Freddy Snowball **Last Nine Days UK**

1,543 - - - - 055 - - - - 3.56%

1,950 - - - - 071 - - - - 3.64%

2,626 - - - - 104 - - - - 3.96%

3,269 - - - - 144 - - - - 4.41%

3,983 - - - - 177 - - - - 4.44%

5,018 - - - - 233 - - - - 4.64%

5,683 - - - - 281 - - - - 4.94%

6,650 - - - - 335 - - - - 5.04%

8,077 - - - - 422 - - - - 5.22%

In 9 days the mortality rate has increased by almost 50%

We seem to heading the way Italy did, closing on 10%

(of confirmed cases)

This isn't really a useful statistic. First of all, there is a difference between a fatality rate and a mortality rate. Second, if you were to try to get a fatality rate, you should probably be comparing deaths with cases from 10+ days ago (as there is a gap between getting the illness and dying). Mostly, as we have restricted testing, these numbers won't make much sense for a while yet.

Which is why I said this rate will increase, probably reaching 9 or 10%

As we progress the Cases-Illnesses-Deaths "in the pipe" will become a smaller proportion

and the growing Mortality rate will slow down, slow down as it approaches (my estimate)

something between 9 and 10%

I'm aware that the current, rolling mortality rate isn't THE rate.

I have been saying so and pointing out that it is climbing as cases "mature"

I guess we could "guess at" an average infection to death time-scale (usually 14 days) and ask

How many Cases were there 14 days ago? and compare that to today's mortalities. Obviously

that will be rough, but over days the variation either way should cancel out. The total cases

will remain accurate as will the recoveries and deaths.

I don't have UK daily new cases before 9 days ago BUT

these numbers are the daily increase in cases x 9% and 10%

037 041 29th

061 068 30th

058 064 31st

064 071 1st April

093 104 2nd April

060 067 3rd April

087 097 4th April

128 143 5th April

They look way too low!

These (below) are the last 9 days' deaths which we would expect to be increasing

for the next 14 days, so taking the new cases for specific days and multiplying

by 9% or 10% don't appear to predict future mortality

16

33

40

37

56

48

54

87

if people reply "Well to pick 14 days is silly" I'd answer

that they are wrong.

For any one day, some might die, say in 10,11,12,13 days, a few in 15, 16, 17 days,

but, presuming mortalities are reasonably consistent this bandwidth would be

cancelled out by underlaps and overlaps from the days before and days after Day X

**Just seen that the average from onset of illness to death is 18.5 days**