Maggie's Farm

Re: Imperial College Model — NOW CONFIRMED AS A COMPLETE FRAUD

It is not a fraud.

So a week ago I was reading tweets from people like John Carmack who had access to the Imperial College/Fergeson model code before its clean up and release on GitHub, and who were pleasantly surprised at how good it was.

Now I see from the usual suspects a report that blasts it, but only at sites where I know from experience that the authors will often use "motivated reasoning"

I was myself looking into the history of this group's modeling, as I had been affected as a UK resident from Government following Neil Fergeson's advice several times now over the past few decades - from the vCJD/BSE crisis onwards.

I was trying to write a post about how this guy has quite a history of being wrong, but when I actually looked back at what was published at the time all I can criticize is the media malpractice in reporting what his model-based studies said.

What I've come to realize is that the reports from his research group always take the form something like:


This gets reported on the BBC as:


Then, as there are enough UK cases for there to be better data from UK health authorities, they release a new report which is a smaller range, a range that does in fact include the number of deaths which actually result.

There's a ton of misreporting in this epidemic about the Imperial College model, e.g. about which big numbers were results accounting for social distancing. But even as I've dug though the "game of telephone" to get all the way back to what it actually predicted, I am still very much not a fan. So don't take my defense of it against poor reportage as any sort of endorsement of the model or the people behind it.