Thread by @KennyMathieson

So, arguably there are two ways you could calculate the doubling time for #Omicron

1. Doubling Time of Total UK cases

and

2 Doubling Time of Daily UK case totals

I’ve just calculated both, using the latest data.

The values are 1.43 and 1.66 days respectively. 😬

Last week, there was much talk about a doubling time of 2-3 days.

I knew nothing about how to calculate this value until Saturday evening, but it’s a simple formula.

The 3 day estimate is (and was) nonsense, unsupported by the data.

It looks to have arisen from a basic error.

The calculation itself is very simple (assuming you know basic algebra) and is outlined in this thread from yesterday.

https://twitter.com/kennymathieson/status/1470093276879458309?s=21

What was the mistake that resulted in a value of 3 days for doubling time?

Once you calculate the Growth Factor, you need to divide that value into ln (2), where ln = the logarithm

The recent Growth Factor was about 0.35

ln (2) = 0.693

This results in 2 days doubling time.

So how did a value of 3 days doubling time come to be quoted?

Schoolboy/schoolgirl error.

Instead of dividing into ln(2), they divided into 1, resulting in a valur of just under 3 days.

Why am I confident about this?

Well, on Saturday night I spotted a post by a prominent poster on all things Covid, who had made just this (honest) mistake.

They acknowledged their error and posted a correction to 2 days (at that time).

The troubling thing is, given the 2-3 days doubling time estimate achieved prominence last week, that range is still being widely reported in the media and the value is now definitely well under 2 days. 😬

3 days is VERY bad.

Around 1.5 days is catastrophic. 😱