COVID-19 June 1 Through Sept. 22

This is kinda weird, kinda not. This is just a relative comparison of the graphs on the three categories over that three-month-three-week period. The rules of thumb are cases precede hospitalizations by one week and hospitalizations precede deaths by two to three weeks. The graphs should be right-shifted, one week hospitalizations from cases, two-three weeks deaths from hospitalizations, 3-4 weeks from case to death.

This is the cases graph between the two dates:

Hospitalizations

Deaths

Now granted, there is reporting lag and we Americans can't count anyhow but you can see there is no apparent right shift between C's and H's. There is a very accurate right shift in D's from H's, not from cases. 

On July 5, 7-day average daily cases bottomed out at 10,608.You'd expect hospitalizations to have bottomed out on July 12. But they didn't. H's bottomed before cases bottomed out, which is like impossible. H's low was 16,790 on June 29. On July 12 (expected crater) H's had already begun their climb and stood at 19,910.  

Deaths' low, 175, was July 7, two days after cases' low. Pause. 175. Know what our average was for the week ending Sept. 22? 2,075. Unpause. Given a three-four week cases-death interval deaths should have bottomed out July 26-Aug. 2. July 26 deaths were 275 ave/day. Not ridiculously far off their actual low. On Aug. 2 though deaths were at 341 ave/day, almost double their actual low twenty-six days earlier, and deaths were already on their climb up which, with the exception of one little hiccup, continues to present. 

Deaths line up better from H's actual low on June 29. Given a two-three week H->D interval D's should have cratered July 5-12. Deaths were 186 ave/daily July 5, very accurate to the low end of the rule of thumb and 258 July 12.

Deaths continued rise is deeply troubling when cases and especially hospitalizations have declined for many consecutive iterations of the 7-day average and the 14-day change. You look at the deaths graph in isolation and you conclude, “Cases and H’s are UP!”, and it’s not so. Eyeballing the graphs cases are about where they were Aug. 1; H’s look to be at the level of mid-August and both categories have been down for all or nearly all of Sept. We’re now well into the third week of Sept. By either the three to four week case-to-death interval or the seemingly more accurate two to three week H-D interval we should have seen deaths drop before now and continue to drop and we’re nor seeing that.