2.2M deaths, if there was no mitigation. Guess how many deaths there would be assuming your 0.67% "mortality rate" and an attack rate of 70% (rough estimate of infection rate needed for herd immunity)? 328M * 0.7 * .0067 = 1.5M Seems pretty accurate to me. Maybe the experts know more about this pandemic than Cerzeca? Something doesn't seem right though, so help me understand. The same White House model projected that 100,000 and 240,000 would die with the mitigation measures in place. Are they speculating that there will be a vaccine available before the US reaches the attack rate? Another possible explanation -- If you flatten the curve, you might be able to avoid some deaths due to the healthcare system overflow, but I expect that the area under the curve would still be similar (compared to when you have a spike). So are they saying that they would save an order of magnitude of lives by keeping the health care system below the threshold? Part of my problem with these projections is that there is no transparency or explanation whatsoever... The reason for the discrepancy is that in the white house models and what we see is that in those models they assumed a 50% compliance with the shelter in place/whatever you may call them orders. What was observed is that the compliance rate is around 90%. So big miss on the assumption there. You basically have the numbers coming below the model because the quarantine is working way better because the people are behaving much, much better than assumed. They are taking it more seriously than the government assumed and that is making the quarantine much more effective. No, that doesn't address my question. I'm not questioning the discrepancy between the projection and actual data here (although that is also important). I'm curious how the experts came to project a much lower number of deaths with the mitigation in place. KCLarkin suggested that the projection of 1.5M-2.2M deaths was based on: 328M * 0.7 * .0067 = 1.5M assuming that people die until the US reaches the herd immunity, without any mitigation measures. But how is the number reduced to 100,000-240,000 with the mitigation measures? Does this mean we do not reach the herd immunity in this case? Then, why would the death number stop at 100,000-240,000? Are they banking on a potential cure or vaccine? If so, have they indicated this at any time? There are other factors that are modeled in mitigation other than compliance such as reasonable ramp up of testing and tracing to bring the overall R0 down during and after the shelter in place is slowly removed. Clearly experts were wrong on how dysfunctional the policy response will be even after 2 months of this. Those measures would slow down the infection rate, not the total number of deaths. Unless you get a vaccine or cure, the number of deaths would simply get stretched out over time. The area under the flattened curve can still the same as a sharp one. So the projections must have taken account some scenario that the virus would be mostly irradicated. Or not? None of the journalists ever asked this question?? Maybe the projections are time bound? What is the time period then?

No, this is horrible! The estimated IFR has been 0.65-1% since at least early March. 0.67%* is devastatingly high. If we want herd immunity (say 70% infected), that would be over 1.5M deaths in the U.S. Yeah, this. My best guess today based on what I've been reading is that it's below 0.5%. Like maybe 0.3%. The argument you need make, Cherzera, is that it's 0.5% at most, and we don't really care about the people who die because most of them were going to die in the near future anyway, while the people who die from the economic impact/shutting down society are more likely to be young. That's a pretty defensible position, I think. And if you really care about equally saving lives, across the world: https://www.cnn.com/2020/04/22/africa/coronavirus-famine-un-warning-intl/index.html

Okay, this is the last time I will correct you for a few weeks. The experts prediction of both mild/asymptomatic cases and IFR are accurate. Maybe it is your lack of expertise that is the issue (for example not knowing the difference between IFR and CFR)? The original model presented by the White House estimated 1.5M-2.2M deaths, if there was no mitigation. Guess how many deaths there would be assuming your 0.67% "mortality rate" and an attack rate of 70% (rough estimate of infection rate needed for herd immunity)? 328M * 0.7 * .0067 = 1.5M Seems pretty accurate to me. Maybe the experts know more about this pandemic than Cerzeca? Something doesn't seem right though, so help me understand. The same White House model projected that 100,000 and 240,000 would die with the mitigation measures in place. Are they speculating that there will be a vaccine available before the US reaches the attack rate? Another possible explanation -- If you flatten the curve, you might be able to avoid some deaths due to the healthcare system overflow, but I expect that the area under the curve would still be similar (compared to when you have a spike). So are they saying that they would save an order of magnitude of lives by keeping the health care system below the threshold? Part of my problem with these projections is that there is no transparency or explanation whatsoever... The reason for the discrepancy is that in the white house models and what we see is that in those models they assumed a 50% compliance with the shelter in place/whatever you may call them orders. What was observed is that the compliance rate is around 90%. So big miss on the assumption there. You basically have the numbers coming below the model because the quarantine is working way better because the people are behaving much, much better than assumed. They are taking it more seriously than the government assumed and that is making the quarantine much more effective. No, that doesn't address my question. I'm not questioning the discrepancy between the projection and actual data here (although that is also important). I'm curious how the experts came to project a much lower number of deaths with the mitigation in place. KCLarkin suggested that the projection of 1.5M-2.2M deaths was based on: 328M * 0.7 * .0067 = 1.5M assuming that people die until the US reaches the herd immunity, without any mitigation measures. But how is the number reduced to 100,000-240,000 with the mitigation measures? Does this mean we do not reach the herd immunity in this case? Then, why would the death number stop at 100,000-240,000? Are they banking on a potential cure or vaccine? If so, have they indicated this at any time?

I hope these antibody tests are conducted across the country (countries) so we can get a much clearer picture and these speculative debates are no longer needed.

Okay, this is the last time I will correct you for a few weeks. The experts prediction of both mild/asymptomatic cases and IFR are accurate. Maybe it is your lack of expertise that is the issue (for example not knowing the difference between IFR and CFR)? The original model presented by the White House estimated 1.5M-2.2M deaths, if there was no mitigation. Guess how many deaths there would be assuming your 0.67% "mortality rate" and an attack rate of 70% (rough estimate of infection rate needed for herd immunity)? 328M * 0.7 * .0067 = 1.5M Seems pretty accurate to me. Maybe the experts know more about this pandemic than Cerzeca? Something doesn't seem right though, so help me understand. The same White House model projected that 100,000 and 240,000 would die with the mitigation measures in place. Are they speculating that there will be a vaccine available before the US reaches the attack rate? Another possible explanation -- If you flatten the curve, you might be able to avoid some deaths due to the healthcare system overflow, but I expect that the area under the curve would still be similar (compared to when you have a spike). So are they saying that they would save an order of magnitude of lives by keeping the health care system below the threshold? Part of my problem with these projections is that there is no transparency or explanation whatsoever...

Nobody ever had a 3-4% IFR estimate. You are conflating CFR with IFR. The estimated IFR was always around 0.65%. My bad. I'd still take these study results as a good sign than a bad one. I'd rather have fewer people die over the period of this pandemic, even if only by fewer percentages.

No, this is horrible! The estimated IFR has been 0.65-1% since at least early March. 0.67%* is devastatingly high. If we want herd immunity (say 70% infected), that would be over 1.5M deaths in the U.S. I'd say that's better than 3-4%.

What is your denominator for those 80,000 regular flu deaths? You make a back-of-the-envelope estimate for COVID IFR and then announce that it is "just like the flu" without having the intellectual curiosity to actually calculate the IFR for "just the flu"? No expert thinks that "the flu" has an IFR anywhere near 0.67%*. The estimates I've seen are around 0.1%. But if you factor in asymptomatic cases, it is likely even lower. https://www.cdc.gov/flu/about/burden/past-seasons.html So Covid is at least 7 times more deadly than "just the flu". -- * even if your Covid denominator is right, your numerator is wrong due to the lag between infection and death. Either way, COVID-19 seems much less deadly than expected, perhaps an order of magnitude, which is a good thing. Would you agree?

Seriously, we should hope that these antibody testing results are accurate. It means that fewer people will die than expected over the course of the pandemic and there is a good chance that we will reach herd immunity before any vaccine is produced. But when you are political, any findings that go against your original narrative is bad news.

Great. This is how scientific experts should be critiquing each other. At least this testing method was transparent, but how come prediction models are never shared with the public so that others could review and critique them?

I sold my VOO (S&P500) puts as we seem to be getting a more clear picture of this pandemic...

Hindsight is 20-20, but my thoughts on what as happened: - Experts were looking at what was happening in Italy and updated their model parameters with the metrics from Italy (e.g., lethality) for the "worst-case" scenario. - Experts advised the governments to prepare the hospitals and resources according to the worst-case scenario. - It turns out, while COVID-19 is highly contagious (due to its ability to spread via asymptomatic people), the number of fatal cases was not as high as expected. Health care systems in many places were rarely under a strain. - Meanwhile, the obvious fact that COVID-19's lethality rate was disproportionately higher among older people was largely overlooked. Most governments did not take enough preemptive measures to protect those in long-term care homes (and sometimes counter-productive actions as in NY). In Europe and Canada, half of the number of COVID-19's death have come from long-term care facilities.

A humanitarian reason for opening up the economies around the world: "Coronavirus pandemic will cause global famines of 'biblical proportions,' UN warns"\ "While dealing with a Covid-19 pandemic, we are also on the brink of a hunger pandemic," David Beasley told the UN's security council. "There is also a real danger that more people could potentially die from the economic impact of Covid-19 than from the virus itself." https://www.cnn.com/2020/04/22/africa/coronavirus-famine-un-warning-intl/index.html Yes, it is from CNN, not Fox.

'Up to half' of Europe deaths in care homes, WHO warns https://www.bbc.com/news/live/world-52391597

FYI, guys, I was talking about the situation in Ontario, Canada... not US or Trump... I guess nobody cares about Canada and I should just leave this thread alone. ;D

Yeah, models aren't perfect, but I don't really see an alternative. Is there one? Make decisions based on the most pertinent observations? E.g., the fact that COVID-19 has significantly higher death rates in older people. In general, scientists should inform decision-makers primarily based on observations/experiments, not prediction models. People are getting delusional regarding these models... thinking that they can accurately predict the future if the models are built by scientists. The only thing that is delusional is to believe that people who have spent all their lives in this area (doctors in the area of infectious diseases, diagnostic, vaccine and drug development experts, and yes, epidemiologist) make decisions solely based on models and not update their hypothesis based on hard data on the ground. Fair point -- if you assume that these scientists were all rational, just like how economists would assume. ;)

Yeah, models aren't perfect, but I don't really see an alternative. Is there one? Make decisions based on the most pertinent observations? E.g., the fact that COVID-19 has significantly higher death rates in older people. In general, scientists should inform decision-makers primarily based on observations/experiments, not prediction models. Easy to say, when the matter at hand is basically unknown from the beginning of the situation. It was well-known for this particular case, which was at the end of March.