Experts Think The U.S. COVID-19 Death Toll Will Hit 50,000 By The End Of April

[Igel]

30. NYC alone added a bunch of deaths

when they were running the calculations for this iteration.

And those deaths needed to be accurately distributed among previous days' numbers--but weren't.

Of course the projection for 60k is wrong. It's based on merging the actual data fit a standard sort of curve. But then the data collection process was altered on a given date, so all the data after that date couldn't be compared with the data before that date. The data that had been used was suddenly declared to be off by fairly large overall number and a quick fix added.

Then you have additional states adopting the new metric over a few days, so the damage isn't just on one day.



They revised it down twice for fairly obvious reasons. The "best guess" numbers were typically less than the daily numbers being reported. Why double down on wrong, when the model itself tries to use current data to adjust later projections. A lot of the daily numbers were wrong--but overall, more were under than over, so the curve was being flattened more than anticipated.