13.5.20
THEY'RE STARTING TO ASK THE RIGHT QUESTIONS.
Senator Rand Paul has a medical degree, and he's going after the simulations that scared public officials into locking down the sick and the healthy alike. "At this point, predictive models have proven to be nearly useless. And Senator Paul hit the head on that nail with a sledgehammer." That part of his questioning of contagion coordinator Anthony Fauci, M.D., has mostly been lost in the usual media foofaraw over personalities and motives. It's easier to posture and preen than to dip into the details of heterogeneous agents let alone higher-order differential equations. And yet, perhaps the after-action review will ask more questions along the lines of "Paul said the models have gotten more wrong than they have gotten right" once we're pretty sure we've pretty much finished slogging through hell.
Subscribe to:
Post Comments (Atom)
2 comments:
Disagree Steve. Dr. Paul also has a model. It's a mental one though. That doesn't make it worse, just less explicit. But it is rather one-sided for Dr. Paul to use his model to criticize those of others, without the predictions of his own model being paraded around for others to vilify.
And, in no way should this be construed as support for the Seattle models over any others ... just acknowledging that this stuff is hard.
I wrote "beginning to" deliberately. The representative's mental model might be as simple as "look at the anomaly." That is, he could have shown a late February forecast of stress on hospitals without and with mitigation, or perhaps several, say, for the United States, Arizona, Iowa, and Minnesota; similar forecasts for mid-March when the initial executive orders (which in Illinois was limited to two weeks ...); then observed cases for mid-April and mid-May, contrasted with the Seattle forecasts and others. Observed results that far off from quarterly forecasts get business gurus fired. Calling the policy maker to account for making a recommendation based on a bad forecast, and asking that policy maker to explain the anomaly, or to explain why the policy goes on despite the anomaly, is what accountability looks like.
Note that I haven't even gotten into the hard stuff, such as those heterogeneous population simulations (some of which include calibrations to previous outbreaks of respiratory syndromes, malaria, and Ebola) that I keep making reference to.
Post a Comment