Ad hominem attacks are the quickest way to lose one’s own credibility, as they are obvious attempts to distract from one’s own untenable position.
Asking what someone’s training is, particularly when they are misinterpreting basic points, is not an ad hominem attack. And please link to your Google Scholar page or a CV listing your publications so anyone else who comes across this exchange can decide for themselves where your expertise lies.
Imagining a unrealistic scenario is meaningless at best and misleading at worst.
What is unrealistic about this scenario? The median detected IFR in Atlanta nursing homes (i.e., this is a ceiling because they might have missed infections) was 13%. I don’t think it’s really debatable at this point that individuals in nursing homes are at a much higher risk of dying when infected by COVID-19 than the general population. If anything, 100x is conservative. Further, given a fixed relative difference in risk (100x in my example), the specific IFRs don’t matter when calculating the overestimate multiple. This is a straightforward algebraic fact that I would expect an expert in statistics to notice, but I’ll leave it as an exercise for the reader to prove why that’s the case.
Your remaining points commit the exact errors I already described in my post and my original reply, namely: conflating empirical IFR estimates with the population IFR, conflating the total number dead with the IFR, and claiming Ioannidis was predicting 1% of the population would be infected (also, hindsight is 20/20, SARS clearly did not infect 25% of the population). Further, you commit other errors such as: neglecting the fact that the risk of COVID-19 roughly doubles every seven years, so your proportional assignment for the 45–54 age group is clearly biased upwards, comparing COVID-19 deaths from what is now effectively 1.5–2 seasons to a single flu season, assuming we have a reasonably accurate estimate of how many people get infected with the flu each year, and assuming the entirety of excess deaths should be attributed to COVID-19. Regarding the flu vs. COVID-19, here is what the CDC says:
312,597 reported COVID-19 deaths as of December 21st.
45 million symptomatic flu infections in the 2017–2018 season.
77 million symptomatic COVID-19 infections as of December 1st.
61,099 flu deaths in the 2017–2018 season.
312,597 * (45 / 77) = 182,687 (the number of COVID-19 deaths we would expect for the same number of symptomatic infections as the 2017–2018 flu season).
182,687 / 61,099 = 2.99
So the CDC’s own estimates suggests the empirical IFR for COVID-19 is on the order of 3x a recent flu season, and, again, that’s not accounting for the fact that nursing homes appear to be suffering higher infection rates with COVID-19 than the general population.
Most regrettably, you make no attempt to engage with what is the most important point made by Ioannidis and Bhattacharya (and Gupta, and Kulldorff, and Baral, etc.), which is that blanket lockdowns are (clearly) not effective at protecting the vulnerable and they are also incredibly harmful, so I will not be replying further.
