The suggestion of vx harms has massive implications, so I really appreciate your thoroughness when investigating the data.
As soon as covid and non-covid deaths are separated, we have some kind of a value judgement going on.
In past years, covid was not present, but there have always been respiratory pathogens. It would seem that covid displaced these over the last couple of years.
If we are calculating non covid Xs deaths, then would we be better using non flu deaths as the reference? Or more reasonably, use all cause deaths (now) and against your calculated reference?
But why use xs deaths at all?
As I understand it, the stats (both ARIMA and xcorrelation) use the variation in the independent and dependent variables. So if there is an effect of doses at a population level, wouldn't we expect to this in all cause deaths?
Why thank you kind sir! I'll squeeze a bit more out of xs but will soon be dropping this as a bad idea. In the same vein I'll drop the somewhat dubious categorisation of COVID and, instead, use multivariate techniques with ARIMA to account for various factors within all cause death. Yes indeed, there are massive implications and the best I can do is to be totally honest as I crank the handle.
Hi I love all this statistical sleuthing. My son, a soon to be PhD, who uses modelling for virus analysis is a total sceptic. I just throw in some of his raw comments. Maybe they might flick the odd light switch and thence might be of some use.
'Do you think the number of excess deaths might be look weird the winter after thousand of old people died a week the previous winter?
Do you think the ongoing open label trials for the vaccines would pick up its killing people after 5 months?
We doesn't he look at high vaccine coverage areas like Scotland, they should have bright red excess deaths after 5 months... /s'
He's right about excess deaths being twisted out of shape and he needs to read the next few articles. As for open label trials all depends on how they're conducted and analysed. The fact that damning Pfizer documents that were supposed to be sealed for 75 years have been released by judicial ruling should speak volumes. High vaccine coverage areas like Israel and Australia do, in fact, show bright red excess - clearly he's not up to speed with everything he should be.
Though we don't have to look at exotic places to realise something is out of kilter. Excess death in both England and Scotland is indeed a bit excessive, especially Scotland, as suggested...
On a personal note he might like to explain the four unexpected deaths among my friends and family after dosing in the last 12 months, with a fifth suffering kidney damage from micro clots. He might also like to speak to the pathologists and GPs I've spoken to. At some point he's going to learn the hard way like many people.
That being said we don't have to resort to personal anecdote (I am in receipt of many from both the public and healthcare professionals) and we don't have to resort to modelling of excess deaths - we can simply take a look at what is happening. In this regard I am fortunate enough to be sitting on medical records of 58k deceased persons and 736k admissions to the emergency departments of one of the largest UK NHS Trusts. Analysis of this confidential dataset has led to 20 restricted circulation reports, little of which has been released to the public. Owing to legalities I have to be careful what I discuss but I can tell you now that vaccine harm is not an illusion.
Since March 2020 I've had dozens and dozens of 'so and so says' conversations like this one, where the sceptic remains sceptical no matter what evidence is mustered and what line of reasoning taken. This is as it should be for I'm not here to convince anybody of anything, I'm here to use my experience as a statistician and apply it to a particularly tricky problem: I'd like to see what the data says rather than various shades of expert and sceptic.
I’m very sympathetic to their being a vax/death relationship, but I’m skeptical of the 23 week relationship here. A real effect should be almost as strong at 22 and 24 weeks as 23 and yet the sign actually flips if it’s off by even a week. That’s very hard to explain as anything other than a statistical artifact.
The problem is the excess death curve is almost flat except for a couple areas of rapid change. This means that any function which has an area of rapid change and is otherwise much flatter will correlate very strongly when the areas of rapid change are aligned.
My thoughts also. I was going to run models with lags ranging from 18 to 28 weeks to show just how unstable things were around that 23 weeks but I found a quicker way around the issue - and caught the culprit! This tray bake will be coming out of the oven at 7am tomorrow.
Great stuff!
The suggestion of vx harms has massive implications, so I really appreciate your thoroughness when investigating the data.
As soon as covid and non-covid deaths are separated, we have some kind of a value judgement going on.
In past years, covid was not present, but there have always been respiratory pathogens. It would seem that covid displaced these over the last couple of years.
If we are calculating non covid Xs deaths, then would we be better using non flu deaths as the reference? Or more reasonably, use all cause deaths (now) and against your calculated reference?
But why use xs deaths at all?
As I understand it, the stats (both ARIMA and xcorrelation) use the variation in the independent and dependent variables. So if there is an effect of doses at a population level, wouldn't we expect to this in all cause deaths?
Why thank you kind sir! I'll squeeze a bit more out of xs but will soon be dropping this as a bad idea. In the same vein I'll drop the somewhat dubious categorisation of COVID and, instead, use multivariate techniques with ARIMA to account for various factors within all cause death. Yes indeed, there are massive implications and the best I can do is to be totally honest as I crank the handle.
Hi I love all this statistical sleuthing. My son, a soon to be PhD, who uses modelling for virus analysis is a total sceptic. I just throw in some of his raw comments. Maybe they might flick the odd light switch and thence might be of some use.
'Do you think the number of excess deaths might be look weird the winter after thousand of old people died a week the previous winter?
Do you think the ongoing open label trials for the vaccines would pick up its killing people after 5 months?
We doesn't he look at high vaccine coverage areas like Scotland, they should have bright red excess deaths after 5 months... /s'
Cheers Glenn
He's right about excess deaths being twisted out of shape and he needs to read the next few articles. As for open label trials all depends on how they're conducted and analysed. The fact that damning Pfizer documents that were supposed to be sealed for 75 years have been released by judicial ruling should speak volumes. High vaccine coverage areas like Israel and Australia do, in fact, show bright red excess - clearly he's not up to speed with everything he should be.
Though we don't have to look at exotic places to realise something is out of kilter. Excess death in both England and Scotland is indeed a bit excessive, especially Scotland, as suggested...
https://twitter.com/ons/status/1544237459105583104
https://www.thetimes.co.uk/article/excess-death-tally-is-worst-in-scotland-2txpwxmb7
On a personal note he might like to explain the four unexpected deaths among my friends and family after dosing in the last 12 months, with a fifth suffering kidney damage from micro clots. He might also like to speak to the pathologists and GPs I've spoken to. At some point he's going to learn the hard way like many people.
That being said we don't have to resort to personal anecdote (I am in receipt of many from both the public and healthcare professionals) and we don't have to resort to modelling of excess deaths - we can simply take a look at what is happening. In this regard I am fortunate enough to be sitting on medical records of 58k deceased persons and 736k admissions to the emergency departments of one of the largest UK NHS Trusts. Analysis of this confidential dataset has led to 20 restricted circulation reports, little of which has been released to the public. Owing to legalities I have to be careful what I discuss but I can tell you now that vaccine harm is not an illusion.
Since March 2020 I've had dozens and dozens of 'so and so says' conversations like this one, where the sceptic remains sceptical no matter what evidence is mustered and what line of reasoning taken. This is as it should be for I'm not here to convince anybody of anything, I'm here to use my experience as a statistician and apply it to a particularly tricky problem: I'd like to see what the data says rather than various shades of expert and sceptic.
I’m very sympathetic to their being a vax/death relationship, but I’m skeptical of the 23 week relationship here. A real effect should be almost as strong at 22 and 24 weeks as 23 and yet the sign actually flips if it’s off by even a week. That’s very hard to explain as anything other than a statistical artifact.
The problem is the excess death curve is almost flat except for a couple areas of rapid change. This means that any function which has an area of rapid change and is otherwise much flatter will correlate very strongly when the areas of rapid change are aligned.
My thoughts also. I was going to run models with lags ranging from 18 to 28 weeks to show just how unstable things were around that 23 weeks but I found a quicker way around the issue - and caught the culprit! This tray bake will be coming out of the oven at 7am tomorrow.