Why does anybody treat Neil Ferguson as an expert?
If we are to believe information on the Internet his previous predictions turned out to be very wrong...
Bird flu: Ferguson predicted 200 million deaths; the eventual number was 282
Swine flu: Ferguson predicted 65,000 deaths; the eventual number was 457
...to name but two.
"Modelling" sounds very clever and earnest but it is clearly not the truth.
Six questions that Neil Ferguson should be asked | The Spectator
Looks like fair comment.
Digging a bit more:
Ferguson et al should be more careful with what they say.
In the document which does have figures:
" There currently remains no strong evidence that Omicron infections are either more or less severe than Delta infections."
Bull.
Why not look at the best evidence there was - South Africa? Percentages going into hospital (low), length of stay...(short), deaths (none)?
This was published Dec 23rd but it refers to stuff we knew well before the gov statement.
https://theconversation.com/south-a...t-tells-us-about-how-deadly-omicron-is-174178
It's balanced, it includes:
"However, most patients had, at worst, mild symptoms, .... These observations substantially differ from the previous waves, including those attributed to the delta variant."
"most hospitalised patients were
unvaccinated."
"South Africa’s omicron wave experiences may follow
very similar patterns in other countries."
Cases in
SA rose by
2x in 10 days from December 1 2021 to December 21 2021
(calculated from their figs, 1.07x per day)
"In the
UK, from December 1 2021 to December 21 2021, cases of COVID per million population have risen
from 634 to 1,280 (a 101% rise),"
(that's not doubling in 2-3 days, that's 2x in 20 days ! It should have doubled several times by the time of the gov report).
(Now we're clear of delta affecting the numbers, omicron cases rates have doubled in the last
13 days (
to Dec 28, https://ourworldindata.org/covid-cases) 839 to 1670 )
"Modelling" sounds very clever and earnest but it is clearly not the truth."
Modelling is fine as long as the inputs are sensible and the results are used sensibly. If you know the people are likely to misuse them then it's irresponsible not to express them carefully.
Modelling should include your best estimate. Experience is a valid input. If there are confidence limits on figures then great, if not then you use your noodle and say so. You use words like "we expect"!
"The team" should have used their model to reproduce the SA results so they could see what the best fit was, not just ignore things which could be different. Assuming a constant 2-3 day doubling rate is stupid when you know there are constraints on it, and you
don't see a 2-3 day doublng rate except for a very short period when the virus arrives. Look athe numbers (left side) here:Fig 1
https://assets.publishing.service.g...44331/20211230_OS__Omicron_Daily_Overview.pdf
Conclusion - what we're told is misleading and has been for weeks. It's not simply a case of "We don't know"; and it's stupid to say that - while scary figures are published..
If you always accounted for the worst possible scenario you wouldn't leave the house. That is NOT the best way to carry on!
