I explicitly linked to the CDC data in my original post on this, where I used the 7 most recent years for which complete (non-preliminary) data is available (2010/11 - 2016/17). The number of flu deaths was about the same or lower than the current number of COVID-19 deaths in 2 out of 7 of those seasons (2011/12: 12,000; 2015/16: 23,000). That makes the current COVID-19 numbers "pretty typical" for a flu season in my book -- certainly not atypical. And what the hell? I cited the actual numbers in my original post, so if I was trying to mislead anybody, that would be a pretty stupid way to do it.
Where are you getting this from? Could you link to the CDC site you are using? The de-facto standard for COVID-19 stats is the JHU collection, available at
COVID-19/csse_covid_19_data at master * CSSEGISandData/COVID-19 * GitHub
I just pulled the latest data from this, in which the maximum number of daily deaths is 2108 occurring on April 10. The Worldometer data set shows a peak in daily deaths on the same date, at 2035 deaths. And, of course, the apparent drop in the two later days is almost certainly due to incomplete reporting in the most recent data. So I don't see any evidence at all that the peak was last week. Do you? If there is something hopeful, it's that the number of new daily *cases* may be plateauing, and those should lead the number of deaths.
First, Sweden vs Norway shows that interventions do something. You quoted the Swedish daily deaths -- the highest number there (114) is only slightly lower than the *total* number of deaths in Norway so far (134). (Incidentally, I'd be extremely wary of trusting any data from the last 3 days or so, just because they always get updated.) Second, everybody keeps saying "the models were wrong" -- but that just shows that nobody bothers actually reading the papers. I did. The Imperial College paper from which the 2.2M estimate for an unchecked outbreak came from, also *explicitly* presented a wide variety of scenarios *with interventions*. And the numbers for those scenarios are pretty well aligned with those we're seeing. You can easily check this for yourself. Third, the IFR estimate spacediver cited from German data (0.4%) is in line with other estimates. In particular, random testing in Iceland shows that about 0.3% to 0.8% of their population are infected, i.e. between about 1000 and 3000 people. They have had 8 deaths so far in Iceland, suggesting an IFR between 0.3% and 0.8%. Small numbers, but consistent. Given a reasonably well established IFR, and the fact that herd immunity happens at ~70% infected, what logic would lets you assume you would NOT have a large number of deaths?
Just as a thought experiment, suppose that some hypothetical disease was really as bad or worse than people like me say COVID-19 is. Epidemiologists tell you that you can dramatically reduce the number of deaths by doing something unpopular. Are you saying you should *never* do because you couldn't ever prove that it was your unpopular action that reduced the number of deaths? I'm a huge fan of randomized controlled trials, but this doesn't seem a good time or place for one.