Games with Models, 6/4/20 Data


Not much was newsworthy in today’s COVID-19 report for Rhode Island, so for a little bit of a hook, I’ll note the strange headline, “Autopsy report shows [George] Floyd tested positive for coronavirus.”

The relevance of that finding (apparently dating to an illness in early April) isn’t clear, although without stretching the connection too far, it does provide a topical example of one of the big measurement questions of this pandemic.  Just as the numbers of hospitalizations in RI began to decrease, the state switched from reporting the numbers of patients in the hospital because of COVID-19 to reporting those who were in the hospital and happened to test positive for the disease.  Not surprisingly, the numbers went way up 41%.

As we watch the gradual descent of hospitalizations in our state, it’s important to remember that it’s actually reflecting many people who would have been in the hospital anyway.

Other than that, the news is more or less the same as it’s been. New cases are slowing down.  Active cases are decreasing.  New hospital admissions are still in the single digits.  And despite daily news media headlines about double-digit deaths, those are actually do to revisions of prior days.  As currently reported, only three of the past 13 days have had double-digit deaths.


(See here for my original methodology and here for a subsequent modification I made.)

Projections versus actuals (date of report).

  • Cases:
    • Projection for 6/4: 15,199
    • Actual for 6/4: 15,325
    • Projection for 6/5: 15,405
  • Hospitalizations:
    • Projection for 6/4: 178
    • Actual for 6/4: 185
    • Projection for 6/5: 177
  • Deaths:
    • Projections for 6/4: 736
    • Actual for 6/4: 756
    • Projection for 6/5: 760

  • Lou

    There was no truth to the rumor that Mr. Floyd had been to Burrillville recently.