One grim way to look at the mortality rate: Due to incubation + hospitalization + ICU/ventilator period, divide the death toll *today* into the number of positive cases *10 days ago*.

U.S. Death Toll as of today: 20,600
U.S. Confirmed cases 10 days ago: 245K
20.6K / 245K = 8.4%.

U.S. death toll as of 4/10: 18,700
U.S. confirmed cases as of 3/31: 188,500
18.7K / 188.5K = 9.9%

death toll 4/09: 16.7K
confirmed 3/30: 163.8K
16.7 / 163.8 = 10.2%

death toll 4/08: 14.8K
confirmed 3/29: 144.3K
14.8K / 144.3K = 10.3%

OK, here's what it looks like since the U.S. broke 1,000 deaths...the "mortality rate" using this methodology has been *dropping* steadily as *testing* has ramped up.

If you divide todays' death total into today's confirmed cases, it's 3.9%.
If you divide today's death total into 10 days ago cases, it's 9.5%.
If you project ahead 10 days from now, you'd get ~51,000 dead, but it'll likely drop somewhat more as testing continues to ramp up.

I should note that @bjdickmayhew is the one who first reminded me of this seemingly-obvious-in-hindsight point.

. @bjdickmayhew recommended going out even further to more like 15 days, and perhaps I should, but there’s so many other factors like false positives/negatives, the same person being tested twice, and of course testing ramping up I decided to stick with 10 days for now.

Still, IF the death toll reaches around 51,000 by 4/22, I’ll know I have it pretty close. We’re at 21K so that’s be around 3,000/day on a average. :(