The Virus anybody else done the maths

[Hobbit]

You know his view on this one is completely crazy.

No they're not. I, as others did, read it differently from you. I, genuinely, can't make the connection you did without going 2+2=347.

Here's an off the cuff thought; if you're the only person that read it as a Tommy Robinson slant you're either very insightful or way off the mark. Surely you question your own metric if you see that you're the only one seeing it that way? There's some very intelligent people on this forum... how many people do you see leaping up and calling him a Tommy Robinson supporter?

 

[Kellfire]

No they're not. I, as others did, read it differently from you. I, genuinely, can't make the connection you did without going 2+2=347.

Here's an off the cuff thought; if you're the only person that read it as a Tommy Robinson slant you're either very insightful or way off the mark. Surely you question your own metric if you see that you're the only one seeing it that way? There's some very intelligent people on this forum... how many people do you see leaping up and calling him a Tommy Robinson supporter?

I read it as Boomer saying we should all have guns. Nothing to do with Tommy Robinson.

 

[SwingsitlikeHogan]

Yet apparently the statistics projected by Imperial College London have been revised down from 500,000 deaths to 5,700. That strikes me as something that was never really going to be going to be as severe we feared, because with all due respect the the Government & the NHS I don't see their actions cutting down the estimates that much. Interestingly one of the comments on the original FT article pointed out that ICL's modelling had previously got a couple of the other infectious outbreaks estimates wildly wrong.

I don't know any more than anyone else, and this is not intended as criticism of the action being taken by the Government, merely as a discussion point.

It's not that difficult to understand. With little or no data, and no spread prevention measures in place, then all you have to start with on which to base your predictions is the basic dynamics of what you are trying to model - and you reflect that by there being very high level of uncertainty associated with what you are trying to model. So at first it is very possible that the basic dynamics of pandemic spread might well give an estimated 500,000 deaths. Now that figure might well have been the 3sigma (99.7% confidence interval) - and indeed the most likely estimate at the outset might have been a lot less as would have been the best possible -3sigma let's say.

As measurements of ACTUAL spread and growth come in, the uncertainty around some of the things you are modelling that drive the number of deaths reduces...and you estimates can be given with greater confidence. Then there are the spread preventative measures. As these are applied one-by-one the impact on spread and growth is assessed through measurement and data. Some will have more impact than others and social separation if we apply it properly can have a very major impact - at the moment we just don;t know as it is too soon for the impact to manifest itself.

So for instance if we perfectly applied absolute separation then it is clear that that would stop spread of the virus dead in it's tracks. Now that isn't going to happen, but you can model it and come out with a very low number of deaths if that is what you want to do.

It is therefore very understandable from a statistical modelling perspective to understand what is happening in respect of the estimates of numbers of deaths.

 

[Ye Olde Boomer]

And that is definitely the most stupid thing ever said on this forum. It also makes me think you must be a troll!

Sorry, Kellfire.
I underestimated your hypersensitivity and thought it was just a little joke.

 

[drdel]

It's not that difficult to understand. With little or no data, and no spread prevention measures in place, then all you have to start with on which to base your predictions is the basic dynamics of what you are trying to model - and you reflect that by there being very high level of uncertainty associated with what you are trying to model. So at first it is very possible that the basic dynamics of pandemic spread might well give an estimated 500,000 deaths. Now that figure might well have been the 3sigma (99.7% confidence interval) - and indeed the most likely estimate at the outset might have been a lot less as would have been the best possible -3sigma let's say.

As measurements of ACTUAL spread and growth come in, the uncertainty around some of the things you are modelling that drive the number of deaths reduces...and you estimates can be given with greater confidence. Then there are the spread preventative measures. As these are applied one-by-one the impact on spread and growth is assessed through measurement and data. Some will have more impact than others and social separation if we apply it properly can have a very major impact - at the moment we just don;t know as it is too soon for the impact to manifest itself.

So for instance if we perfectly applied absolute separation then it is clear that that would stop spread of the virus dead in it's tracks. Now that isn't going to happen, but you can model it and come out with a very low number of deaths if that is what you want to do.

It is therefore very understandable from a statistical modelling perspective to understand what is happening in respect of the estimates of numbers of deaths.

I think you'll find it is far from setting simplistic 3sigma confidence boundaries. Using models that sought to explain the SARS outbreak based on uniform distributions from which you could use 3sigma boundaries were of little use. There are many refereed papers that review the behaviour of interactions in populations, controlled or uncontrolled and/or constrained by interventions when using various sorts of pdf's and the specific suitability of geometric, Poisson and binomial distributions. Have squint at the Bulletin of Mathematical Biology. I think you'll find that statistical modelling has given way to discrete event simulation and other process based evaluations.

 

[IainP]

Just an observation:
On the "by population size" stats

France, Netherlands, Belgium & Switzerland are featuring higher. Not sure what that says other than as we know Europe is a hot spot.

I expect the USA will be climbing this week also.

 

[Hobbit]

Just an observation:
On the "by population size" stats

France, Netherlands, Belgium & Switzerland are featuring higher. Not sure what that says other than as we know Europe is a hot spot.

I expect the USA will be climbing this week also.

The UK has a natural barrier, the Channel.

 

[Stuart_C]

The UK has a natural barrier, the Channel.

Thankfully.

 

[williamalex1]

Thankfully.

Kay sera sera, C'est la vie.

 

[IainP]

A alternative view on the maths from the Peter Hitchens column;

I shall come to these, to underline the fact that it is not I, alone, who have these doubts. I do not claim to be an expert. But I refer to those who definitely are experts, who doubt the wisdom of what we are doing.
It is sad that far too little of this is being reported as prominently as it should be by our supposedly diverse and free media, especially the BBC, which has largely closed its mind and its airwaves to dissent. It is quite funny that a statue of George Orwell stands by the entrance to the BBC, bearing the inscription: ‘If liberty means anything at all it means the right to tell people what they do not want to hear.’
Obviously, they should take it down, as nobody inside the building appears to believe that.
Crucially, those who began by claiming that we faced half a million deaths from the coronavirus in this country have now greatly lowered their estimate. Professor Neil Ferguson was one of those largely responsible for the original panic. He or others from Imperial college have twice revised his terrifying prophecy, first to fewer than 20,000 and then on Friday to 5,700.
...

Re the bit in bold, apparently it assumed this: "His paper predicted that at its highest, Britain would have 260 deaths a day. "

Seems the total figure has already been revised upwards by the same group

 

[SwingsitlikeHogan]

I think you'll find it is far from setting simplistic 3sigma confidence boundaries. Using models that sought to explain the SARS outbreak based on uniform distributions from which you could use 3sigma boundaries were of little use. There are many refereed papers that review the behaviour of interactions in populations, controlled or uncontrolled and/or constrained by interventions when using various sorts of pdf's and the specific suitability of geometric, Poisson and binomial distributions. Have squint at the Bulletin of Mathematical Biology. I think you'll find that statistical modelling has given way to discrete event simulation and other process based evaluations.

I was only talking very simplistically around why it is quite possible to get very significant changes in estimates from the initial estimates

 

[IainP]

I think that Mr boomer may be having a gentle poke at American values.
His grasp of irony makes me think that his most troll-y activity is that he's not actually American at all...

And maybe that poke was well made...
https://www.bbc.co.uk/news/world-us-canada-52108162

 

[Crazyface]

The UK has a natural barrier, the Channel.

Like that's working for us. Immigrants are still floating over and still we take them in. Utterly bonkers.

 

[IainP]

As a few have commented, the counting variations are complex. Seems France has done a "catch up"
1355 new deaths in France, including 884 fatalities in nursing homes that occurred over a period of several weeks and that were announced only today. NOTE: If and when the French government determines the correct distribution of these additional deaths over time, we will adjust historical data accordingly

 

[IainP]

Know we're not meant to look at days in isolation, however noting the UK reported more today than Spain. Better news for Spain if now falling, sobering for UK as Spain was elevated for 10 days or so.

France posting a large number again, so far without the explanation like yesterday.

Sorting by deaths per population the ten most impacted are all European countries (including the UK)

 

[IainP]

^ edit to above, looks like Spain's figures for the 3rd were edited upwards by 300 or so.
Also the US (with the time difference I suppose) increased greatly.
Sad times.

 

[harpo_72]

Still people are not heeding the request and huge social gatherings occurring.
This means we won’t see an improvement at all as they will infect each .. having isolated properly from day 1 this going to be a long time!! Every time there is a big gathering they are adding at least a 3 week period of required isolation based on the infection and infectious period.

 

[User62651]

One for the analysts.
Looking at IainPs recent posts the stats show UK now has around 10,000 deaths of a worldwide total of 100,000 reported so roughly 10%.
Our UK population however is close to 1% of the world population I think.
So is that 10% to be expected given UK was next affected along with France after China and then Italy/Spain so it'll drop right back as other continents catch up or are our UK numbers going to stay proportionately very high? If so why?
Just read the awful numbers today and seems awfully high here before we've reportedly even reached a peak.
Sad times indeed.

 

[Jacko_G]

A couple of bawbags on the golf course in my village today when the wife and I went for a walk, brazen as foxtrot as well with bags.

 

[harpo_72]

One for the analysts.
Looking at IainPs recent posts the stats show UK now has around 10,000 deaths of a worldwide total of 100,000 reported so roughly 10%.
Our UK population however is close to 1% of the world population I think.
So is that 10% to be expected given UK was next affected along with France after China and then Italy/Spain so it'll drop right back as other continents catch up or are our UK numbers going to stay proportionately very high? If so why?
Just read the awful numbers today and seems awfully high here before we've reportedly even reached a peak.
Sad times indeed.

Think you have to also consider cities there size and how the accommodation is .. I would be interested in Tokyo as they are totally rammed, Monaco as well is rammed ... They were saying that in New York it was the poor areas that were hit, but they to are close together.
Population density is key surely?