Spreadsheet Phil

Spreadsheet Phil was Philip Hammond's nickname amongst his staff at the Treasury. I've purloined his monicker for the day and put together one of my punk spreadsheets to try to understand the NHS data for Covid 19 deaths by day of occurrence. As I expected the data comes through slowly and a bit erratically. If you take the raw data from the NHS England website*  you can understand why the experts are saying we're not at the peak yet and why they'll need to be some way past it to be sure.

Taking yesterday's data as an example, for England 740 deaths were announced of which 151 took place on 15 April, 314 on the 14th, 122 on the 13th and the remainder on earlier dates stretching back to early March.

Here's an extract from my cumulative table showing how the deaths we knew about for a particular day (horizontal heading) tot up over the following days (left side column):

APRIL	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15
2	84	0	0
3	263	87	0
4	397	286	99	0
5	448	400	280	97	0
6	462	425	350	285	69	0
7	495	462	440	464	341	81	0
8	512	485	503	557	503	387	135	0
9	531	515	534	604	560	487	419	140	0
10	544	530	564	643	603	557	573	496	117	0
11	553	543	581	671	635	606	634	657	442	115	0
12	555	544	584	675	640	616	660	709	562	423	121	0
13	557	549	590	680	645	625	673	737	611	516	443	118	0
14	558	554	599	693	655	631	692	750	636	579	575	437	122	0
15	558	557	599	698	657	638	699	771	653	606	637	540	399	113	0
16	559	558	604	699	660	643	704	782	674	624	660	589	521	427	151
17	559	558	604	699	660	643	704	782	674	624	660	589	521	427	151

It looks to me like it takes about 4 days for the deaths on a particular day to be known within 20% and more like 8 days for the data to become fairly settled.  Given that we'll want to see the figures falling for several days in a row that means we'll be more like 10 days or more after the event before the stattos are sure that we're past the peak.

And then the relaxations that make sense depend on whether we got R down to 0.9 (so not much scope) or 0.5 (quite a bit of scope), which Whitty hinted at on Thursday.

What is striking is that (if I've understood the NHS data and manipulated it correctly) the total deaths by day (bottom row of figures, not reliable after about April 10) aren't anything like as peaky as most folk seem to think. You can see why the experts aren't sure we're at the peak yet.

I also think that, from the comments of some journalists, they expect the cumulative deaths curve to flatten and then turn down and decline. Which ain't exactly going to happen unless the Easter resurrection becomes a widespread phenomenon.

They'd be better leaving the politicians to concentrate on consulting their experts and getting on with the work that needs to be done.

Fat chance but I think I'll do that for now.

* https://www.england.nhs.uk/statistics/statistical-work-areas/covid-19-daily-deaths/