Another Week, Another Nonsense Graph from the Government!

So another week passes by with yet more truly awful science from the UK Government

(Yes, it may well be true that the real Covid-19 scientists are working well – and properly – in the background but it’s becoming increasing obvious that ministers in front of the camera either don’t – or won’t – understand what they’re being told.)

This probably sums it up as well as anything could:  https://www.thepoke.co.uk/2020/05/10/boris-johnsons-covid-19-update-speech-went-down-like-a-cough-in-a-lift/ but the focus of this particular post is that graph above.

What the hell is that?

There are all sorts of nice images dropped in there but the most important element of any graph is the data it’s supposed to represent.  Often this is a function of something but at the very least it’s some sort of plot, or line, or curve.

Well, we’ve got a curve on that graph.

But what is it?

It’s difficult, of course because neither axis (x or y) is labelled properly and there are no meaningful values anywhere but let’s try …

The curve rises steeply, with the area under it coloured red (so presumably that’s bad?), then flattens out and drops less steeply (blue: better?).  The dividing line between red and blue is (sort of) labelled ‘R=1’, although – unhelpfully – not adjacent to either axis.  R is obviously this critical infection rate the Government has been guessing at recently (but with no real idea of what it might be because they haven’t been testing at sufficient levels to feedback into – and correct – their models).

So is the curve R?

No, it can’t be because that curve starts from zero, goes up to some peak, then falls.  If R=1 represents the peak then presumably it’s saying the maximum R value we’ve ever had has been 1, which is almost certainly wrong.  Best guesses at R values of the past month have ranged between 2 and 4.  So that curve can’t be R and R=1 can’t be the value on the y-axis.

Perhaps it’s the rate of increase of R?  So the up-slope means R is rising and the down-slope, falling?  No, but then the point labelled R=1 is actually where the curve is flat so that’s R=0.  That doesn’t work.

So is R=1 measured on the x-axis?  In which case it’s some point in time?

Well it could be but there are two problems with this.  Firstly, we still don’t know how far along that x-axis we are.

Are we there yet?

And secondly, we’re back to square one …

What the hell is that curve?

It’s NOT R.  it’s not labelled.  It’s got not data.

We’re going to have to guess!

This really isn’t the way to present statistics!  Although hardly the first time the government’s done that badly recently.

Seriously, what is it?  Is it perhaps the number of deaths that have occurred?  Could be but the slope up to the peak doesn’t look right.

Is it the number of infections?  Perhaps but no-one has any idea what that might be because we haven’t been testing.

So, we’d integrate this unknown function between two indeterminate points to calculate the … what exactly?

Or is the point actually to tell the public, ‘this is all too hard for you to understand’ … because it’s bloody well shouldn’t be!

Or is just something that’s been made up on the spur of the moment to look half-decent like everything else this dreadful government has produced so far?