Thank you for your candour - this is really useful demonstration that statistically significant results can still occur as from artefacts in the data sets.
Given the vagaries of the data recordings, I wonder if it is really possible to use the high frequency parts of the signal at all?
If not, then aren't we back to comparing rates over discrete time periods between the vaxed and unvaxed. Then we have the problem of what denominators to use in each group, let alone trying to match the two groups to avoid potential biases.
Data recording vagaries (noise and time shift) should be roughly the same for both vaxed and unvaxed. So could we either use the ratio of the two (thus cancelling the noise) or perhaps look at CCF of deaths and jabs of the vaxed and unvaxed groups separately? If there is a vax effect, then the unvaxed results should serve as a reference.
Why thank you kind sir! This was a useful experience for me and a salutary reminder of the precarious position of the statistician. We are thinking along the same lines in that I'm now favouring discrete period analysis instead of 'curve matching'. I've also been exploring temporal causal modelling (Granger causality) to see if that offers anything reliable. I haven't got data for vaccination status as yet but I'll probably take a look at trends for age bands over time first. What got me tugging my beard most was having the 2015-2019 baseline series correlate with dosing - that really took the biscuit!
Something new every day - Granger causality test - thanks!
I was really pleased you looked at the jabs to deaths CCF in the period prior to vax - that was to be one of my 'sanity check' suggestions. I recall my signal processing lecturer saying there are very few truly random signals!
Is it possible that noise in the form of time shifting was responsible for the really strange results you got from the EPR record analyses that you did where you randomised the patients and still got a +ve vax effect?
Yes, that had to be done to reassure - and it didn't! It's fascinating how things changing over time can play tricks and, yes, I'm guessing the EPR simulation was proof of this. It was only when records were fully randomised did the effect disappear - to see an effect at 60 - 70% randomisation was pretty wild.
I really appreciate that you put the English translation in every so often so I can kind of track what you are talking about.
I find the excess death data extremely curious and don't know what to make of it. I was hoping that this 9 part thread would be able to shed light on what might be causing the excess deaths. But alas, it has not. (through no fault of yours)
Your attempts to tackle the problem went down seemingly promising paths each of which then didn't work out because the data is "not clean" or "has hidden features that are just artifacts of how society functions". This is my summary of your nine super interesting posts. :-)
I have a question, which I think I know the answer to, but maybe not. You had a hypothesis you were testing: were there excess deaths five months after the jabs? Your conclusion seems to be that you currently have no data to support the hypothesis. Is the inverse true as well, that there is no data to suggest that the vaccines are mitigating the excess deaths? I am not talking about specific results that you kept on seeming to find showing the vaccines seemed to be working, but then the models blew up. Rather I am talking about the messiness of the data not allowing one to make either assertion: the vaccines are causing excess deaths nor the vaccines are mitigating excess deaths.
Yes, this makes sense. There are no fast and easy answers owing to muddled data, changing definitions and a raft of confounding factors longer than your arm! It takes a great deal of time, thinking and effort to tease out a robust answer in retrospective studies of the real world. Whilst the vaccine series stopped at part 9 the work hasn't stopped - I am now considering other factors that may contribute to excess death, closure of hospital beds/shortage of staff being but one more factor to dial in. Then we have to tackle the confounding factor of long-COVID and the long term impact of lockdown (other than bed closure). Until you nail all the key factors you can't say anything sensible about vaccine harm/benefit. At some point I shall return to the vaccine study and kick-off with part 10. At this stage I have no idea how much more work needs to be put in but tricky topics like this usually take 6 - 8 months of solid work to unravel.
I highly recommend you watch the latest Dr Drew Youtube where he interviews Ed Dowd. He is an analyst from Wall Street who has been convinced about the vaccine effects for some time and has a book 'Cause unknown". He has set up a website called the humanity project where he and other mathematicians and analysts show the catastrophic effects on mortality and disability. In the interview he invites other scientists to contribute to this website. I think it may be worth you considering getting in touch about your work. For example the data he shows for Denmark reveals increase in XS mortality in each age group which follows the cumulative dose of vaccine which results in the Danish authorities now stopping vaccines for under 50yr olds because the risk of jab worse than the disease. Also in the video he mentions the extraordinary stats on sudden death in athletes. The Olympic Committee did survey pre covid which showed World Wide about 29 deaths per year (mostly cardiac) but post vaccine it is now 50 per month. All of this is ignored by the regulatory authorities who should be taking action
The website recently created mentioned in the interview is phinancetecnologies.com. The reason he says for using phi rather than F was because of the mathematicians in his team. I think personally its a mistake as difficult to remember! You might be interested in the excess death methodology they use for UK data which is under heading of 'Resources' on the website menu. They use a different method for calculating baseline death ie rates rather than numbers. But they show lots of graphs which I have not really looked at yet. Best wishes
That is quite fabulous and is giving me ideas! Grateful thanks.
LOL - now that takes me back to my youth and sitting around the radio with all the other undergrads instead of ploughing through essays!
Thank you for your candour - this is really useful demonstration that statistically significant results can still occur as from artefacts in the data sets.
Given the vagaries of the data recordings, I wonder if it is really possible to use the high frequency parts of the signal at all?
If not, then aren't we back to comparing rates over discrete time periods between the vaxed and unvaxed. Then we have the problem of what denominators to use in each group, let alone trying to match the two groups to avoid potential biases.
Data recording vagaries (noise and time shift) should be roughly the same for both vaxed and unvaxed. So could we either use the ratio of the two (thus cancelling the noise) or perhaps look at CCF of deaths and jabs of the vaxed and unvaxed groups separately? If there is a vax effect, then the unvaxed results should serve as a reference.
Why thank you kind sir! This was a useful experience for me and a salutary reminder of the precarious position of the statistician. We are thinking along the same lines in that I'm now favouring discrete period analysis instead of 'curve matching'. I've also been exploring temporal causal modelling (Granger causality) to see if that offers anything reliable. I haven't got data for vaccination status as yet but I'll probably take a look at trends for age bands over time first. What got me tugging my beard most was having the 2015-2019 baseline series correlate with dosing - that really took the biscuit!
Something new every day - Granger causality test - thanks!
I was really pleased you looked at the jabs to deaths CCF in the period prior to vax - that was to be one of my 'sanity check' suggestions. I recall my signal processing lecturer saying there are very few truly random signals!
Is it possible that noise in the form of time shifting was responsible for the really strange results you got from the EPR record analyses that you did where you randomised the patients and still got a +ve vax effect?
Yes, that had to be done to reassure - and it didn't! It's fascinating how things changing over time can play tricks and, yes, I'm guessing the EPR simulation was proof of this. It was only when records were fully randomised did the effect disappear - to see an effect at 60 - 70% randomisation was pretty wild.
I really appreciate that you put the English translation in every so often so I can kind of track what you are talking about.
I find the excess death data extremely curious and don't know what to make of it. I was hoping that this 9 part thread would be able to shed light on what might be causing the excess deaths. But alas, it has not. (through no fault of yours)
Your attempts to tackle the problem went down seemingly promising paths each of which then didn't work out because the data is "not clean" or "has hidden features that are just artifacts of how society functions". This is my summary of your nine super interesting posts. :-)
I have a question, which I think I know the answer to, but maybe not. You had a hypothesis you were testing: were there excess deaths five months after the jabs? Your conclusion seems to be that you currently have no data to support the hypothesis. Is the inverse true as well, that there is no data to suggest that the vaccines are mitigating the excess deaths? I am not talking about specific results that you kept on seeming to find showing the vaccines seemed to be working, but then the models blew up. Rather I am talking about the messiness of the data not allowing one to make either assertion: the vaccines are causing excess deaths nor the vaccines are mitigating excess deaths.
Does this make sense?
Yes, this makes sense. There are no fast and easy answers owing to muddled data, changing definitions and a raft of confounding factors longer than your arm! It takes a great deal of time, thinking and effort to tease out a robust answer in retrospective studies of the real world. Whilst the vaccine series stopped at part 9 the work hasn't stopped - I am now considering other factors that may contribute to excess death, closure of hospital beds/shortage of staff being but one more factor to dial in. Then we have to tackle the confounding factor of long-COVID and the long term impact of lockdown (other than bed closure). Until you nail all the key factors you can't say anything sensible about vaccine harm/benefit. At some point I shall return to the vaccine study and kick-off with part 10. At this stage I have no idea how much more work needs to be put in but tricky topics like this usually take 6 - 8 months of solid work to unravel.
I highly recommend you watch the latest Dr Drew Youtube where he interviews Ed Dowd. He is an analyst from Wall Street who has been convinced about the vaccine effects for some time and has a book 'Cause unknown". He has set up a website called the humanity project where he and other mathematicians and analysts show the catastrophic effects on mortality and disability. In the interview he invites other scientists to contribute to this website. I think it may be worth you considering getting in touch about your work. For example the data he shows for Denmark reveals increase in XS mortality in each age group which follows the cumulative dose of vaccine which results in the Danish authorities now stopping vaccines for under 50yr olds because the risk of jab worse than the disease. Also in the video he mentions the extraordinary stats on sudden death in athletes. The Olympic Committee did survey pre covid which showed World Wide about 29 deaths per year (mostly cardiac) but post vaccine it is now 50 per month. All of this is ignored by the regulatory authorities who should be taking action
The link is https://www.youtube.com/watch?v=ZAoIYJKMEgc&t=3300s and the title is 'Ed Dowd exposes sudden Adult Deaths. Hope you find it interesting
Fabulous - thank you very much!
The website recently created mentioned in the interview is phinancetecnologies.com. The reason he says for using phi rather than F was because of the mathematicians in his team. I think personally its a mistake as difficult to remember! You might be interested in the excess death methodology they use for UK data which is under heading of 'Resources' on the website menu. They use a different method for calculating baseline death ie rates rather than numbers. But they show lots of graphs which I have not really looked at yet. Best wishes
Smashing stuff - will take a look.