We Should Ask the Modellers to Prove their Models Can Accurately Predict the Future Before Following their Advice Again

We’re publishing a guest post today by Dr. David Livermore, Professor of Medical Microbiology at the University of East Anglia, arguing that the SAGE modellers should be asked to prove their bona fides before we follow their advice again. He has come up with a simple test: build a model that can accurately predict the course of the 1889-94 ‘Russian Flu’ pandemic based on data available in 1891. If they can’t – if there are simply too many variables – then we shouldn’t base policy decisions on their predictions of how the current Covid epidemic is likely to play out in England. Here is an extract in which he lays out his rationale for setting this challenge:

Since Freedom Day in July, we’ve had predictions from Prof Ferguson alone that “cases would hit 100,000 per day” and then that the pandemic would be “over by October”. More recently we’ve had a boast that while suppression policies had their “downsides” they had proved “very effective” at reducing mortality over the last two years. Then, this week, when the pandemic was supposed to be over, Prof F has warned that Plan B, with masks, vaccine passports and WFH will soon be needed. All this strikes me as ad hoc guesswork based on short term trends, not modelling at all.

Closures and restrictions, if enforced, cause untold damage to real lives and real livelihoods. So, it is time to call the modellers out and to set a test, before they have the chance to inflict more harm. Medics, nurses, physios, medical lab scientists, and clinical scientists all must demonstrate competency to be licensed. There is no such regulation for modellers. Yet their advice, adopted by the Government, affects millions whereas a bad doctor can only harm one patient at a time.

So, in the absence of regulation, let me set a challenge.

This is an excellent piece from one of the country’s leading experts on infectious diseases. Very much worth reading in full.

As David says, if any modellers want to take up his challenge I’d be happy to publish the results in the Daily Sceptic.

Notify of
Newest Most Voted
Inline Feedbacks
View all comments
Would love your thoughts, please comment.x