# Predict the number dead from COVID-19 in America by end of 2020 **URL:** \ **Category:** The Quarantine Zone\ **Tags:** covid-quarantine\ **Created:** [March 29, 2020, 2:56am UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561 "2020-03-29T02:56:24Z")\ **Posts on this page:** 20\ **Page:** 10
**Author:** ![CarnalK](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/carnalk/32/486_2.png) [@CarnalK](https://boards.straightdope.com/u/CarnalK)\ **Post date:** [April 19, 2020, 3:59pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/181 "2020-04-19T15:59:00Z")
I actually doubt it would have doubled every 3 days indefinitely without intervention. That is rather beyond doubt on a worldwide basis. Japan didn’t do much of anything with confirmed cases back in January, nowhere near doubling deaths every 3days. Hyper dense India hasn’t showing that growth. The various jurisdictions in the US with a variety of lockdowns don’t show the obviousness of such rates. People seem to have completely forgotten that these were extreme worst case scenarios based on very little information. I bet not a single person here, including me, even remotely remembers what the best case scenarios were without intervention. I don’t think they were ever reported. ---
**Author:** ![BigAppleBucky](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/bigapplebucky/32/178_2.png) [@BigAppleBucky](https://boards.straightdope.com/u/BigAppleBucky)\ **Post date:** [April 19, 2020, 5:26pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/182 "2020-04-19T17:26:57Z")
> [@CarnalK](#): > > I actually doubt it would have doubled every 3 days indefinitely without intervention. That is rather beyond doubt on a worldwide basis. Japan didn’t do much of anything with confirmed cases back in January, nowhere near doubling deaths every 3days. Hyper dense India hasn’t showing that growth. The various jurisdictions in the US with a variety of lockdowns don’t show the obviousness of such rates. . . Of course. If we take yesterday’s rates of increase (4.1% cases, 5% deaths) and project those out from yesterday, we’d reach 1 million cases on the 26th and 100,000 deaths on May 8. Clearly, those rates of growth are still too high. Reducing the rate of increase of deaths to just 1% projects to 503K deaths by the end of the year. Sadly, with so much “reopen now” stupidity happening I think that number is possible. ---
**Author:** ![CarnalK](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/carnalk/32/486_2.png) [@CarnalK](https://boards.straightdope.com/u/CarnalK)\ **Post date:** [April 19, 2020, 6:02pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/183 "2020-04-19T18:02:59Z")
Why assume the death rate will increase at all from here? Looks to me, ignoring the couple of adjustment of method days, it’s not obviously increasing now. I don’t know what numbers you’re looking at, either. There was a drop in the daily death toll yesterday according to Worldometer. 1867 yesterday, 2528 for Friday, 2176 Thursday. ---
**Author:** ![casdave](https://avatars.discourse-cdn.com/v4/letter/c/c6cbf5/32.png) [@casdave](https://boards.straightdope.com/u/casdave)\ **Post date:** [April 22, 2020, 10:23am UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/184 "2020-04-22T10:23:19Z")
There are bound to be anomolies and its the trend that matters, we can see from Sweden that they seem to have a weekly cycle of reporting which has regualr peaks and troughs - probably due to reporting mechanisms and legal requirements for evaluating cause of death. US seems to have some sort of irregularity but it is not as predictable - so the last couples of days have seen significant increases, these might well iron out. I’d suggest that the main concern is the lag time of infection, symptoms and hospitalizations - those protests took place only a few days ago and if they spread the infection it is likely to take a good 10days to two weeks to show - just as todays deaths relate to infections that were recorded some time ago. ---
**Author:** ![RTFirefly](https://avatars.discourse-cdn.com/v4/letter/r/c77e96/32.png) [@RTFirefly](https://boards.straightdope.com/u/RTFirefly)\ **Post date:** [April 23, 2020, 9:27pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/185 "2020-04-23T21:27:19Z")
Here’s a model I think has lost its way: [COVID-19](https://covid19.healthdata.org/united-states-of-america) They just updated it yesterday, but they’re showing total U.S. deaths leveling off at under 67,000. We’ll have our 50,000th death tonight sometime. Their model says the U.S. covid-19 death rate will drop by 50% over the next week. That would be wonderful, and this could be the week it starts dropping for real, but not anything like that. (I would love, love, _love_ to be wrong.) The current death rate reflects the closings and other restrictions and advisories we’d done through the first few days of April - and AFAICT, most of what we’ve done, we’d done by then\*. So it’s hard to see that there’s a sudden drop in the death rate only days away. Hopefully the additional restrictions since then will start making enough of a difference for the death rate to start dropping shortly. Let’s hope - and continue to wash our hands thoroughly, wear masks outside the home, minimize our contact with other people, etc. Those numbers won’t go down by themselves. \*Once you’ve closed schools, nonessential businesses, and stuff like that, there’s not many places to go anyway, so that’s most of the impact of a stay-at-home regime anyway, AFAIAC. ---
**Author:** ![Sage\_Rat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/sage_rat/32/399_2.png) [@Sage\_Rat](https://boards.straightdope.com/u/Sage_Rat)\ **Post date:** [April 23, 2020, 9:38pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/186 "2020-04-23T21:38:04Z")
> [@RTFirefly](#): > > They just updated it yesterday, but they’re showing total U.S. deaths leveling off at under 67,000. We’ll have our 50,000th death tonight sometime. Their model says the U.S. covid-19 death rate will drop by 50% over the next week. That would be wonderful, and this could be the week it starts dropping for real, but not anything like that. (I would love, love, _love_ to be wrong.) Most other experts think that May will be the worst month (though similar in total deaths to April). Based on data from the UK, they only got about half of their deaths through hospitals and another half seem to have died mysteriously, at home or elsewhere, presumably from Corona (though, it could also be fatalities from other illnesses due to shortages of supplies and help). Given that most of our 50k happened this month, that implies that we’re actually at 100k (including deaths at home) and that we should expect around a similar amount next month, putting us up to 200k by end of May. It seems unlikely that we couldn’t get another 100k added in over the remaining 7 months of the year after that. ---
**Author:** ![RTFirefly](https://avatars.discourse-cdn.com/v4/letter/r/c77e96/32.png) [@RTFirefly](https://boards.straightdope.com/u/RTFirefly)\ **Post date:** [April 23, 2020, 10:52pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/187 "2020-04-23T22:52:39Z")
> [@Sage\_Rat](#): > > Most other experts think that May will be the worst month (though similar in total deaths to April). > > Based on data from the UK, they only got about half of their deaths through hospitals and another half seem to have died mysteriously, at home or elsewhere, presumably from Corona (though, it could also be fatalities from other illnesses due to shortages of supplies and help). When the overall fatality counts by month are known, they’re going to be able to compare with the average or trend of recent years, and get a good estimate of excess deaths due to the coronavirus. And of course, many won’t have been counted as a coronavirus death. Some will be coronavirus deaths at home, some will be excess deaths due to other causes in a hospital that’s been stretched too thin, and some will be deaths at home due to other causes, where people were more afraid of going to the hospital and catching the coronavirus than they were of their existing conditions killing them at home. If I had to guess, I’d guess somewhere between 150,000 and 300,000 deaths counted as coronavirus deaths at the time: even if the deaths start declining over the next week, it’s hard for me to see how we don’t get another 50-60K deaths by the end of May, and then another 40-50K over the rest of the summer. And the ultimate count of excess deaths during the time of the plague: like you said, probably about double that range. ---
**Author:** ![RTFirefly](https://avatars.discourse-cdn.com/v4/letter/r/c77e96/32.png) [@RTFirefly](https://boards.straightdope.com/u/RTFirefly)\ **Post date:** [April 23, 2020, 10:59pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/188 "2020-04-23T22:59:37Z")
> [@casdave](#): > > There are bound to be anomolies and its the trend that matters, we can see from Sweden that they seem to have a weekly cycle of reporting which has regualr peaks and troughs - probably due to reporting mechanisms and legal requirements for evaluating cause of death. > > US seems to have some sort of irregularity but it is not as predictable - so the last couples of days have seen significant increases, these might well iron out. Yeah, I wouldn’t put much stock in a 3-4 day trend. We’ve had those before that turned out to be illusory. A 5-7 day moving average is good. > [@](#): > > I’d suggest that the main concern is the lag time of infection, symptoms and hospitalizations - those protests took place only a few days ago and if they spread the infection it is likely to take a good 10days to two weeks to show - just as todays deaths relate to infections that were recorded some time ago. Fortunately in terms of spreading the virus, those protests were pretty small. OTOH, a metric ton of people voted in person in Wisconsin on April 7, and [it may already be showing up in the body count](https://talkingpointsmemo.com/edblog/covid19-and-the-wisconsin-election). ---
**Author:** ![DSeid](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/dseid/32/20194_2.png) [@DSeid](https://boards.straightdope.com/u/DSeid)\ **Post date:** [April 24, 2020, 12:42am UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/189 "2020-04-24T00:42:37Z")
> [@RTFirefly](#): > > Here’s a model I think has lost its way: [COVID-19](https://covid19.healthdata.org/united-states-of-america) > > They just updated it yesterday, but they’re showing total U.S. deaths leveling off at under 67,000. We’ll have our 50,000th death tonight sometime. Their model says the U.S. covid-19 death rate will drop by 50% over the next week. That would be wonderful, and this could be the week it starts dropping for real, but not anything like that. (I would love, love, _love_ to be wrong.) > > The current death rate reflects the closings and other restrictions and advisories we’d done through the first few days of April - and AFAICT, most of what we’ve done, we’d done by then\*. So it’s hard to see that there’s a sudden drop in the death rate only days away. Hopefully the additional restrictions since then will start making enough of a difference for the death rate to start dropping shortly. > > Let’s hope - and continue to wash our hands thoroughly, wear masks outside the home, minimize our contact with other people, etc. Those numbers won’t go down by themselves. > \*Once you’ve closed schools, nonessential businesses, and stuff like that, there’s not many places to go anyway, so that’s most of the impact of a stay-at-home regime anyway, AFAIAC. Their model is a more sophisticated version of my idea to just assume that we follow the curves of those before us. Those who have gone before us have started to drop off on new deaths after new deaths have plateaued for about a week. We are as of the entered 4/22 numbers at 5 days into the one week moving average of daily death rate staying flat. New York’s death rate in particular is falling fast, that same 7 day moving daily average dropping 68 to 57 deaths/million/day and the daily number from 165 peak to 16. True 165 may have been a reporting glitch with a change of definition but the current daily deaths that include “probable deaths” is now not much more than a third of what the number was before that counted only “confirmed deaths”. I’m personally sticking with my 80somethingK but given how much NYC has been driving the U.S. numbers, and how their model has tracked results significantly better than most others have (follow [the time elapse graphic in this article](https://www.nytimes.com/interactive/2020/04/22/upshot/coronavirus-models.html?algo=top_conversion&fellback=false&imp_id=764826643&imp_id=873080876&action=click&module=trending&pgtype=Article®ion=Footer)), I personally wouldn’t _bet_ on their having lost their way _too_ far. ---
**Author:** ![Sage\_Rat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/sage_rat/32/399_2.png) [@Sage\_Rat](https://boards.straightdope.com/u/Sage_Rat)\ **Post date:** [April 24, 2020, 12:58am UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/190 "2020-04-24T00:58:59Z")
> [@RTFirefly](#): > > When the overall fatality counts by month are known, they’re going to be able to compare with the average or trend of recent years, and get a good estimate of excess deaths due to the coronavirus. That’s what the UK did and why I say that we should expect that about 50% of our current fatalities are probably not yet recorded. ---
**Author:** ![RTFirefly](https://avatars.discourse-cdn.com/v4/letter/r/c77e96/32.png) [@RTFirefly](https://boards.straightdope.com/u/RTFirefly)\ **Post date:** [April 24, 2020, 9:55am UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/191 "2020-04-24T09:55:37Z")
> [@DSeid](#): > > Their model is a more sophisticated version of my idea to just assume that we follow the curves of those before us. On this score, [the Financial Times’ graph](https://www.ft.com/coronavirus-latest) says American exceptionalism is still very much a thing. ☹ We’re still rising at a point where deaths in even basket cases like Italy and Spain were clearly on the decline, and had been for awhile. > [@](#): > > I’m personally sticking with my 80somethingK but given how much NYC has been driving the U.S. numbers, and how their model has tracked results significantly better than most others have (follow [the time elapse graphic in this article](https://www.nytimes.com/interactive/2020/04/22/upshot/coronavirus-models.html?algo=top_conversion&fellback=false&imp_id=764826643&imp_id=873080876&action=click&module=trending&pgtype=Article®ion=Footer)), I personally wouldn’t _bet_ on their having lost their way _too_ far. I’ve been following the NY+NJ totals for a couple of weeks, and they keep on being a just a bit over half of the national deaths. (This morning, it’s ~26,300 v. ~23,900.) So even if NY is plummeting, NJ is making up the slack all by itself. At any rate, IHME will be put to the test right away. If its model is basically sound, we should see a major plummet in the U.S. death rate by the middle of next week. ---
**Author:** ![Sage\_Rat](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/sage_rat/32/399_2.png) [@Sage\_Rat](https://boards.straightdope.com/u/Sage_Rat)\ **Post date:** [April 24, 2020, 4:38pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/192 "2020-04-24T16:38:34Z")
> [@RTFirefly](#): > > At any rate, IHME will be put to the test right away. If its model is basically sound, we should see a major plummet in the U.S. death rate by the middle of next week. Math, China, Italy, and Spain all say that you quickly spike and then things slowly decline at a much more gradual pace. I’m not sure why IHME and the media came to accept the idea of a bell curve epidemic progression, but that’s not how it happens. > **[Compartmental models in epidemiology](https://en.wikipedia.org/wiki/Compartmental_models_in_epidemiology)** > > Compartmental models are a very general modelling technique. They are often applied to the mathematical modelling of infectious diseases. The population is assigned to compartments with labels – for example, S, I, or R, (Susceptible, Infectious, or Recovered). People may progress between compartments. The order of the labels usually shows the flow patterns between the compartments; for example SEIS means susceptible, exposed, infectious, then susceptible again. > The origin of such models is the e... [COVID-19 Map - Johns Hopkins Coronavirus Resource Center](https://coronavirus.jhu.edu/map.html) (Daily Cases) ---
**Author:** ![DSeid](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/dseid/32/20194_2.png) [@DSeid](https://boards.straightdope.com/u/DSeid)\ **Post date:** [April 28, 2020, 3:00am UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/193 "2020-04-28T03:00:58Z")
> [@RTFirefly](#): > > Here’s a model I think has lost its way: [COVID-19](https://covid19.healthdata.org/united-states-of-america) > > They just updated it yesterday, but they’re showing total U.S. deaths leveling off at under 67,000. We’ll have our 50,000th death tonight sometime. Their model says the U.S. covid-19 death rate will drop by 50% over the next week. That would be wonderful, and this could be the week it starts dropping for real, but not anything like that. (I would love, love, _love_ to be wrong.) … > [@RTFirefly](#): > > … At any rate, IHME will be put to the test right away. If its model is basically sound, we should see a major plummet in the U.S. death rate by the middle of next week. Using Worldometer numbers … On the date that was written daily deaths rate was 2340 and the seven day average was 2190. Four days later, middle of the next week I’d call it, dropped to a two day average of 1271, pretty close to the 50% dropped they called for over the whole of the next week and certainly qualifying as “a major plummet” by the middle of the next week. Spookily their very first set of curves out looks like it predicted about 1500 deaths/day at this precise point. Yeah two days _could_ still be noise. But so far it’s looking like dismissing them as a model that has “lost its way” may have been premature … Their model, with wide confidence intervals, has predicted about 1100 deaths a day on May 1. No other model is even close to having their prediction value. Most of them were predicting 2200 to 3700 on this date. ---
**Author:** ![DSeid](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/dseid/32/20194_2.png) [@DSeid](https://boards.straightdope.com/u/DSeid)\ **Post date:** [April 29, 2020, 12:40am UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/194 "2020-04-29T00:40:53Z")
AND maybe noise! Today back up. ☹ ---
**Author:** ![orcenio](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/orcenio/32/1036_2.png) [@orcenio](https://boards.straightdope.com/u/orcenio)\ **Post date:** [April 29, 2020, 12:49am UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/195 "2020-04-29T00:49:03Z")
Grim milestone. I see that the official [US covid death toll](https://coronavirus.jhu.edu/map.html) has surpassed the official US Vietnam War death toll. ---
**Author:** ![casdave](https://avatars.discourse-cdn.com/v4/letter/c/c6cbf5/32.png) [@casdave](https://boards.straightdope.com/u/casdave)\ **Post date:** [April 29, 2020, 11:12am UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/196 "2020-04-29T11:12:56Z")
If you go back 2 to 3 weeks of infections you do not see a corresponding drop with infections over the last 3 days, I would have expected death rate would be flat at best, so it appears to me that those two days of lower mortality was likely so be optimistic. It seems that yesterdays total has averaged it out, maybe it is lower overall but we won’t know for a few days - the 7 day average is a better indicator and this does look better. ---
**Author:** ![casdave](https://avatars.discourse-cdn.com/v4/letter/c/c6cbf5/32.png) [@casdave](https://boards.straightdope.com/u/casdave)\ **Post date:** [April 29, 2020, 2:39pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/197 "2020-04-29T14:39:08Z")
Need to correct that first sentence, If you go back 2 to 3 weeks of infections you do not see a corresponding drop with infections over the last 3 days change it to, If you go back 2 to 3 weeks of infections you do not see a corresponding drop with **Deaths** over the last 3 days. You would expect a drop in deaths would follow around a drop of infection two to three weeks previous - we have not seen this, at best its flat, hope this is the case but more time is needed. USA has is not likely to hit 80k by the end of the month, which was within bounds of possibility, its going to be mid 60k ish but when you look at Italy, Spain the drop is quite slow, it has taken 21 days to reduce the death rate by 50%, assume something similar and its looking like around a further 40k deaths in a similar period unless improvements in medical practice make a significant difference. ---
**Author:** ![RTFirefly](https://avatars.discourse-cdn.com/v4/letter/r/c77e96/32.png) [@RTFirefly](https://boards.straightdope.com/u/RTFirefly)\ **Post date:** [April 29, 2020, 3:57pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/198 "2020-04-29T15:57:13Z")
> [@DSeid](#): > > AND maybe noise! Today back up. ☹ One thing that’s guided my thinking here is that the number of new cases was more or less steady through the first few weeks of April, and the testing level was also steady through most of that period. Maybe we were catching more of the cases over that period despite not increasing testing, but given that treatments are only improving at the margins, deaths should mostly increase or decrease based on the number of new cases 2-3 weeks earlier. It’s true that the number of new cases isn’t a very good measure for what’s really happening, but little reason to think it got much better or much worse as a measure during that time. ---
**Author:** ![beowulff](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/beowulff/32/542_2.png) [@beowulff](https://boards.straightdope.com/u/beowulff)\ **Post date:** [April 29, 2020, 4:23pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/199 "2020-04-29T16:23:11Z")
[FWIW, here is an expert who is predicting 800K US deaths in the next 18 months.](https://www.cnn.com/2020/04/21/opinions/bergen-osterholm-interview-two-opinion/index.html) ---
**Author:** ![SlackerInc](https://sea3.discourse-cdn.com/straightdope/user_avatar/boards.straightdope.com/slackerinc/32/275_2.png) [@SlackerInc](https://boards.straightdope.com/u/SlackerInc)\ **Post date:** [April 29, 2020, 11:07pm UTC](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561/200 "2020-04-29T23:07:22Z")
I’m seeing people, including those who should know better (like my wife’s colleagues, all with masters degrees, who are planning a get-together this weekend in direct violation of Minnesota’s stay-at-home order) losing their patience with social distancing guidelines, so that does not augur well for the future. [Previous page](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561.md?page=9) [Next page](https://boards.straightdope.com/t/predict-the-number-dead-from-covid-19-in-america-by-end-of-2020/850561.md?page=11)