Some numbers which mean nothing
May 9, 2020•387 words
[This is almost entirely wrong: pay no attention to it.]
It is obviously complicated to know what the important statistics are for COVID-19, not least because the situation is rapidly changing, there are exponential processes involved, different countries are not always counting things the same way and are not all as good as each other at counting things at all.
But I do know what the wrong statistics are: any number which is of the form 'n per country' is almost always meaningless and often actively misleading.
To give an example: on the 9th of May 2020, JHS CSSE is reporting that
- 78,084 people have died in the US;
- 31,662 people have died in the UK.
So, obviously, things are a lot worse in the US than they are in the UK, right? In particular the chance of you having died of CV19 are much higher in the US than the UK, right?
No, they're not:
| where | deaths | population | deaths per million |
|---|---|---|---|
| US | 78,084 | 324,500,000 | 241 |
| UK | 31,662 | 66,200,000 | 478 |
On the 9th of May 2019, the chance of you having died of CV19 if you lived in the UK was about double the chance of you having died if you lived in the US.
That doesn't mean the US is doing better than the UK: the UK may be run by clowns but the US is run by idiot clowns. What it probably means is that the US is much larger and much more thinly populated so the virus will take longer to do its work there. Almost certainly the death rate in the US is going to be awful: it just isn't yet.
But a country which has a population twice as large as another country is, other things being equal, going to have numbers of deaths which are twice as large at any given time. Numbers which are quoted per country mean nothing, because countries don't exist, other than in the heads of humans: the only numbers which mean anything are rates per person or per million people. It doesn't matter how many people are dying in a given country: what matters is what proportion of your friends and family will die, or how likely you are to die.
Those are the numbers which you should be watching.