Specifically, the Imperial College model with the "two million deaths" conclusion assumed an IFR of 0.9% (adjusted for age distribution) in every scenario. Maybe not quite a delusion, but the only IFR number they modeled is starting to look pretty close to one.
The derived IFR is still <1% (0.66%), but I'll admit it seems unjustifiably high given the inconsistent reporting out of China. It would have been better characterized, at least qualitatively, as somewhat pessimistic, given all the assumptions.
Ah right, the 0.9% is after adjusting for the age distribution of Great Britain:
> These estimates were corrected for non-uniform attack
rates by age and when applied to the GB population result in an IFR of 0.9% with 4.4% of infections hospitalised (Table 1).
