June 1, 2020: Out And About

According to reports from friends, it seems like there are a bunch of weirdos running around everyone's neighborhoods. So, on top of everything else, that sounds fun!

37 thoughts on “June 1, 2020: Out And About”

    1. that piece just reinforces what a travesty the Hall voting (or rather, lack of voting) was for Kenny Lofton. 10th all-time in rWAR for CFers, placing him just short of the average HOF CFer on rWAR, JAWS and WAR7, 4 Fielding Grammys, 15 Black Ink and 62 Gray Ink.

  1. So I was looking at COVID-19 data. Had been focused on deaths, since deaths seems more likely to be accurate than cases, and combined deaths with some case data today on a state level--and yeah, the case data is all over the map and hardly seems indicative of anything. Excess deaths is probably even more reliable than deaths, but is lagged by quite a few weeks.

    If you average deaths per day by state over the last 7 days, then add them up, I get 952 deaths per day. If you assume that the next 28 days will be similar to the last 7 days, that gets you to 26.6K additional deaths over the next 28 days.

    Then I went over to the 538 summary on COVID-19 forecasts. 28 days out from 5/31, the mean estimated US cumulative deaths from each forecast ranges from 120K to 128K, which is a range of 16-24K additional deaths over the next 28 days.

    I find it interesting as a spot check that essentially the default forecast method of tomorrow's weather being the same as today's gets that close to models of various sophistication and assumption-making.

    Further, if I was a betting man, I'd bet the over on every one of those forecasts -- it's possible that there are some factors out there working in our favor, but I think it's fair to say that no state is planning to tighten restrictions on interpersonal contact, many are loosening them, and it's unlikely we'll see any states backtrack directionally on stay-at-home orders. More or less, I don't see how we rationally expect anything over the next 28 days to be better than it was the last 7 days.

    And then I don't see how it will be any better the month after that, or the month after that. To pick on one forecast, IHME's estimate of an 80% drop in daily deaths between 6/1 and 8/1 seems like complete wishful thinking (or terrible assumptions, pick your poison.) Optimistically, 10-20M out of 330M+ have been infected, so it looks like we have a long way to go from where I'm sitting.

    1. I looked at CDC data that said we're at 102% of expected deaths (all deaths). So, I'd assume 2% additional deaths from Covid. Then I see that Minnesota is saying deaths are up 20%!

      I don't know if I'm looking at two different timeframes, or Minnesota is that much worse than the nation, or what is going on.

      1. The CDC excess death data I know of is here, though I don't especially like the way the data is presented. From the charts, I've gleaned:

        406-976 excess deaths cumulative, in Minnesota since Feb 1, if we don't assign COVID-reported deaths to any special category
        0-217 excess deaths cumulative, in Minnesota since Feb 1, if we exclude COVID-reported deaths

        This generally suggests that Minnesota has likely been good at identifying most/all COVID-related deaths as COVID-related deaths and not cooking the books to call them pneumonia or something like that. It's possible that your 102% number is referring to everything except COVID deaths (All deaths - Reported COVID deaths)/(Expected Deaths - Reported COVID deaths) and the 20% is (Reported COVID deaths)/(Expected Deaths). Looking through the Minnesota breakdown, the only category that seems suspicious is Alzheimer's and dementia deaths. For the 19th week of the year, the baseline is 95, but reported is 159 -- the most recent data might change, though, so it's possible it's just noise.

        As a comparison, Pennsylvania has 3,500-4,500 excess deaths when excluding COVID-related deaths, so they potentially have had 9,000-10,000 COVID-related deaths as opposed to the 5,500 reported COVID-related deaths. To be fair, a lot of that excess is estimated through their weighted deaths metric as opposed to unweighted, which take 1-8 weeks to come in. Strictly from a statistical standpoint, the testing and reporting is an absolute mess.

    2. I'd bet the over on every one of those forecasts -- it's possible that there are some factors out there working in our favor, but I think it's fair to say that no state is planning to tighten restrictions on interpersonal contact, many are loosening them, and it's unlikely we'll see any states backtrack directionally on stay-at-home orders. More or less, I don't see how we rationally expect anything over the next 28 days to be better than it was the last 7 days.

      Things we have in our favor are warmer weather, sunlight, and a better understanding of how to reduce the risk of infection. The other, more unfortunate thing in our favor is that many of the most vulnerable to COVID-19 were the first to be exposed (the nursing home data from the Northeast states is so, so sad).
      I agree that you're probably right that the forecasts are off, and this projection for Minnesota looks like it lines up with your estimates for additional deaths pretty closely. COVID isn't going away, not anytime soon and probably not ever.

      To pick on one forecast, IHME's estimate of an 80% drop in daily deaths between 6/1 and 8/1 seems like complete wishful thinking (or terrible assumptions, pick your poison.) Optimistically, 10-20M out of 330M+ have been infected, so it looks like we have a long way to go from where I'm sitting

      The same site I linked above has a tab laying out the issues with IHME's model- and there are many. The covid19-projections model, which seems to be more accurate, has an estimate on the number of cases that also falls right in the middle of your range.

      And even given all those big numbers of cases and infections and deaths, there's still this- a different perspective on the severity of this pandemic.

      Spoiler SelectShow
      1. Spoiler SelectShow
        1. Spoiler SelectShow
                1. Spoiler SelectShow
          1. Spoiler SelectShow
            1. Yeah. For the last 8 weeks of data from the CDC, the mortality rate is about 23% above average. If we look at all-cause mortality here, and divide the leading causes of death by 4 to get a rough approximation of how many you'd expect in 3 months, you'd have:

              161K -- Heart Disease
              150K -- Cancer
              40K -- Chronic lower respiratory diseases
              37K -- Stroke
              30K -- Alzheimer's disease
              21K -- Diabetes
              14K -- Flu and Pneumonia
              13K -- Nephritis, nephrotic syndrome and nephrosis
              12K -- Suicide

              And in the last 3 months, there are over 100K deaths related to COVID-19, during a time which relatively extreme measures have been taken to prevent its spread. So suddenly we appear to have a new 3rd-leading cause of death, potentially better or worse depending how seasonality or other factors play out.

          2. Spoiler SelectShow
    1. I have a friend who is a former JAG, and she is pretty much apoplectic about the chance a whole bunch of military is about to violate federal law.

  2. Spent all day dealing with an issue in the house and trying to get some work done, so I was significantly more unplugged than usual. Just catching up now, and what a day, just too many where I have to shake my head and wonder where everything is going. I can't even imagine what July will be like at this point.

    My kid is excited that his baseball team is cleared to practice at least.

Leave a Reply