Games with Models, April 20 Data

Of course, because the trend of this weekend was a leveling-off of COVID-19 hospitalizations, it was tempting to think we’d already hit the peak, and today’s increase to 272 people in the hospital proves that wasn’t so.  Nonetheless, the increase was still less than my model had predicted.

(It bears repeating that these projections are based entirely on recent trends, so there isn’t any assessment of how viruses spread.)

For today, the model had projected 284 hospitalizations, and the actual number turned out to be 272.



Here’s how the one-day predictions stacked up to reality and where they’re going next:

  • Cases:
    • Projection for 4/20: 4,931
    • Actual for 4/20: 5,090
    • Projection for 4/21: 5,311
  • Hospitalizations:
    • Projection for 4/20: 284
    • Actual for 4/20: 272
    • Projection for 4/21: 296
  • Deaths:
    • Projection for 4/20: 157
    • Actual for 4/20: 155
    • Projection for 4/21: 162

Naturally, to the extent that cases can follow a different trend than hospitalizations and deaths, it’s great to see the latter two going down.  Increased testing would tend to mean we’re recording more cases leaning toward the less-serious side and would also tend to make the model overestimate for hospitalizations and deaths.

These trends move the projected peak of active cases to 4,082 tomorrow, with 1,067 Rhode Islanders’ having recovered.

  • No products in the cart.