3 Comments

Great news about Stu!!!

Expand full comment

Have they learned nothing about models over the past two plus years?

Expand full comment

I think they do "set it and forget it"!

To me, I'm not sure why they created an unreported deaths model in the first place. It never was their strength -- I think they were trying to anticipate COVID death spikes, but the epidemiological models didn't perform very well, either (which are different from the weighting models I demonstrated in the most recent post).

I ignored all these models when it was obvious that not only did these models perform poorly, they weren't even adjusting parameters or structure. I tried to make sure I downloaded only actual reported numbers and threw out others' modeled results. (I do my own models for other purposes... and generally modeled a range of possible outcomes)

Expand full comment