I think your example of the 1/100 error rate in a communication channel describes what is known as "positive predictive value" in the medical world, at least as I understand it from watching this video:
I had to work through the examples in the video several times on my own before I could understand it well enough to write a small program to calculate the positive predictive value of a PCR test, given its sensitivity, specificity, and the seroprevalence. When I try to explain this to people without diagrams and math, I use a greatly simplified example: if a test has a 5% false positive rate, but only 5% of people actually have the disease, when you get a positive test, it could just as likely be real as fake.
yeah - there are lots of different ways to approach this problem. I like to think in terms of "inference". What inferences can be drawn from the data? The best book I have on this kind of stuff is MacKay's "Information Theory, Inference and Learning Algorithms" - it's the book that first convinced me I really ought to be more Bayesian haha.
In terms of communication we receive a message - OK - what can we infer about the *sent* message from the *received* message.
In terms of medical testing we receive a positive test result - OK - what can we infer about the state of our health from this result?
PCR testing is aimed at the spike protein, so vaxxed people can show up as "positive" for an unknown number of months post-vax. So there's that. Then there are the <3 months post-recovery cohort - they can show up as positive, too. The statistics using the PCR test detect a lot more people than just people with active disease who are infectious - that's what rapid antigen tests pick up. So with PCR you get a huge haystack of useless results, with very few "needles" - infectious people with active disease - or many "needles" - you don't know, because the PCR tests won't tell you. Pre-vax, in absence of a certain complex of symptoms, the PCR test, tells you that the person may have active disease at a low level, or is recently recovered. Post-vax, PCR testing tells you *nothing*. So at this point, after over a year of vaccination, stats based on PCR testing are worthless - they tell you nothing. So the only kind of testing which counts is rapid antigen testing - and if positive, and after six days for most people without complications, that's going to be a recovered person with natural immunity. The old adage "if you can't bedazzle them with your brains, then baffle them with bullshit" applies here. This last two years has been an absolute barrage of garbage science, and if you think about it, it should be obvious to those pushing the vaccines, and every other scientist, or reasonable person. In other words, the proponents are knowingly and intentionally lying to push highly profitable defective products for which they have no civil liability. The "statistics" are just a smokescreen.
Thanks Nova - it's really annoying me I haven't properly gotten to the bottom of all this yet - I really want to be able to derive some neat (and easily explainable) *limits* on when and how we can get +ve efficacies for subgroups from an overall -ve efficacy. All I have so far is an algebraic gloop.
It struck me when trying to get my head around this "Why are we still explaining ourselves, or even still having this conversation". This is not an insult or meant to take anything away from your efforts, which I, and many others greatly appreciate, though if im honest i prefer your humourous philosophical pieces, but thats just my preference. The point I am trying to make, badly as it happens, is that we shouldn't be in this bloody position. I said "No thanks" to whatever it is in the vial. I said "I'm not going to live my life in fear of a virus, or anything else that could make me poorly for that matter". I'm still the really lovely, pleasant, huggable, caring guy (so my mum tells me) that I was 3 years ago, how can the 2 opinions above make me an enemy of the state. Where is this mass awakening? I don't want to make out that your hard work is futile, but I don't see anything stopping whoever is leading this from getting their way. How many of those who refused boosters, that are supposedly slowing waking up, will join the queue for the booster once their ability to live depends on it. Sorry Rudolf, i'm just having a bad day. I think i'm just analysed out, its been 28 months of looking at charts, trying to learn about virulogy, immunology, epidemiology, psychology, all of which has made me 1000% more intelligent than I was before, but I'm still in the same place, facing whats to come, and it's still coming.
"It’s not quite as holy a cow as before, but it gives to charity and does good deeds on every other day." :)
And it always wears a mask to protect the vulnerable.
I like the sliding-boxes visualisation. When is the interactive tool coming out?
about the same time as I learn to code :-)
I think your example of the 1/100 error rate in a communication channel describes what is known as "positive predictive value" in the medical world, at least as I understand it from watching this video:
https://www.youtube.com/watch?v=NSRK41UbTEU&t=140s
I had to work through the examples in the video several times on my own before I could understand it well enough to write a small program to calculate the positive predictive value of a PCR test, given its sensitivity, specificity, and the seroprevalence. When I try to explain this to people without diagrams and math, I use a greatly simplified example: if a test has a 5% false positive rate, but only 5% of people actually have the disease, when you get a positive test, it could just as likely be real as fake.
yeah - there are lots of different ways to approach this problem. I like to think in terms of "inference". What inferences can be drawn from the data? The best book I have on this kind of stuff is MacKay's "Information Theory, Inference and Learning Algorithms" - it's the book that first convinced me I really ought to be more Bayesian haha.
In terms of communication we receive a message - OK - what can we infer about the *sent* message from the *received* message.
In terms of medical testing we receive a positive test result - OK - what can we infer about the state of our health from this result?
PCR testing is aimed at the spike protein, so vaxxed people can show up as "positive" for an unknown number of months post-vax. So there's that. Then there are the <3 months post-recovery cohort - they can show up as positive, too. The statistics using the PCR test detect a lot more people than just people with active disease who are infectious - that's what rapid antigen tests pick up. So with PCR you get a huge haystack of useless results, with very few "needles" - infectious people with active disease - or many "needles" - you don't know, because the PCR tests won't tell you. Pre-vax, in absence of a certain complex of symptoms, the PCR test, tells you that the person may have active disease at a low level, or is recently recovered. Post-vax, PCR testing tells you *nothing*. So at this point, after over a year of vaccination, stats based on PCR testing are worthless - they tell you nothing. So the only kind of testing which counts is rapid antigen testing - and if positive, and after six days for most people without complications, that's going to be a recovered person with natural immunity. The old adage "if you can't bedazzle them with your brains, then baffle them with bullshit" applies here. This last two years has been an absolute barrage of garbage science, and if you think about it, it should be obvious to those pushing the vaccines, and every other scientist, or reasonable person. In other words, the proponents are knowingly and intentionally lying to push highly profitable defective products for which they have no civil liability. The "statistics" are just a smokescreen.
Very creative way of looking at the stats. (Creative in a good way) 👍🏽
Thanks Nova - it's really annoying me I haven't properly gotten to the bottom of all this yet - I really want to be able to derive some neat (and easily explainable) *limits* on when and how we can get +ve efficacies for subgroups from an overall -ve efficacy. All I have so far is an algebraic gloop.
It must be hard with this covid data that can't be trusted.
It struck me when trying to get my head around this "Why are we still explaining ourselves, or even still having this conversation". This is not an insult or meant to take anything away from your efforts, which I, and many others greatly appreciate, though if im honest i prefer your humourous philosophical pieces, but thats just my preference. The point I am trying to make, badly as it happens, is that we shouldn't be in this bloody position. I said "No thanks" to whatever it is in the vial. I said "I'm not going to live my life in fear of a virus, or anything else that could make me poorly for that matter". I'm still the really lovely, pleasant, huggable, caring guy (so my mum tells me) that I was 3 years ago, how can the 2 opinions above make me an enemy of the state. Where is this mass awakening? I don't want to make out that your hard work is futile, but I don't see anything stopping whoever is leading this from getting their way. How many of those who refused boosters, that are supposedly slowing waking up, will join the queue for the booster once their ability to live depends on it. Sorry Rudolf, i'm just having a bad day. I think i'm just analysed out, its been 28 months of looking at charts, trying to learn about virulogy, immunology, epidemiology, psychology, all of which has made me 1000% more intelligent than I was before, but I'm still in the same place, facing whats to come, and it's still coming.