In Spring 2020 a novel coronavirus swept across the world: novel, but related to other viruses. In the UK, unknown at the time, around 50% of the population were already immune. The evidence for this is unequivocal and arose due to prior infection by common cold-causing coronaviruses (of which four are endemic). This prior immunity has been confirmed around the world by top cellular immunologists. There is even a very recent paper from Public Health England on the topic of prior immunity and a wealth of other evidence from studies on memory T-cells, studies on household transmission and on antibodies.
Because of the extent of the prior immunity, and as a result of heterogeneity of contacts, once only a low percentage of the population, perhaps as low as 10-20% had been infected, “herd immunity” was established. This is why daily deaths, which were rising exponentially, turned abruptly and began to fall, uninterrupted by street protests, the return to work, the reopening of pubs and crowded beaches during the summer.
In Spring, however, this virus did kill or hasten the end for approximately 40,000 vulnerable people, who were mostly old (median age 83, which is longer than that cohort’s life expectancy when born) and many of whom had multiple other medical conditions. There were some rare and very unfortunate younger people who also died, but age is clearly the strongest risk factor.
But due to extraordinary errors in modelling created by unaccountable academics at Imperial College, the country was told to expect over a half a million deaths. Three Nobel prize-winning scientists wrote to that modelling team in February correcting their errors. This was done confidentially. This expert, third-party estimate was remarkably accurate – it predicted that there would be a total of 40k deaths from COVID-19. I believe this is in fact correct and is what has happened. While I have no proficiency in modelling, I can distinguish predictions that are biological plausible from those which are literally incredible. When inputs to a model are wrong or missing, their outputs cannot be trusted. The Imperial model made the extreme assumption that there was zero prior immunity in the population or social contact heterogeneity.