The striking thing about this survey of infectious disease experts' predictions is not that the predictions are dire, though some are, but how much variation there is. And if these people don't know what's going to happen, that means no one does.https://fivethirtyeight.com/features/infectious-disease-experts-dont-know-how-bad-the-coronavirus-is-going-to-get-either/ …
-
Infectious disease and statistical inference (or probabilistically predicting the future) are drastically different skills. I see no reason to expect these professionals, who are probably great at their job, to be able to predict accurately.
Interesting point. I wonder if that's true. It may not be, because experts on infectious diseases have a lot of data about the shapes of past epidemics, and it's the shape of the epidemic you're predicting, not simply x random future event.
-
-
Formal models, like epidemiological spread models, tend to outperform other methods in geopolitical forecasting
@PTetlock, but unsure how much correlation that will have with COVID forecastingpic.twitter.com/OTWHsIyCei
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
There's so much that can be done that a model, especially one targeted to "ops normal" living, is not going to account for such as the impact of isolation, whether voluntary or enforced, which is varied by nation and region and age and stubbornness and (willful) ignorance.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I’d expect predictive modeling to be a core competency required to study infectious diseases
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Considering the novel resilience of covid19 both in air and on external surfaces, I doubt past epidemiological models could be accurate.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Front line experts already know how to bend the curve beyond what western media is focusing on. A happy exceptionhttps://www.wsj.com/articles/these-drugs-are-helping-our-coronavirus-patients-11584899438 …
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
In my experience, don't talk epidemics with a virologist/microbiologist. They're great at what they do (churning papers), but generally not smart at all regarding epidemic maths.
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I think their models are pretty good in a cause-and-effect sense but that there are huge unknowables in parameters that describe people’s behavior. Certainly a huge spread of behavior in the US by geo, age, party…to put in “All models are wrong but some are useful.”—George Box
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
The article does not cite whom they surveyed. If I can't review a person's qualifications and publication history then I do not assume they know what they are talking about. And even then, I still want to understand the rationale for their opinion.
-
Geography also important. Actual experience on the ground will vary. An expert in Seattle may bias differently than one in Fargo even if looking at similar hard data.
End of conversation
New conversation -
-
-
I can tell you that shapes of past epidemics is not in medical school curricula and certainly not something that doctor's walk around discussing
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.