Let's Crunch the Dead

330 million x 40% infection rate x 1% death rate = 1,320,000 dead.

330 million x 30% infection rate x 0.5% death rate = 495,000 dead.

One is more optimistic than the other, but is either realistic?

We just don't have enough data to know. Testing is ramping up, but so are deaths, and we don't know how far behind testing is to the spread of the virus among the population of the US. We know that "social isolation" has worked to-date in other countries, but we don't know how effective social isolation is in the US. And there's no hard cutoff date for the beginning of social isolation because different states have done different things, and the states that have acted the most still acted in stages over a couple of weeks.

My guess (on 2020-03-25) is the total directly attributable deaths in the US will be in the range of 20,000-30,000 - if the social isolation holds instead of being short-circuited by the stock-pumper-in-chief. But that is completely a guess.

For now, we should treat all estimates - high or low - with skepticism.