Read on Twitter
OK, I have tried to model this more, and the bottom line is: without a test and trace system that absolutely jumps on rising cases when the number of cases is really low, the government has a really hard task. Brace for some ugly graphs: 1/ https://twitter.com/Gilesyb/status/1338752791913885698
This models a Spring surge, harsh measures in April-May that crush the virus infection down, and leads to loosening in June and July.
Basically, the R rate rises fast when lockdowns end in June - but you can hardly see it .... Table shows measures (low = restrictive) 2/
Basically, the R rate rises fast when lockdowns end in June - but you can hardly see it .... Table shows measures (low = restrictive) 2/
and so further lockdowns etc are *politically impossible* at the point they may stop it rising. Imagine - 3 months of almost no cases, not many deaths - but the exponential force is building, and then BANG 3/
Basically, it is easy to construct a model where the most rational time to throw in an extra 10% of Lockdown measures (in terms of reducing deaths) is August - but it means having to do it when it looks like the pandemic is *barely there*. 4/
Looking backwards at the real data now, having modelled how it works, it appears clearer that mid Sep was when to get really alarmed. And when did Patrick Vallance get really alarmed? MID SEP 5/
So, bottom line: it is really hard, and perhaps politically impossible, to impose actual measures when the R is undetectably high and the cases low - like it must have been in August at some point. But the moment the cases confirmed it - the excuses go away, IMHO.
And Patrick Vallance was right. /ends
