6 ms·

Let's compare apples to apples and stick to the US for a minute, which is what I believe the WSJ was doing. The winter of 2017-2018 was a bad flu year, and the

by xscott 7y ago

Let's compare apples to apples and stick to the US for a minute, which is what I believe the WSJ was doing. The winter of 2017-2018 was a bad flu year, and there were about 61,000 deaths from it in the US (uncertainty interval: 46K to 95K) [0]. According to the worldometers site, there are 785 deaths so far in the US from corona virus [1]. It's probably a poor estimate, but if we could extrapolate from the US population to the world population, 1.28 million died from the flu that year (61e3*7.8e9/372e6). [0] https://www.cdc.gov/flu/about/burden/index.html https://www.cdc.gov/flu/about/burden/index.html [1] https://www.worldometers.info/coronavirus/country/us/ https://www.worldometers.info/coronavirus/country/us/ EDIT: I found another link with a better estimate for world deaths from the flu. According to the World Health Organization, between 290,000 and 650,000 die from the flu each year [2]. So my estimate above was off by a factor of two. [2] https://www.who.int/news-room/fact-sheets/detail/influenza-(seasonal) https://www.who.int/news-room/fact-sheets/detail/influenza-(...

DanBC 7y ago
The death rate in the US is doubling every 3 days. (A bit less than 3 days.) What will it look like in 21 days? 0 785 3 1570 6 3140 9 6280 12 12560 15 25120 18 50240 21 100480
- xscott 7y ago
  Yes, exponential curves are scary. So are logistics curves. We should know in 21 days if your prediction is right.
  - DiogenesKynikos 7y ago
    The only things that will bend the curve downward are: 1. The virus begins to exhaust the available fuel (i.e., something approaching 60-70% of the population gets sick), or 2. Strict social distancing measures are implemented. Option 1 means that ~200 million Americans contract the virus, ~20 million need hospitalization, and ~2 million die. At that level of hospitalization, though, the healthcare system will collapse, and the death rate (not only for coronavirus patients) may well go up significantly. That's the logistic curve you'll get if you leave things to run their course.
    - xscott 7y ago
      My post above just stated the facts as they are today or a couple years ago, with citations and uncertainty ranges. Do you have a strong prediction for the future with a date and count? I'm not sure what qualifies as "strict", and I'm only interested in counting deaths.
      - DiogenesKynikos 7y ago
        Your post above compared deaths so far due to COVID-19 to yearly deaths due to influenza. That's a very silly comparison, because we're only in the very early stages of the outbreak. The outbreak is growing exponentially at this point, and basic epidemiological models say that it will continue to do so until a significant fraction of the population becomes infected. COVID-19 is estimated to have an R_0 of around 2.4, meaning that if nothing is done to lower transmission, the epidemic will only begin to recede once approximately 1-1/2.4 = 0.58 (58%) of the population is infected. The current doubling time for the epidemic is 3 days in the US, meaning that the US will reach the saturation level of ~58% of the population infected sometime this May. In the face of the cold logic of exponential growth in epidemiology, you can't try to reassure people by pointing to the current numbers of infections. You have to look at R_0 (~2.4), the level of preexisting immunity (0%), and the mortality rate (~1%), and act accordingly. I don't want to see 2 million Americans dead by the end of May. Pointing out that only a few hundred have died so far and asking me to ignore basic epidemiology doesn't reassure me.
        xscott 7y ago
        No one asked you to ignore anything, and my goal is certainly not to reassure you. Very silly or not, I'm not making predictions - I'm just keeping track of a few of you who are. Are you predicting 2 million Americans dead from COVID-19 by the end of May? If you feel confident in your predictions, make a strong claim. If you're less confident, put some error margins around your claim. If you need to put some caveats (i.e. "strict"), be specific. If you're uncertain, it would be more intellectually honest if you indicated that when you post.
  - wilgertvelinga 7y ago
    Five days later it doesn't look far off: 3,178 deaths in the USA according to worldometer.
    - xscott 7y ago
      I keep checking back too... One way or another, we'll see.
    - xscott 7y ago
      Not sure if you're still checking back, but at day 15 the US deaths are at 16k where his prediction says 25k.
      - DanBC 7y ago
        That's good. The cumulative deaths are now doubling at a slower rate (something like 5.5 days?), partly because stronger social distancing measures have been put in place. But also, let's see what the CDC says next week, because those figures are laggier but more accurate. https://www.cdc.gov/nchs/nvss/vsrr/COVID19/index.htm https://www.cdc.gov/nchs/nvss/vsrr/COVID19/index.htm
        xscott 7y ago
        And how long do you predict a 5.5 day doubling period to continue?
        DanBC 7y ago
        For clarity, I'm not predicting. 15 days ago was just after New York and California implemented shelter-in-place orders, so we'd expect to see levelling in the numbers of daily new infections starting about two weeks after (now), with levelling in daily new deaths about 10 days after that (about 20th April). Here's what other people are predicting - a peak on April 12th with over 2000 deaths in one day, and total cumulative deaths just over 23000 . https://covid19.healthdata.org/united-states-of-america https://covid19.healthdata.org/united-states-of-america
- zepolen 7y ago
  In just 2 months it will have killed 10 billion people!
  - mythrwy 7y ago
    Sounds like a friend who in early 2018 told me that if Ethereum kept going up at even half the rate it had been for the next 3 years he would be retired! I sadly had to pop his bubble by pointing out that value represented more than the sum of money in the entire world.
    - zepolen 7y ago
      That was the joke.
      - mythrwy 7y ago
        Yep. I found it funny too. (Sadly my example wasn't a joke).
  - DiogenesKynikos 7y ago
    You have to ask yourself what drives exponential growth, and what eventually breaks the exponential. What drives exponential growth in an epidemic is the fact that each person infects some number of other people, on average. The more people are sick, the more people get infected each day. Growth is proportional to size, so you've got an exponential curve. What stops exponential growth in an epidemic is when so many people are immune that the average sick person isn't able to infect more than one other person. At that point, the epidemic dies off exponentially. You only reach the exponential decrease phase once a significant fraction of the population has immunity. That only happens once they've gotten sick and recovered. Based on the transmissibility of COVID-19, it's expected to eventually infect 60% of the world population, unless a vaccine is developed before it does.
    - zepolen 7y ago
      Woosh.
- Ancalagon 7y ago
  FYI 785 is current TOTAL deaths in the US, not daily deaths (as of writing this comment)...
  - xscott 7y ago
    That's a good catch, but I assumed he was just being sloppy in his terminology.

by xscott 7y ago

DanBC 7y ago

The death rate in the US is doubling every 3 days. (A bit less than 3 days.) What will it look like in 21 days? 0 785 3 1570 6 3140 9 6280 12 12560 15 25120 18 50240 21 100480

xscott 7y ago
Yes, exponential curves are scary. So are logistics curves. We should know in 21 days if your prediction is right.
- DiogenesKynikos 7y ago
  The only things that will bend the curve downward are: 1. The virus begins to exhaust the available fuel (i.e., something approaching 60-70% of the population gets sick), or 2. Strict social distancing measures are implemented. Option 1 means that ~200 million Americans contract the virus, ~20 million need hospitalization, and ~2 million die. At that level of hospitalization, though, the healthcare system will collapse, and the death rate (not only for coronavirus patients) may well go up significantly. That's the logistic curve you'll get if you leave things to run their course.
  - xscott 7y ago
    My post above just stated the facts as they are today or a couple years ago, with citations and uncertainty ranges. Do you have a strong prediction for the future with a date and count? I'm not sure what qualifies as "strict", and I'm only interested in counting deaths.
    - DiogenesKynikos 7y ago
      Your post above compared deaths so far due to COVID-19 to yearly deaths due to influenza. That's a very silly comparison, because we're only in the very early stages of the outbreak. The outbreak is growing exponentially at this point, and basic epidemiological models say that it will continue to do so until a significant fraction of the population becomes infected. COVID-19 is estimated to have an R_0 of around 2.4, meaning that if nothing is done to lower transmission, the epidemic will only begin to recede once approximately 1-1/2.4 = 0.58 (58%) of the population is infected. The current doubling time for the epidemic is 3 days in the US, meaning that the US will reach the saturation level of ~58% of the population infected sometime this May. In the face of the cold logic of exponential growth in epidemiology, you can't try to reassure people by pointing to the current numbers of infections. You have to look at R_0 (~2.4), the level of preexisting immunity (0%), and the mortality rate (~1%), and act accordingly. I don't want to see 2 million Americans dead by the end of May. Pointing out that only a few hundred have died so far and asking me to ignore basic epidemiology doesn't reassure me.
      - xscott 7y ago
        No one asked you to ignore anything, and my goal is certainly not to reassure you. Very silly or not, I'm not making predictions - I'm just keeping track of a few of you who are. Are you predicting 2 million Americans dead from COVID-19 by the end of May? If you feel confident in your predictions, make a strong claim. If you're less confident, put some error margins around your claim. If you need to put some caveats (i.e. "strict"), be specific. If you're uncertain, it would be more intellectually honest if you indicated that when you post.
- wilgertvelinga 7y ago
  Five days later it doesn't look far off: 3,178 deaths in the USA according to worldometer.
  - xscott 7y ago
    I keep checking back too... One way or another, we'll see.
  - xscott 7y ago
    Not sure if you're still checking back, but at day 15 the US deaths are at 16k where his prediction says 25k.
    - DanBC 7y ago
      That's good. The cumulative deaths are now doubling at a slower rate (something like 5.5 days?), partly because stronger social distancing measures have been put in place. But also, let's see what the CDC says next week, because those figures are laggier but more accurate. https://www.cdc.gov/nchs/nvss/vsrr/COVID19/index.htm https://www.cdc.gov/nchs/nvss/vsrr/COVID19/index.htm
      - xscott 7y ago
        And how long do you predict a 5.5 day doubling period to continue?
        DanBC 7y ago
        For clarity, I'm not predicting. 15 days ago was just after New York and California implemented shelter-in-place orders, so we'd expect to see levelling in the numbers of daily new infections starting about two weeks after (now), with levelling in daily new deaths about 10 days after that (about 20th April). Here's what other people are predicting - a peak on April 12th with over 2000 deaths in one day, and total cumulative deaths just over 23000 . https://covid19.healthdata.org/united-states-of-america https://covid19.healthdata.org/united-states-of-america
zepolen 7y ago
In just 2 months it will have killed 10 billion people!
- mythrwy 7y ago
  Sounds like a friend who in early 2018 told me that if Ethereum kept going up at even half the rate it had been for the next 3 years he would be retired! I sadly had to pop his bubble by pointing out that value represented more than the sum of money in the entire world.
  - zepolen 7y ago
    That was the joke.
    - mythrwy 7y ago
      Yep. I found it funny too. (Sadly my example wasn't a joke).
- DiogenesKynikos 7y ago
  You have to ask yourself what drives exponential growth, and what eventually breaks the exponential. What drives exponential growth in an epidemic is the fact that each person infects some number of other people, on average. The more people are sick, the more people get infected each day. Growth is proportional to size, so you've got an exponential curve. What stops exponential growth in an epidemic is when so many people are immune that the average sick person isn't able to infect more than one other person. At that point, the epidemic dies off exponentially. You only reach the exponential decrease phase once a significant fraction of the population has immunity. That only happens once they've gotten sick and recovered. Based on the transmissibility of COVID-19, it's expected to eventually infect 60% of the world population, unless a vaccine is developed before it does.
  - zepolen 7y ago
    Woosh.
Ancalagon 7y ago
FYI 785 is current TOTAL deaths in the US, not daily deaths (as of writing this comment)...
- xscott 7y ago
  That's a good catch, but I assumed he was just being sloppy in his terminology.