Coronavirus days: fog of war

Things are getting somewhat confusing. In our own parochial UK... no wait, in our own parochial East, things aren't too bad; we're one of the lower regions. But in the UK as a whole, there's uncertainty. There are so many clamouring voices saying so many things, it's no wonder people don't really know what is going on. Some of the voices are genuinely trying to help, some are just looking for the oxygen of publicity, but the end result is unclarity.

I think the prime example of this is Covid-19: Growth in cases may be slowing in England:

The growth in cases of coronavirus may be slowing down, the largest study of the infection in England suggests. A team at Imperial College London analysed samples from 84,000 people chosen at random from across the country. They said the R number, the virus's reproduction number, appears to have fallen since measures including the "rule of six" were introduced. However, they warn cases are high, with one in every 200 people infected. The React study is highly influential, both due to its size and because it gives an up-to-date picture of how the virus is spreading. The last samples used in the analysis were collected as recently as Saturday. It was the previous React report that found infections were doubling every seven to eight days in late August and early September... Then the research group estimated the R number for their study - the average number of people each infected person is passing the virus on to - was 1.7. The latest analysis, of swab samples collected between 19 and 26 September, suggests the R number has fallen to about 1.1 - although the precise figure is uncertain.

This is from Dear Aunty Beeb, you can trust her even in times of war, and all that gumpf. However is it true? Probably not. Consider the pic, stolen from Oliver Johnson's Twat. It-was-1.7-it-is-now-1.1 is based on interpretation B. Whereas interpretation A seems more natural.  Apart from anything else, B is discontinuous, which is unphysical. Also I just don't trust their underlying "explanation" for the slowdown: that the Glorious Leader's "rule of six" has pulled down R. That may have made some difference, but not a lot; and the return-to-school and return-to-university has certainly pulled the numbers in the opposite direction; finger-in-the-air, I'd say those latter two will have made more difference. JA is barely able to believe that people are still falling for this stuff, and yet they do. People want the Bad Thing to go away.

[Update: alerted by Twatter, I bothered to read further down the article, and find However, Prof Oliver Johnson, from the University of Bristol, said the conclusion that cases were slowing down was "wrong and dangerous". And he doubts both the old and the new estimates of the R value. He said: "I suspect they were both wrong, and it was actually more like R=1.4 each time.". So not finding that the first time was a bit crap of me. But burying it so far down was even crapper of the Beeb.

More people not being dead impressed can be found at [E]xpert reaction to preprint with the latest interim data from the REACT-1 study on COVID-19 spread across England, but even there they lead with someone liking it.]

Also, version A is consistent with James Annan's daily-updated modelling, which looks vastly more sane than anything Imperial have been able to do. I admit that I did lose faith a little when deaths clearly fell below the curve towards the end of August, but happily the corpses have started stacking up since then, and my trust is restored.

The govt of course has not helped the aura of confusion by being a pack of incompetent clowns; but there's more blame to go round. The media have been irresponsible too, and not a few of my fellow citizens have been dumb enough to go out partying, the tossers¹.

It is conventional to compare Science in Covid and Global Warming. When Science delivers us a vaccine, all will be well let us hope, but at the moment Science isn't doing a brilliant job on Covid, except in a rather confused, muddling-along, ants-moving-a-leaf kind of way. Certainly in the UK we haven't managed to get any competent unified voice speaking sense. SAGE is too tied to the govt; and anyway doesn't seem to manage to be particularly sensible, and doesn't really speak in public. Unlike, say, the IPCC. Though the IPCC has the luxury of much longer timescales; and doesn't I think say much that is influential on the economics, instead grinding over long-solved problems in ever greater detail². 

Indeed the competence of (local?) govt seems to be more important than Science; though the latest FT Covid pix blow the idea of some Socialist Miracle; France and Spain now have deaths well over the UK; only the Squareheads are looking good.

Away from whingeing, via Paul, This Overlooked Variable Is the Key to the Pandemic from The Atlantic. Oversold, of course, but perhaps correct in that looking at the average too much hides important information.

Refs

* Today's random restrictions, by region - DailyMash
* Self-Help Is Like a Vaccine by Bryan Caplan
* Ridiculous Widespread Beliefs by DON BOUDREAUX and Expert Failure to Know
* CAN CARBON-NEGATIVE SOLAR CREMATION ECLIPSE THE ANTINOMY OF CASTE-NEUTRAL SUTTEE? 
* Jeez People, Get This Right - Timmy
* Adding to my JA Twit collection; shame he is so restrained.

* The Dunning-Kruger effect: Misunderstood, misrepresented, overused and … non-existent? Just stop using it!

* Opinion: The case for voting against presidential candidates by GEORGE LEEF. But, why only for presidential voting? I've advocated similar, but can't now find where. Related: Why Can't They Both Lose?

Notes

1. Anecdote: a friend of my daughter's is at St Andrews, now isolating in his household, because another member of the household has got Covid, due to going out and screwing around. This is irresponsible, but on a statistical level that's going to happen when young folk go away from home.

2. I exaggerate for effect, you understand.