I was thinking when I looked at gov.uk curve that the rate of decay was slowing and that therefore the R rate was rising towards 1.0 but thinking about it if you have 1000 cases at time = 0 and the virus is declining at a fixed rate then each time step is going to trigger a smaller fall. In other words a population rises exponentially and declines exponentially. I should have realised that. A flattening curve does not mean that we should be worried that something fundamental is happening it's just how populations grow and fall.
To illustrate what I mean here is a population of radioactive atoms that decays at a fixed rate. The curve flattens but the decay rate is fixed.
View attachment 10963
It feels illogical but is how this works and why I say in the reports - eg about Trafford - that once you get down to low numbers there is much further to rise than to fall. So it is easy for us to be deceived by numbers.
Really percentage falls week to week are the key.
Look at the England hospital data
23, 042 to 17, 694 to 14, 315 are the falls week to week in patient number Saturday 2 weeks ago to last Saturday to yesterday.
It is falling fast. That much is obvious. But it fell 5348 in week one and 3379 in week two.
Obviously it is slowing down and that looks like by quite a big drop.
But in fact it is only from 23.2% to 19.1% week to week in the term that matters - the proportion of the change not the raw number which tends to deceive us.
North West patient numbers in that same two weeks fell from 3292 to 2631 in the first Sat to Sat data and then to 2071 yesterday
These are falls of 661 first week and 560 last.
So obviously it is slowing?
Um no - actually its increased.
From a 20.1% fall to 21.3%
Raw numbers can easily fool our logic.
And the key thing to take from the above is that England hospital numbers - ALL regions - now including the North West - which was slower to get to this level than some - are falling at about 20% a week.
In mid winter in a pandemic that cannot be spun any way but good.
