I have a very basic question about the @IHME_UW model. To me, the cone of uncertainty is very odd here. It leaps to a maximum one day into the future and declines thereafter.

Any insight into why this is happening?

Another way to look at it.

Normally when you predict things, uncertainty increases into the future. Now, when you know the final state with certainty, it will turn around at some point. Uncertainty very rarely decreases monotonically into the future. Yet it does here.