STATISTICAL INFERENCE IS HARD.

Marginal Revolution's Tyler Cowen reports on some off-the-web conversation with an epidemiologist.  The practitioner is out of the academy, and understands what the academicians are up against.

1. Data is limited and often inaccurate in the critical forecasting window, leading to very large confidence bands for predictions
2. Unless the disease has been seen before, the underlying dynamics may be sufficiently vague to make your predictions totally useless if you do not correctly specify the model structure
3. Modeling is secondary to the governmental response (e.g., effective contact tracing) and individual action (e.g., social distancing, wearing masks)

The good news is public conversation appears to be considering how best to proceed to restoring some sort of expanded human interaction: masks and taking temperatures and what have you, rather than closing the paint aisle at the home improvement stores.

That might be a good thing, as epidemiologists are like any other discipline, good at what they do but not necessarily sensitive to what might be going on in the rest of life.  "Realistically though, there’s a dominant strain of thought that the economic costs of an action are secondary to stopping the spread of an epidemic. To summarize the [disciplinary social welfare function]: damn the torpedoes, full steam ahead!" It's more like "all back emergency," but you get the idea.