This was originally a Facebook post of mine from 2020-8-1, and is archived here as a curiosity. Minor changes to formatting, as well as basic copy-edits, may have been made in the transition from Facebook post to web format.
Today, the United States reported 60,264 cases of the novel coronavirus, down from 65,413 cases a week ago (Covid Tracking Project).
We reported 1172 deaths today, up from 1037.
So it's more of the same: cases are still going down, while deaths are still going up. Now, cases can't keep going down while deaths keep going up forever. How long can this go on? I can't say for sure, but I can do a little back-of-the-envelope calculation.
Today, in my little computer program I use to track my cases, is something like day 143 of the epidemic in the US. I started tracking 143 days ago, and I can't remember how I picked that particular date.
Anyhow, the last big rise in cases started on about day 94, while the last big rise in deaths started on about day 117. In other words, deaths, at least for the last big flip in deaths and cases, followed cases, but about 23 days later. That makes sense, because people who do die of this thing often die a couple weeks after they start experiencing symptoms.
Now, I'm not some kind of Covid forecaster, but let's just assume for the sake of the exercise that deaths will follow cases on a 23-day delay going forward. There are several reasons this assumption could be wrong, but let's run with it. It was about eleven days ago that cases started falling again, so on those assumptions we'd guess that we're about twelve days away from seeing deaths start to drop again, too.
But it's a bad idea to trust a little calculation like that in the middle of an extremely confusing outbreak. It's just a mental exercise based on tracking the very simplest variables.
Now, there's a guy named Youyang Gu who is, according to just about everybody, pretty darn good at this kind of thing. His website is at covid19-projections.com/, and everybody from Nate Silver to the CDC watches his numbers. He also has the number of deaths per day hitting a peak on August 13th, and starting a gentle decline after that.
So let's watch and see what happens the next two weeks, and if deaths start falling around August 13th, that'll be one point in favor of simple calculations from the graphs of deaths and cases, and one point in favor of paying attention to Youyang Gu.
One thing that makes me think this Gu is a smart cookie is that Gu's projections don't go out past the end of October. I respect someone who doesn't pretend to know more than they do.
The usual graphs are attached.
This page is released under the CC0 1.0 license.