Post #1244: Final COVID-19 update for the week, new cases continue to decline

Posted on September 11, 2021

The U.S. stands at 45.1 new COVID-19 cases per 100,000 population per day.  That’s down more than 0.5 cases from yesterday.  The seven-day change is -10%, not very different from yesterday.

Just one state (TN) exceeds 100 new cases per 100K per day.  Only 14 states saw an increase in daily new cases over the past week, the rest saw declines.

For the U.S. as a whole, the peak of the Delta wave still looks disorganized, but there is a lot of coherence within regions.

I’m rebasing my short-term graphs so that the begin with the start of the Delta wave.  Next update will be Tuesday, owing the lack of data reporting over the weekend.

A curveball

I’m retiring this graph, and this view, today.  So far, the peak of the Delta wave has been well-behaved, in the sense of matching the shape and duration of a typical peak in this pandemic.  Which is what I sketched in, a month ago.  Here, the “new start date” marks where the short-term graphs will start from today going forward.

Data source for this and other graphs of new case counts:  Calculated from The New York Times. (2021). Coronavirus (Covid-19) Data in the United States. Retrieved 9/11/2021, from”  The NY Times U.S. tracking page may be found at

Before this passes into history, I want to take the time to emphasize the fundamental weirdness of that graph, because I think it’s not obvious to most people. In a nutshell, there’s nothing in the textbooks to suggest that my naive prediction method should work.  And yet, I don’t think it was a matter of luck, which is why I published that prediction far in advance of the peak.

COVID-19 in the U.S. appears to come in self-limiting waves.  These waves tend to have roughly the same duration in every state, but different heights.  States that start the wave early end it early, giving us information on the typical duration for the current wave.  It’s as if, once you get really rapid and widespread infections, the virus runs out of targets, and ebbs.  For now.

So that’s the first bit of weirdness — how neat and orderly these most recent waves have been.

It appears to me that this nice, neat Delta wave was driven by the inherent dynamics of the spread of the virus.  The level of vaccination and level of COVID hygiene in a state might have affected the height of that state’s wave.  The first bit of weirdness is that the timing of it — from start to peak — seems to be remarkably uniform across the states, regardless of what they did and did not do to prevent spread of disease.

That’s the second weirdness.  We have an incredible diversity across the states in terms of level of COVID-19 circulating, mask use, school mask mandates, vaccination levels, genetic makeup of the population, occupational mix, and on and on.  That seems to have no effect on the wave length, just the amplitude.

Epidemiologists tell that “running out of targets” story for the pandemic as a whole.  That process is what generates the classic “sigmoid” (S) curve, if you plot total infections over time.  It’s what generates herd immunity.  Once you get enough people immune to the disease, new infections start to decline.

But under the standard theory, if the peak is lower, it should take less time to get there.  And that does not appear to be the case.  There’s nothing in the textbooks to suggest that all the little sub-waves, all over the U.S., should all have about the same length.  But that appears to be true.

The standard epidemiological story for a peak in a pandemic relies on reaching herd immunity.  We have never approached herd immunity yet in this pandemic, if you define that as having enough people immune to the disease that life can return to mask-less normal.

And that’s the third weirdness.  We keep having these peaks without having herd immunity.  And, let me emphasize that in the Delta wave, we got that peak in some states with almost no measurable change in the behavior of the population.  It’s as if the virus is going to peak, without herd immunity, no matter what.  It’s going to run its course.  All you can do is modulate the amplitude of the peak by taking precautions.

Finally, a curveball and maybe a little ray of hope. 

Each wave after the first one appears to have been set off by a change in the conditions that govern the spread of the disease.  Something changes, the disease becomes easier to spread, and the wave takes off.  Those changing conditions were:

  • Second wave:  Dry air conditioned air in hot-climate states.
  • Third wave:  Dry winter indoor air, particularly in the Midwest and Mountain states.
  • Fourth wave:  Alpha variant, more contagious than prior variants, truncated by mass immunization.
  • Fifth wave:  Delta variant, more contagious than Alpha variant, greatest spread in hot-climate states.

Read Post #894 if you want to know why I emphasize indoor relative humidity.

This all took place against a backdrop of an ever-increasing share of the population that was immune to COVID-19, either from prior infection, or from vaccination.

The question of the moment is, what’s going to happen this winter?  Assuming we don’t get yet-another-variant taking over from Delta, is the transition to drier indoor air going to be enough to set off another wave?

In this blog, I keep talking as if there must be a winter wave this year, because we had one last year.  And I’d still say the odds favor it.  But it’s not a slam-dunk because conditions continue to change.

For one thing, the disease is now concentrated into a smaller subset of the population.  Here’s the nice, clean Commonwealth of Virginia data on COVID-19 infections by vaccination status:

Source:  Virginia Department of Health COVID-19 dashboard.

Converting those weekly rates to daily rates, that 5.5 cases / 100K / day for the vaccinated, versus 84 cases / 100K / day for the unvaccinated.  If Virginia’s vaccinated population were its own state, it would rank last in the U.S. in terms of new COVID-19 cases.  (By contrast, the unvaccinated population would rank 6th, just above Wyoming.)

For another, the Delta variant has just taken out a big chunk of people who would be vulnerable to infection by Delta.  Look at the graph above — the Delta wave dwarfs all but last year’s winter wave.  If there is any substance to the “running out of targets” theory of waves, Delta already seems to have run out of targets.

And finally, we’re aware that a winter wave is likely.  Governments, institutions, and individuals may choose not to act.  But we aren’t going to be taken by surprise.  I would like to think that a lot of entities will be more proactive this winter than they were last winter.

I, for one, have already stocked this winter’s supplies for my preferred humidifier (Post #895).  And I’ve managed to find a high-filtration mask that my daughter will enthusiastically wear full-time (see next post).  She had been a cloth-mask devotee before this, and would only wear the parentally-supplied N95s in high-risk situations.

As the phrase goes, history doesn’t repeat itself, but it rhymes.  Seems like Delta has worn out its welcome, meaning, infected the people who are apt to be infected by it.  Seems like an increasing share of the population is immune, and that share is getting larger.  The data show that it is clearly hard to spread Delta among the immunized.  And we’re not as naive as we were one year ago.  We know there’s a rhythm and a seasonality to these waves.

Maybe this time around things will be different.

The new graph and table for today

Same as the old graph, just starts later.  Same as the old table, just starts later.