I understand how important new deaths are in deciphering all this. I use the figures Wikipedia posts at:
https://en.wikipedia.org/wiki/Template:2019–20_coronavirus_pandemic_data/United_States_medical_cases
It accumulates daily data on deaths and new infections.
And, yes, this figure is very solid. Trying to use other solid data and then using other information that is out there but admittedly inaccurate, sort it and balance it, might be used as a way to predict the future. Pretty standard methodology.
For me, as a retiree, all I want to know is “where we are”, i get bored pretty easy after that. Induction, exponential, and recovery. I am not really interested in what the numbers actually are just on which date do those events occur. There are a million questions to ask and smart people a plenty to answer them. “Not my table’ as we used to say when I was a Resident.
I have been looking at worldometers.