I think you're putting forward a weirdly deceptive view that the two possibilities (under-diagnosis and long-dying-process) are exclusive. It seems extremely likely both factors are involved in current statistics - nothing in one excludes the other.
A nation that's systematically testing and finding all the cases will start with a lower death rate and watch it creep upward - but for the manageable number of cases they will see as they get patient quarrantine. A nation or region that's not systematically testing will have a high death and high recovery rate as they only see cases after the fact but then they'll see all cases spike, their health system overwhelmed and things in a mess (ie, Italy now, Hubei earlier).
I'm heartened that the US is now finding more cases rather than just seeing more death but clearly the US slacking on testing and surveillance.
>I'd say nobody is going to shut down whole megalopolis cities if it was just a more severe flu.
I mean a severe flu by a factor of five or so is basically what this is, it's just that this in and of itself is really, really bad because nobody has the capacity to handle that many cases.
Neither does this virus, typically. However some people end up having severe complications, as do people with the flu which is why the latter causes a few ten thousand deaths each year in the US, and half a million worldwide.
The issue here really is a matter of quantity. Covid-19 is not a mysterious super disease, but it is dangerous enough to overstress an unprepared healthcare system.
Also, if the entire population of Northern Italy was infected and mortality rate was .5%, you would be seeing 13K deaths (which isn't impossible given the situation but far from 6-700 deaths so far).
Which is to say, you're right that infection on a large scale is far more problematic than infection on a small scale but I don't think infection on a large necessarily contradicts the Diamond Princess results. A lot of people can die and it is social catastrophe even with percent of the entire population.
But I want to be clear that nothing I said above is anything like a solid prediction, just a pure back-of-the-envelope rough comparison. There's nothing I'm going to predict here.
That's one possibility; the other is what is happening in Italy right now, i.e. initial 2 weeks of infections are done and now the severe cases remain and pass away. Like in South Korea 98% cases are in progress, so the outcome is not known yet and making any stats out of those is premature (i.e. too optimistic). I'd say nobody is going to shut down whole megalopolis cities if it was just a more severe flu...