"US life expectancy stalls due to cardiovascular disease, not drug deaths"
"‘Sushi parasites’ have increased 283-fold in past 40 years"

Not so "super" forecasting?

UPDATE: Thanks to a couple of commenters--thanks Albert, thanks Tim--I made a really stupid error. The forecasts are for March 31, 2021. So the post below--I'll leave it up as reminder to me not to be stupid--is premature, at the least.

A week ago I linked to short-term forecasts for the virus made by "superforecasters". (Superforecasters "qualified by being in the most accurate 2% of forecasters from a large-scale, government-funded series of forecasting tournaments that ran from 2011-2015 . . .").

Since the forecasts are for March 31 we can now evaluate how well they did. Most of the forecasts seem to have been made on March 13, but a few may have been made later. Here are the results (all data from the Johns Hopkins tabulations, here and here):

Cases, worldwide. Actual, 857,957; forecast as "most likely," "more than 53 million but less than 530 million".

Deaths, worldwide. Actual, 42,139; forecast as "most likely," "more than 800,000 but less than 8 million".

Cases, U.S. Actual, 188,547; forecast as "most likely," "more than 2.3 million but less than 23 million".

Deaths, U.S. Actual, 3873; forecast as "most likely," "more than 35,000 but less than 350,000".

While I have a healthy respects for the evil power of the jinx and while the forecasters may turn out to be, as they say on Wall Street, not wrong but simply early, for right now, they seem to have missed by an order of magnitude or more.

 

Comments