It's quite possible there is some vulnerability/exposure bias early on in the outbreak, but it's almost impossible to imagine this could get the IFR (infection fatality ratio) below 0.5%, let alone down to 0.2%, as per below. Also, as I said to T2K before, he'll never provide any data or analysis supporting his 0.2% number - he just says stuff. And for those who don't read all my posts, I'll repeat the basics here. Look at NYC or NY State, where we actually have seroprevalence data on who has antibodies to the virus from past infection, meaning we actually can calculate reasonably accurate infection fatality ratios.
Being conservative, let's just look at deaths through 5/1, when the antibody testing was completed (in actuality deaths for at least the next week or two would've been from infections prior/up to 5/1). The antibody testing (shared on 5/2) showed 12.3% of NY (2.46MM of 20MM) with antibodies and 19.9% of NYC (1.67MM of 8.4MM) with antibodies, while as of 5/1, there were 24.1K fatalities in NY and 18.3K fatalities in NYC (using the Worldometers data). It's simple math to then calculate the IFR for NY of 1.0% and an IFR for NYC of 1.1%. And we also have a strong piece of corroborating data now, as Spain has ~5% with antibodies (46.7MM people, as of testing last week and 27.2K deaths at that time for an IFR of 1.2%.
The thought that NY's or Spain's IFR could be 0.2% eventually is unrealistic. If one assumes 60% of the population eventually becomes infected (herd immunity: most estimates on that range from 50-80%), then NY's IFR if nobody else dies between now and then would be 28,300 deaths (current)/12MM (60% of NY) or 0.25%, which is already greater than 0.2%. I do think it's likely the IFR will come down some from the 1.0-1.1% we're seeing in NY/NYC, due to early bias towards more of the most vulnerable being infected and perhaps other demographic reasons, which is why I've been saying I think the eventual IFR could be in the 0.5-1.0% range (but more likely towards the 1% side of that range). But let's just say 0.5-1.0% IFR for argument's sake and that's still horrible.
And if the IFR is 0.5-1.0% and ~60% eventually become infected in the US, which will happen without a vaccine and will happen over 6-24 months depending on interventions and population density (those things only affect transmission rate, which affects when 60% infections is achieved, not the endpoint reached), then the estimated US deaths is a simple calculation of 0.5-1.0% x 60% x 330MM which equals 0.82-1.65MM US deaths, eventually (assuming no vaccine or cure for this analysis).
Note that Worldometers in the link below calculates a NYC IFR of 1.4%, as they also count in "excess deaths" (as per the CDC analysis) and assume those are due to coronavirus. I'm only using the confirmed + probable deaths in my calculations, as these are more conservative. If one used only confirmed deaths for NY/NYC, the IFR would be about 0.8% and it's hard to imagine that dropping below 0.5%.
The only way 0.2% IFR is even close to in play is if somehow herd immunity is far less than 60% as a few have speculated or if a large percentage of people have "built-in" immunity to the virus without having antibodies for it yet (the cross reactivity I posted about last night); both of these are very unlikely though.
https://www.worldometers.info/coronavirus/coronavirus-death-rate/