Cold Covid - part the seventeenth

Well, another professional heard from. And his piece reads like Constant Power^© applies. I certainly feel better, but don't expect a call from the Nobel Committee, as deserved as it may be.

Our models project that the United States is likely to document more Covid-19 cases in January than in any previous month of the pandemic, but a smaller fraction of those cases will require hospitalization. Whether hospitals experience more or less strain than they did in January 2021 will depend on case numbers and how severe they are. For example, if twice as many people become infected but these people are half as likely to be hospitalized, the demand for hospital beds would be the same. This calculus also applies to estimated deaths from the virus, as well as expected disruptions to the work force.

Alas, he doesn't show his work, math/stats behind his conclusions. Suffice to say, he supports the notion that estimating beyond the data is fraught with danger. A few weeks at most. Interesting that he skips the PI/CI differences, and simply analogizes to hurricane path predictions: cone of uncertainty. I like that.

[NOTE: The NYT tells me that I can link to 10 articles/per month without paywall block on you, dear reader. We'll see.]