There was a lot of reporting on the inner workings of the IMHE model earlier this week but I didn’t see anybody mention that their data is also just kind of a mess? It erroneously says that there were zero deaths in Colorado on these three days & is undercounting the total by 50

separate from the data issues, their model is clearly continuing to evolve (which is fine). Notice the big change in the shaded projection range in the last two days, from a potential maximum peak of ~45 deaths per day to over 100