Thread by @AdamJKucharski

Any writer who claims 'UK models always overestimate COVID numbers' is either lazy or trying to mislead their readers. A few illustrations... 1/

A surprisingly under-reported example is early UK scenario models (although @anthonybmasters & @d_spiegel correctly flag it in 'Covid by Numbers'). Some point estimates from initial 'recurrent lockdown' scenarios were lower than what actually occurred:

https://twitter.com/AdamJKucharski/status/1324836700133105665?s=20

2/

Then there’s the July 2020 report from @acmedsci, which used Imperial modelling. It was designed to be a plausible worst case scenario, not a forecast, but ended up strikingly close to reality (with Alpha increasing risk & vaccines decreasing it): 3/

https://twitter.com/AdamJKucharski/status/1347490849836433410?s=20

Estimates of increased transmissibility of Alpha and Delta also came from modelling, which implied they'd become dominant across Europe. Despite a lot of criticism at the time (and lots of abusive messages in my inbox), this is exactly what happened:

https://twitter.com/AdamJKucharski/status/1393253535584821252?s=20

3/

Also, July 2021 reopening scenarios. There's been much focus on specific epidemic curves (often stripping away uncertainty intervals & full range of scenarios), but good faith readers can compare scenarios with subsequent known values for mobility, vaccine effectiveness etc...4/

E.g. Below scenarios from LSHTM model, with range of assumptions about mobility and vaccine effectiveness (gov.uk/government/pub…). In reality there have been around 16k COVID deaths in England since 1st July, and around 120k admissions. 5/

It’s important to have critiques of modelling (e.g. what's Q being addressed? What unknowns make most difference to results?) But if an article suggests models always cherry pick overly high numbers, bear in mind it's the writer who is really doing the cherry picking. 6/6