Scrutable

lpm wrote: ↑

Fri Aug 14, 2020 2:20 pm

Ah, some good news. The government's official UK R number is unchanged at 0.8 to 1.0. Cases are falling at -4% to -1% per day.




IMPORTANT GOVERNMENT NOTICE:
DO NOT LOOK AT ANY CHARTS OF CASES. ANY UPWARDS LINE ON CHARTS IS A STATISTICAL ARTEFACT. CASES ARE FALLING NOT RISING.

Covid-19 app by Kings College and Zoe has had daily cases falling for the last two weeks. c2200 then c1800 and today c1400.

I wonder if mask wearing is starting to have an effect?

discovolante wrote: ↑

Fri Aug 14, 2020 2:32 pm

Grumble wrote: ↑

Fri Aug 14, 2020 2:07 pm

discovolante wrote: ↑

Fri Aug 14, 2020 1:50 pm

Going out for meals isn't breaking the rules though?

You’re only meant to go out with your household aren’t you?

https://www.gov.uk/guidance/meeting-peo ... rom-4-july

When eating or drinking out with people you do not live with, you should only meet one other household if you are seated indoors.

If you are eating or drinking outdoors, you can do so with one other household or in a group of up to 6 people from different households. You should take care to limit your interactions with anyone outside the group you visit these places with.

Get away from here with your facts, I want to be cross!

headshot wrote: ↑

Fri Aug 14, 2020 3:48 pm

lpm wrote: ↑

Fri Aug 14, 2020 2:20 pm

Ah, some good news. The government's official UK R number is unchanged at 0.8 to 1.0. Cases are falling at -4% to -1% per day.




IMPORTANT GOVERNMENT NOTICE:
DO NOT LOOK AT ANY CHARTS OF CASES. ANY UPWARDS LINE ON CHARTS IS A STATISTICAL ARTEFACT. CASES ARE FALLING NOT RISING.

Covid-19 app by Kings College and Zoe has had daily cases falling for the last two weeks. c2200 then c1800 and today c1400.

I wonder if mask wearing is starting to have an effect?

Those numbers are clearly bollocks. To be true, Actual Cases would now equal Official Confirmed Through Testing Cases.

You could hypothesise that last month Actual was 3x Official, but is now only 2x, hence Actual stable or falling while Official rises. ONS has the Actual as stable. But in no country in the world can they be 1x. Even 2x seems pretty implausible, given fundamental problem of asymptomatic cases plus the sh.t UK testing.

The Official cases have been 7,000 for the past week, 1,000 per day. Actual cases must be at least 2,000, probably more like 3,000.

headshot wrote: ↑

Fri Aug 14, 2020 3:48 pm

lpm wrote: ↑

Fri Aug 14, 2020 2:20 pm

Ah, some good news. The government's official UK R number is unchanged at 0.8 to 1.0. Cases are falling at -4% to -1% per day.




IMPORTANT GOVERNMENT NOTICE:
DO NOT LOOK AT ANY CHARTS OF CASES. ANY UPWARDS LINE ON CHARTS IS A STATISTICAL ARTEFACT. CASES ARE FALLING NOT RISING.

Covid-19 app by Kings College and Zoe has had daily cases falling for the last two weeks. c2200 then c1800 and today c1400.

I wonder if mask wearing is starting to have an effect?

Today's official number is 1441.

headshot wrote: ↑

Fri Aug 14, 2020 3:48 pm

lpm wrote: ↑

Fri Aug 14, 2020 2:20 pm

Ah, some good news. The government's official UK R number is unchanged at 0.8 to 1.0. Cases are falling at -4% to -1% per day.




IMPORTANT GOVERNMENT NOTICE:
DO NOT LOOK AT ANY CHARTS OF CASES. ANY UPWARDS LINE ON CHARTS IS A STATISTICAL ARTEFACT. CASES ARE FALLING NOT RISING.

Covid-19 app by Kings College and Zoe has had daily cases falling for the last two weeks. c2200 then c1800 and today c1400.

I wonder if mask wearing is starting to have an effect?

The Cambridge/PHE 6 August estimate is 3200 infections per day.

https://www.mrc-bsu.cam.ac.uk/now-casti ... gust-2020/

R is between circa 1 and 0.9.

Case report from Brazil on likely reinfection. Suggests that a mild initial infection may not have built up enough antibodies. https://onlinelibrary.wiley.com/doi/10.1002/jmv.26432

No prizes for guessing which pub chain is failing to collect customers' mobile phone numbers and observe social distancing.

If you want to play along at home you can also try to guess which prominent member of its senior management has weighed into the argument and also his opinion on the relative risk COVID represents.

The results may surprise absolutely nobody

France:

- case numbers going up fast
- but concentrated in the 20-29 age bracket
- and more women than men

The 20-29 age is understandable - young adults eager to resume mating rituals. But the women > men is interesting. Most likely reason is women are more diligent about getting tested when noticing slight symptoms?

Simultaneously there's a very low case rate for the 80-89 and 90+ age brackets. Same in Belgium, Spain and Netherlands. Much better procedures in care homes? Or even a bit of herd immunity among survivors in care homes?

This secondwaving (I think we can now classify Spain and France as secondwaving) has a very different profile to the first. With school reopenings, it's going to look like a young person's epidemic, with correspondingly low hospitalisation and death rates, but maybe some nasty Long Covid implications (triggering diabetes in children being the latest). Eventually, of course, high cases in the young are going to filter through to the rest of the population - unless action is taken to stop young adults spreading it, i.e. closing down pubs again.

Netherlands and Belgium are running a little experiment.

Both started off on the upwards path at the same time, the lines tracking together.

- Netherlands responded with a bit of tutting about obeying the rules, plus getting people in restaurants and bars to leave their phone numbers. No masks, except on public transport

- Belgium brought back restrictions, cut size of household bubble to 5 people per week, mandatory masks (incl outdoor markets) and strict quarantines

Netherlands continues upwards, at risk of a full-on second wave.

Belgium curve turned downwards again, looks like getting back under control.

lpm wrote: ↑

Wed Aug 19, 2020 9:14 am

.... Eventually, of course, high cases in the young are going to filter through to the rest of the population -

Learning the lessons from Florida (or not)

The Florida heat map on that thread is a great visual - on fire down the bottom of the age chart, flames spread upwards...

From one of the BTL comments, the French heat map. The title is in foreign but perhaps someone could translate it. Remarkable how it matches Florida, an intense fire down the bottom, now indications it's starting to spread upwards.

It may be an artefact of the scales, and I'm also certainly overinterpreting from scanty data, but it looks to me as though the France heatmap shows much more age stratification than the Florida one. France is consistent with (a) high-risk social behaviour among 20-30s (b) medium-low-risk behaviour among working-age people (i.e. moderately effective workplace controls) (c) low risk behaviour among retired people (i.e. shielding). Florida suggests high-risk behaviour among all age groups.

If my interpretation is correct, that would provide useful (if not particularly surprising) information on how the virus spreads and what behaviours feed it.

the average age of new covid positives has dropped to 39 in Italy but I haven't been able to find raw data to plot a heat map.

lpm wrote: ↑

Wed Aug 19, 2020 8:59 am

France:

- case numbers going up fast
- but concentrated in the 20-29 age bracket
- and more women than men

The 20-29 age is understandable - young adults eager to resume mating rituals. But the women > men is interesting. Most likely reason is women are more diligent about getting tested when noticing slight symptoms?

Switzerland is men > women. So probably just a random thing?

Little waster wrote: ↑

Mon Aug 17, 2020 10:50 am

No prizes for guessing which pub chain is failing to collect customers' mobile phone numbers and observe social distancing.

If you want to play along at home you can also try to guess which prominent member of its senior management has weighed into the argument and also his opinion on the relative risk COVID represents.

The results may surprise absolutely nobody

Would that be the same pub chain as this one which won't close despite a staff member testing positive?

Kings College Covid-19 app now has 1265 new cases per day. It’s been steadily trending down for a month.

Kings College are a bunch of f.cking idiots.

Actually, I've now read their stuff, and it's you who's getting it badly wrong. They show cases rising and far higher than 1,265 per day - that's an estimate of just symptomatic cases.

Yeah, because the Covid-19 symptom tracker app picks up loads of asymptomatic cases.

They’re tracking over 4m people contributing to the app. Their estimate is that the number of daily new symptomatic cases has reduced week on week since July. My assumption is that the proportion of symptomatic vs asymptomatic cases would remain roughly the same throughout the pandemic, but I suppose that if younger people are getting infected and larger proportion of those people would be less prone to show symptoms.

So, instead of just dismissing this study and calling people f.cking idiots, why don’t you explain why you disagree with their findings?

National COVID-19 figures remain low despite local outbreaks

It's the numbers you are providing that are duds. Where did you get them from? Your numbers cannot possibly right - unless the UK track and trace service miraculously doubled its success rate over the last 3 or 4 weeks.

Per your numbers:

Week to 31 July

Official cases 553 per day
Kings College 2,200 per day
= Official testing service captures 25%

This week

Official cases 1,042 per day
Kings College 1,265 per day
= Official testing service captures 82%

This implied transformation in the UK testing service is clearly bollocks. UK test and trace quantity has gone up a bit while quality has gone down (% results within 24 hours is down). To go from 25% to 82% would be an amazing story with the governments of the world flocking to Dido Harding's door to ask how she did it. Kings College figures of almost halving while confirmed cases almost double should have rang more alarm bells.

Kings College are a bunch of f.cking idiots because they are failing to present their numbers clearly. That link is a terribly written report that throws in unnecessary shadows instead of illumination. There are no straightforward charts or tables of numbers, so it's not surprised your interpretation went awry somewhere. The usual rule applies: if they can't explain something clearly, they probably don't understand it. They are not getting basic presentation right. Where are their weekly case reports saved, where's their chart of the ups and downs, where did you get the 2,200 per day figure from?

Would it help if I provided a link? (Imagine that)

https://covid.joinzoe.com/data

headshot wrote: ↑

Sat Aug 22, 2020 12:51 pm

Would it help if I provided a link? (Imagine that)

https://covid.joinzoe.com/data

I have to say that I have quite a bit of doubt about some aspects of their results

Looking at the map for the NW of England, there are LAs with estimates of >100 cases per 100K directly adjacent to LAs with zero.
Which seems unlikely