arcube Posted April 8, 2020 Posted April 8, 2020 May be Japan and Singapore will start shutdown. India started shutdown on March 22nd. See the numbers since March 22nd. https://www.worldometers.info/coronavirus/country/india/ As opposed to? As opposed to before March 22nd. Isnt the shutdown supposed to reduce cases and deaths, at least after 2 weeks? That will make sense if the same amount of testing exists before and after. The shut down (one you are questioning) is supposed to flatten the curve and that will take time as testing ramps up. Not sure how aware are you about India. The country could be literally decimated if this shut down was not placed. It is one third the size of U.S. with 4x+ the population. It would have taken decades to recover. India paid a huge price for swine flu a few years ago. These are tough choices, not a menu of great choices to be exact. It is easy to shout in a forum that the know nothing epidemiologists are basically nerds with models, its another thing to be one where for years you are trained to make these choices from a menu of fking hard options. Deaths are usually tested correctly. Deaths also went up after 2 weeks of shutdown. How long is it supposed to take for this flatening to happen and why? You just summarized the gist of this whole damn thread. I will let you ponder over that one.
Investor20 Posted April 8, 2020 Posted April 8, 2020 May be Japan and Singapore will start shutdown. India started shutdown on March 22nd. See the numbers since March 22nd. https://www.worldometers.info/coronavirus/country/india/ As opposed to? As opposed to before March 22nd. Isnt the shutdown supposed to reduce cases and deaths, at least after 2 weeks? That will make sense if the same amount of testing exists before and after. The shut down (one you are questioning) is supposed to flatten the curve and that will take time as testing ramps up. Not sure how aware are you about India. The country could be literally decimated if this shut down was not placed. It is one third the size of U.S. with 4x+ the population. It would have taken decades to recover. India paid a huge price for swine flu a few years ago. These are tough choices, not a menu of great choices to be exact. It is easy to shout in a forum that the know nothing epidemiologists are basically nerds with models, its another thing to be one where for years you are trained to make these choices from a menu of fking hard options. Deaths are usually tested correctly. Deaths also went up after 2 weeks of shutdown. How long is it supposed to take for this flatening to happen and why? You just summarized the gist of this whole damn thread. I will let you ponder over that one. So its sure millions will loose jobs. Not sure why and when this flattening will occur with a shutdown. While many countries in South East Asia went atleast till now without shutdown managing low deaths. The option is not clear?
Liberty Posted April 8, 2020 Posted April 8, 2020 How do you interpret 0 cases, 0 deaths in every country? It's possible that's a stand-in for a trivial number or cases before burning out (near-0 is modeled as 0 given difficulty in approximating such complex dynamics) but the model is quite clearly suggesting this is the first and only wave. I don't think this tool does what you think it does. I think you are wrong. I referenced the part of the FAQs that explains the question I posed to try to avoid this response. You misinterpreted my reference to AIDS as well. The point of my original post was that experts recommend different actions for different threats. You said I was wrong and then repeated what I wrote in a specific instance (as opposed to general). I think you are fighting with idiots too much 8). My point isn't to suggest that the model is saying we are eradicating the disease in 2 months of social distancing. My point was that expectations have been lowered to the point that the public is going to be disappointed in the future. I didn't misinterpret your reference to AIDS, but maybe mine wasn't clear. I was also pointing out that different threats necessitate different response, reinforcing that with an illustration. And if your point was that "that expectations have been lowered to the point that the public is going to be disappointed in the future", then I didn't get that from what you wrote at all, but I agree with it if that's what people are taking from it. But I hope people are smart enough (well, futile hope, maybe) to use the model to try to get a ballpark for the peaks in deaths, resource usage, etc, but don't use it as a comprehensive model for the whole epidemic.
cwericb Posted April 8, 2020 Posted April 8, 2020 "While many in South East Asian countries went at least till now without shutdown and low deaths?" Perhaps, but what will it look like in a month's time?
ERICOPOLY Posted April 8, 2020 Posted April 8, 2020 May be Japan and Singapore will start shutdown. India started shutdown on March 22nd. See the numbers since March 22nd. https://www.worldometers.info/coronavirus/country/india/ As opposed to? As opposed to before March 22nd. Isnt the shutdown supposed to reduce cases and deaths, at least after 2 weeks? If it went down, the curve wouldn't be flattened. Their goal was to flatten it. That implies no decline in the case load.
Liberty Posted April 8, 2020 Posted April 8, 2020 May be Japan and Singapore will start shutdown. India started shutdown on March 22nd. See the numbers since March 22nd. https://www.worldometers.info/coronavirus/country/india/ As opposed to? As opposed to before March 22nd. Isnt the shutdown supposed to reduce cases and deaths, at least after 2 weeks? If it went down, the curve wouldn't be flattened. Their goal was to flatten it. That implies no decline in the case load. It takes more than two weeks to see an effect because no shutdown is perfect, and those who catch it then infect their families with whom they're quarantined (they stay asymptomatic for days), and it then takes weeks for those people to fight through the disease, and hopefully they don't in turn don't infect too many other people.. For India: Looks like as of 2 days ago, India did 18,000 tests a day. So even if somehow 100% of tests came back positive, they couldn't have more than a few tens of thousands of reported cases in a country of 1.35bn. People shouldn't confuse reported cases with actual cases. https://www.bloombergquint.com/coronavirus-outbreak/india-aims-to-scale-up-covid-19-testing-to-one-lakh-daily India aims to ramp-up its Covid-19 testing by fivefold as the number of cases of infection mounts in Asia’s third-largest economy. The Indian Council for Medical Research is working to scale-up the testing capacity to 1 lakh a day in the upcoming months as part of its preparedness for a worst case scenario, according to a statement. That compares with 18,000 tests daily at present. This here has even lower numbers: https://www.statista.com/statistics/1107370/india-coronavirus-covid-19-tests-per-day/ 12k at peak, and in the 1-2k range just a week before that. This official release seems to say that capacity is now 18k, and that total cumulative tests so far for India is 96k. https://www.icmr.nic.in/sites/default/files/press_realease_files/Press_Brief_6April2020.pdf
Liberty Posted April 8, 2020 Posted April 8, 2020 The president has been live-tweeting Fox segments about the coronavirus, telling the public that he relies on the network for information about its spread, and even bringing Fox personalities like Ingraham into the White House to advise his administration. He told Sean Hannity just last night that he has been watching Hannity’s show for guidance on which states need additional ventilators from the federal government. https://www.mediamatters.org/fox-news/fox-news-dangerous-coronavirus-mission-accomplished-moment
RichardGibbons Posted April 8, 2020 Posted April 8, 2020 Let me repost my 5 things that people don't understand. [*]Exponential growth [*]It takes weeks between infection and hospitalization and death [*]Hospitals are not infinitely expandable--if enough people come in, hospitals run out of resources [*]If you're in the ICU with this, you are likely in there for weeks [*]That without ventilators, the death rate increases dramatically Deaths are usually tested correctly. Deaths also went up after 2 weeks of shutdown from 2 on March 22nd to 24 now. How long is it supposed to take for this flatening to happen and why? The things you don't understand are items 2, 3, 4, and 5. When you put in the lockdown, it seems like it's about 2 weeks until cases fall (assuming similar testing criteria, which seems to rarely be true.) Once someone is symptomatic, it seems to take somewhere in the range of 2-3 weeks for them to die. Hospital resources, and in particular the ICU, fill up and the death rate increases as this happens. So, if your ICU isn't filled, you'd expect to see deaths decline between 4 and 5 weeks after lockdown. If the ICU reaches capacity, you'd expect deaths to decline even later than that, since the percentage of people dying because they can't get access to ICU treatments will increase.
Dalal.Holdings Posted April 8, 2020 Posted April 8, 2020 Let me repost my 5 things that people don't understand. [*]Exponential growth [*]It takes weeks between infection and hospitalization and death [*]Hospitals are not infinitely expandable--if enough people come in, hospitals run out of resources [*]If you're in the ICU with this, you are likely in there for weeks [*]That without ventilators, the death rate increases dramatically Deaths are usually tested correctly. Deaths also went up after 2 weeks of shutdown from 2 on March 22nd to 24 now. How long is it supposed to take for this flatening to happen and why? The things you don't understand are items 2, 3, 4, and 5. When you put in the lockdown, it seems like it's about 2 weeks until cases fall (assuming similar testing criteria, which seems to rarely be true.) Once someone is symptomatic, it seems to take somewhere in the range of 2-3 weeks for them to die. Hospital resources, and in particular the ICU, fill up and the death rate increases as this happens. So, if your ICU isn't filled, you'd expect to see deaths decline between 4 and 5 weeks after lockdown. If the ICU reaches capacity, you'd expect deaths to decline even later than that, since the percentage of people dying because they can't get access to ICU treatments will increase. Thanks for posting this. Please have it ready to repost when we need it again a few days from now. It’s like we have to reinvent the wheel each time.
arcube Posted April 8, 2020 Posted April 8, 2020 Let me repost my 5 things that people don't understand. [*]Exponential growth [*]It takes weeks between infection and hospitalization and death [*]Hospitals are not infinitely expandable--if enough people come in, hospitals run out of resources [*]If you're in the ICU with this, you are likely in there for weeks [*]That without ventilators, the death rate increases dramatically Deaths are usually tested correctly. Deaths also went up after 2 weeks of shutdown from 2 on March 22nd to 24 now. How long is it supposed to take for this flatening to happen and why? The things you don't understand are items 2, 3, 4, and 5. When you put in the lockdown, it seems like it's about 2 weeks until cases fall (assuming similar testing criteria, which seems to rarely be true.) Once someone is symptomatic, it seems to take somewhere in the range of 2-3 weeks for them to die. Hospital resources, and in particular the ICU, fill up and the death rate increases as this happens. So, if your ICU isn't filled, you'd expect to see deaths decline between 4 and 5 weeks after lockdown. If the ICU reaches capacity, you'd expect deaths to decline even later than that, since the percentage of people dying because they can't get access to ICU treatments will increase. Thanks for posting this. Please have it ready to repost when we need it again a few days from now. It’s like we have to reinvent the wheel each time. +1. I support the motion to keep it handy.
Guest Schwab711 Posted April 8, 2020 Posted April 8, 2020 But I hope people are smart enough (well, futile hope, maybe) to use the model to try to get a ballpark for the peaks in deaths, resource usage, etc (assuming current lockdown conditions remaining through the end of May), but don't use it as a comprehensive model for the whole epidemic. To further hammer home the point for others: Trump quoted this model in reference to reopening the economy. I don't know what the right strategy is for the country but the model is not relevant if we make any changes for the next 2 months since all calculations are based on that assumption. We'd expect cases/deaths to be higher if we change course at any point over the next 7 weeks. As we went back and forth on, this is a model to predict the first wave death totals, under lockdown conditions. This does not represent an estimate for the whole pandemic. To bring this back to investing, stock prices are too high if you are bullish because of case/death statistics.
Investor20 Posted April 8, 2020 Posted April 8, 2020 Let me repost my 5 things that people don't understand. [*]Exponential growth [*]It takes weeks between infection and hospitalization and death [*]Hospitals are not infinitely expandable--if enough people come in, hospitals run out of resources [*]If you're in the ICU with this, you are likely in there for weeks [*]That without ventilators, the death rate increases dramatically Deaths are usually tested correctly. Deaths also went up after 2 weeks of shutdown from 2 on March 22nd to 24 now. How long is it supposed to take for this flatening to happen and why? The things you don't understand are items 2, 3, 4, and 5. When you put in the lockdown, it seems like it's about 2 weeks until cases fall (assuming similar testing criteria, which seems to rarely be true.) Once someone is symptomatic, it seems to take somewhere in the range of 2-3 weeks for them to die. Hospital resources, and in particular the ICU, fill up and the death rate increases as this happens. So, if your ICU isn't filled, you'd expect to see deaths decline between 4 and 5 weeks after lockdown. If the ICU reaches capacity, you'd expect deaths to decline even later than that, since the percentage of people dying because they can't get access to ICU treatments will increase. 1) There was no exponential growth in India before shutdown. Hardly 1-2 deaths a day till March 22nd. In fact on March 21st its ZERO deaths. The exponential growth came after shutdown - I didn't know that is the purpose of shutdown. "When you put in the lockdown, it seems like it's about 2 weeks until cases fall" -2.5 weeks later zero deaths on March 21st came to about 24 deaths on April 07th 3, 4, 5 are irrelavant to India as the absolute cases are small for a country of billion+ people.
Liberty Posted April 8, 2020 Posted April 8, 2020 1) There was no exponential growth in India before shutdown. Hardly 1-2 deaths a day till March 22nd. In fact on March 21st its ZERO deaths. The exponential growth came after shutdown - I didn't know that is the purpose of shutdown. "When you put in the lockdown, it seems like it's about 2 weeks until cases fall" -2.5 weeks later zero deaths on March 21st came to about 24 deaths on April 07th 3, 4, 5 are irrelavant to India as the absolute cases are small for a country of billion+ people. You know that every other country started with just 1 case and 1 death, right?
pocoapoco Posted April 8, 2020 Posted April 8, 2020 Does anyone have decent statistics on survival rate after being put on a ventilator? With all the urgency to get ventilators perhaps we take it for granted that a ventilator will save people but there’s a moment when you decide to use a ventilator on someone At that time , if you (1) choose not to put them on it (or don’t have one available) then they have a certain mortality rate. If you (2) put them on one they have a certain mortality rate. I would like to know how these compare to evaluate the marginal benefit of having a ventilator. I’ve seen bits and pieces anecdotally suggesting only a very small percent of people who go on a ventilator come out the other side alive, and if they do come out, many are seemingly permanently impaired (aka still need some sort of life support). That’s why I’m seeking more concrete data.
cubsfan Posted April 8, 2020 Posted April 8, 2020 Can someone answer these 2 questions for me? - What is the real CFR for Covid19? - How many people actually have or have had Covid19? I'd like to plug those into my model..
minten Posted April 8, 2020 Posted April 8, 2020 Does anyone have decent statistics on survival rate after being put on a ventilator? My country published survival rates from those admitted to the IC. Not completely the same as being put on a ventilator, as apparently there are those that are admitted to the IC without a ventilator (like Boris Johnson), but it gets close. Under 40 there are few admissions and virtually everyone surives. From 40-60 most surive, from 60-70 it's a coin flip, over 70 (by far the largest group) most die. The overall surival rate is about 50%. As a side note, old people are very often not admitted to the IC here (because it's a really lonely, terrible way to die), unlike for example in Italy where the culture is different. IC survival rate over there must be much much lower, but I haven't found any proof of that in stats.
Liberty Posted April 8, 2020 Posted April 8, 2020 Can someone answer these 2 questions for me? - What is the real CFR for Covid19? - How many people actually have or have had Covid19? I'd like to plug those into my model.. What is the real forward total return on your portfolio? How many people working for companies you own have done something illegal in the past month? Thanks. (It's not because you can't know something with precision that you know nothing about it and should just do random things... you can look at past experience, base rates, mechanistic understanding of the situation, etc, to determine what's the best course of action based on the best available data at the time, while adjusting constantly as new data comes in and as what you're doing has an impact on the trajectory, and with probabilistic models (note the wide outcome distribution). Otherwise, if you really believe the line that you're trying to sell here, why are you investing at all? You can't know everything about what you invest in and you can't know the future, so might as well give up, right?)
RichardGibbons Posted April 8, 2020 Posted April 8, 2020 +1. I can't believe he thought that people were ignoring those questions because they were so good that they refuted everyone. (As opposed to what actually happened--everyone recognizing that the questions were so stupid that everyone would know they were stupid, so nobody bothered to reply.)
cubsfan Posted April 8, 2020 Posted April 8, 2020 Can someone answer these 2 questions for me? - What is the real CFR for Covid19? - How many people actually have or have had Covid19? I'd like to plug those into my model.. What is the real forward total return on your portfolio? How many people working for companies you own have done something illegal in the past month? Thanks. (It's not because you can't know something with precision that you know nothing about it and should just do random things... you can look at past experience, base rates, mechanistic understanding of the situation, etc, to determine what's the best course of action based on the best available data at the time, while adjusting constantly as new data comes in and as what you're doing has an impact on the trajectory, and with probabilistic models (note the wide outcome distribution). Otherwise, if you really believe the line that you're trying to sell here, why are you investing at all? You can't know everything about what you invest in and you can't know the future, so might as well give up, right?) Just an honest question on how much trust you can put into a model where the underlying data is unknown. I know it might humble some people. At least it should. Does it look like we were headed to 500,000 deaths in the USA? Seems like we will save thousands of lives as well by the severe reduction in driving that is happening. Should we just stay on lockdown? Be nice to understand the real tradeoffs taking place by our esteemed scientific community.
RichardGibbons Posted April 8, 2020 Posted April 8, 2020 1) There was no exponential growth in India before shutdown. Hardly 1-2 deaths a day till March 22nd. In fact on March 21st its ZERO deaths. The exponential growth came after shutdown - I didn't know that is the purpose of shutdown. So you're saying that the shutdown caused exponential growth of deaths? Like, your expectation would be if you isolate people into groups of 4 and never let them have any sort of contact with any other group, your death rate from COVID-19 would be higher than allowing everyone to intermix to their heart's content? Huh. That's kind of a strange theory, but OK. "When you put in the lockdown, it seems like it's about 2 weeks until cases fall" -2.5 weeks later zero deaths on March 21st came to about 24 deaths on April 07th 3, 4, 5 are irrelavant to India as the absolute cases are small for a country of billion+ people. Yeah, I included those in there because I thought you wanted an explanation of when deaths should be decreasing, not to discredit your "quarantining causes order-of-magnitude spikes in COVID-19 deaths" theory. My mistake.
Liberty Posted April 8, 2020 Posted April 8, 2020 Just an honest question on how much trust you can put into a model where the underlying data is unknown. I know it might humble some people. At least it should. Does it look like we were headed to 500,000 deaths in the USA? Seems like we will save thousands of lives as well by the severe reduction in driving that is happening. Should we just stay on lockdown? Be nice to understand the real tradeoffs taking place by our esteemed scientific community. 1) You're not doing what you're saying your doing. You're not some epistemology geek trying to refine our thinking. You're just trying to score points with a few questions that don't really add any insight to the discussion, because we're clearly already aware of what you're saying here and then some. Meanwhile, you ignore all the other points that have been made to you and have no response for them and haven't updated your thinking on anything even when people point out the obvious flaws in what you say, so this isn't good faith discussion, just BS trolling on your part. Deaths from driving don't compound daily for weeks and weeks and risk destroying the healthcare sector of multiple countries by overwhelming it with patients while shutting down international travel and supply chains, btw. 2) The US definitely was on track for something in the magnitude of 500k deaths or more a few weeks ago when those numbers were floated, but changes in policy have changed the trajectory (mostly thanks to governors and mayors deciding to shut down in the face of a vacuum of leadership at the federal level, but many states were still much slower than they should've been if there had been leadership at the top, and they'll have bigger outbreaks than they otherwise would've -- was it the governor of Georgia who recently said he didn't know symptomatic people could transmit it and then shut down because of it?). There's reflexivity here, as high predictions lead to bigger actions which lead to fewer deaths (and then to dumbass saying "well, we did it for nothing, see how it wasn't as bad as initially predicted?"). It's Y2K all over again. If we do it right, the early predictions don't come to pass, but it doesn't mean that we should've done nothing. The variables are linked, not independent.
RichardGibbons Posted April 8, 2020 Posted April 8, 2020 Does it look like we were headed to 500,000 deaths in the USA? It actually kind of does look like we were heading for 500,000 deaths in the USA. New York City has 4,111 deaths. And daily "at home" deaths spiked from 20/day to 200 per day when NYC was having 240 recorded COVID-19 deaths. If we assume another 2/3rds of "unattributed deaths at home", then we're talking about 6,500-7000 COVID-19 deaths just in New York while taking precautionary measures. So yeah, when you start extrapolating it to the rest of the country, 500K seems easily achievable if no preventative actions were taken. Low, in fact. This demonstration of how broken your mental model is might humble some people. At least it should.
Investor20 Posted April 8, 2020 Posted April 8, 2020 1) There was no exponential growth in India before shutdown. Hardly 1-2 deaths a day till March 22nd. In fact on March 21st its ZERO deaths. The exponential growth came after shutdown - I didn't know that is the purpose of shutdown. So you're saying that the shutdown caused exponential growth of deaths? Like, your expectation would be if you isolate people into groups of 4 and never let them have any sort of contact with any other group, your death rate from COVID-19 would be higher than allowing everyone to intermix to their heart's content? Huh. That's kind of a strange theory, but OK. "When you put in the lockdown, it seems like it's about 2 weeks until cases fall" -2.5 weeks later zero deaths on March 21st came to about 24 deaths on April 07th 3, 4, 5 are irrelavant to India as the absolute cases are small for a country of billion+ people. Yeah, I included those in there because I thought you wanted an explanation of when deaths should be decreasing, not to discredit your "quarantining causes order-of-magnitude spikes in COVID-19 deaths" theory. My mistake. I dont have any theory. I dont think there is enough information to make a theoretical model. Empirically the theory YOU proposed doesnt match the data at hand. In India the exponential growth started after lockdown. Thats not a theory. Thats data. Also, the deaths all of South East Asia are low where there were no lockdowns till now. Thats not a theory.. Not a model. Thats data.
arcube Posted April 8, 2020 Posted April 8, 2020 1) There was no exponential growth in India before shutdown. Hardly 1-2 deaths a day till March 22nd. In fact on March 21st its ZERO deaths. The exponential growth came after shutdown - I didn't know that is the purpose of shutdown. So you're saying that the shutdown caused exponential growth of deaths? Like, your expectation would be if you isolate people into groups of 4 and never let them have any sort of contact with any other group, your death rate from COVID-19 would be higher than allowing everyone to intermix to their heart's content? Huh. That's kind of a strange theory, but OK. "When you put in the lockdown, it seems like it's about 2 weeks until cases fall" -2.5 weeks later zero deaths on March 21st came to about 24 deaths on April 07th 3, 4, 5 are irrelavant to India as the absolute cases are small for a country of billion+ people. Yeah, I included those in there because I thought you wanted an explanation of when deaths should be decreasing, not to discredit your "quarantining causes order-of-magnitude spikes in COVID-19 deaths" theory. My mistake. I dont have any theory. I dont think there is enough information to make a theoretical model. Empirically the theory YOU proposed doesnt match the data at hand. In India the exponential growth started after lockdown. Thats not a theory. Thats data. Also, the deaths all of South East Asia are low where there were no lockdowns till now. Thats not a theory.. Not a model. Thats data. That sure sounds like a logical cry you are making - Yes that is DATA. In itself it is useless, and has to be interpreted in some manner within a context. You keep talking about South East Asia - What countries are you counting in those?
Investor20 Posted April 8, 2020 Posted April 8, 2020 1) There was no exponential growth in India before shutdown. Hardly 1-2 deaths a day till March 22nd. In fact on March 21st its ZERO deaths. The exponential growth came after shutdown - I didn't know that is the purpose of shutdown. So you're saying that the shutdown caused exponential growth of deaths? Like, your expectation would be if you isolate people into groups of 4 and never let them have any sort of contact with any other group, your death rate from COVID-19 would be higher than allowing everyone to intermix to their heart's content? Huh. That's kind of a strange theory, but OK. "When you put in the lockdown, it seems like it's about 2 weeks until cases fall" -2.5 weeks later zero deaths on March 21st came to about 24 deaths on April 07th 3, 4, 5 are irrelavant to India as the absolute cases are small for a country of billion+ people. Yeah, I included those in there because I thought you wanted an explanation of when deaths should be decreasing, not to discredit your "quarantining causes order-of-magnitude spikes in COVID-19 deaths" theory. My mistake. I dont have any theory. I dont think there is enough information to make a theoretical model. Empirically the theory YOU proposed doesnt match the data at hand. In India the exponential growth started after lockdown. Thats not a theory. Thats data. Also, the deaths all of South East Asia are low where there were no lockdowns till now. Thats not a theory.. Not a model. Thats data. That sure sounds like a logical cry you are making - Yes that is DATA. In itself it is useless, and has to be interpreted in some manner within a context. You keep talking about South East Asia - What countries are you counting in those? Eventhough not separate country, Hongkong, and then Taiwan, Japan, Malaysia, Singapore. There were a bunch of articles about Taiwan how well they managed. And they didnt have shutdown.
Recommended Posts