C19 Update

Observations in several countries past their peak indicate a broad top and slow decline, suggesting a lognormal fit would better represent the data than the narrow peak, symmetric logistic fit given here. I'm working on lognormal fit, but nothing worth showing yet. Oh well, maybe tomorrow....


Spoke too soon, got the lognormal fit to work, but in fact it is a far worse fit to the data so far than logistic. Is this because on world data we do not yet have any points past the peak? Maybe after world peak the lognormal will become the better solution. My logistic fit suggests world peak in the next 2-3 days. We shall see.

Each day I will post best fit first (logistic for now) and the other below (lognormal for now).

Logistic fit

Lognormal fit