Immortality & Bayesian Statistics

[Jabba]

- I still think that my priors are generous in favor of the H side. The following are a couple more sites that I think support my suspicion.
- http://www.pewforum.org/2009/11/05/scientists-and-belief/
- https://phys.org/news/2015-12-worldwide-survey-religion-science-scientists.html
- I just figure that we should be open-minded enough to accept that atheists just might be missing something -- especially, when a lot of scientists are not atheists and do think that the atheists are missing something.

- So far, I can't figure out why I'm wrong about immortality, but have to suspect that I am, since it would seem that no "official" Statistician has ever proposed what I am.
- I'm hoping that the explanation is that official Statisticians tend to be too analytic to believe in transcendence and to thereby think to apply Bayes to it. And, whereas Many "Scientists" seem to accept transcendence, maybe the same is not true re Statisticians -- and then, most scientists aren't familiar enough with Bayesian Statistics in order to notice (what I think are) the implications ...

 

[Dason]

It's probably because statisticians like data. And all data points to being able to feel pain and experience illness (so even if I am "immortal" how does that change anything) and we've seen people we know experience death and have no reason to believe that we are different. So yeah... I stopped actually reading your ramblings a while ago but if you don't address those issues I doubt you'll convince anyone.

 

[Jabba]

- Hopefully, someone is still tuning in. If so, please let me know where you disagree.
- Now, I'll re-read the whole thread and try to answer any unanswered questions or objections.

 

[Jabba]

It's probably because statisticians like data. And all data points to being able to feel pain and experience illness (so even if I am "immortal" how does that change anything) and we've seen people we know experience death and have no reason to believe that we are different. So yeah... I stopp actually edreading your ramblings a while ago but if you don't address those issues I doubt you'll convince anyone.

Dason,
- The "data" I'm using for P(E|H) is the evidence determining what is needed to result in my current existence.
- The data I'm using for P(E|~H) is the evidence supporting the different possible specific hypotheses included in ~H.
- The data I'm using for the prior probabilities is the evidence supporting them.

 

[Jabba]

Assume that X is a random variable with a standard normal distribution. Lets say you observe a draw from this variable (call the observed value x). Do you know what P(X = x) is? Regardless of the observed value the probability is 0. So just observing a small probability doesn't mean that something is impossible.

Dason,
- I'm not sure that I understand your question, but my best guess is that you're asking for the likelihood of drawing x -- given the hypothesis (population) of X. If so, we don't know anything about X or x and don't have a clue as to the likelihood of drawing x from X.

 

[Jabba]

I like a good philosophical debate, but I think this was easily squashed by Dason on the exact same thing I was thinking: P(X=x) = 0. This argument is also not that different to me than the anthropic principle in that it tries to use limited information (observations) to infer global principles about things to which the observation provides no basis for. In this case, P(X=x) = 0, therefore the entire domain (support) must have 0 probability! That wouldn't make sense ...

Bryan,
- It is the same logic as used in the anthropic principle, but how do you support "that it tries to use limited information (observations) to infer global principles about things to which the observation provides no basis for"? The numbers used to determine the likelihood of life in a universe have been scientifically determined.

 

[Dason]

I'm done.

 

[GretaGarbo]

Not that I have any answers to Jabbas statements. But for curiosity:

What is the "hypothesis" H?
What is the data?
What is the distribution for the prior?
and what is P(H)?
and where have you got that probability from?
What is the distribution used for the likelihood?

 

[Jabba]

Greta,
- H is that I have only one finite life (at most). Abbreviated as OOFLam.
- P(H) = .99
- That prior is an estimate based upon scientific conclusions. I suspect that I'm being generous.
- As to data and distributions, I don't really know the right answers... Maybe, these concepts are more relevant to frequentist inference than Bayesian inference? Maybe, "degrees of belief" replaces these concepts in Bayesian inference?
- Re P(E|H), the estimate is based upon the prior probability of each of the different events required to result in my current existence.
- Re P(E|~H), I'll have to get back to you on that...

 

[Abaddon]

Greta,
- H is that I have only one finite life (at most). Abbreviated as OOFLam.

Explain how exactly it is possible to have 0.5 of a life? or 0.63333 of a life? It is a binary condition. You have a life or you do not. Your OOFLam is blatantly a false premise.

- P(H) = .99

And how exactly did you derive that number?

- That prior is an estimate based upon scientific conclusions. I suspect that I'm being generous.

Provide that data to support that random number.
.

- As to data and distributions, I don't really know the right answers...

You don't? Why are you therefore insisting that you do? Have you a basis for those numbers?

Maybe, these concepts are more relevant to frequentist inference than Bayesian inference? Maybe, "degrees of belief" replaces these concepts in Bayesian inference?

No. Just no.

- Re P(E|H), the estimate is based upon the prior probability of each of the different events required to result in my current existence.

And what exactly is that prior probability? From whence did you glean those numbers? Surely you cannot be simply making them up out of whole cloth? Surely you would not descend into such randomised numbers? Surely you must have some basis for those numbers? Well, here you are on a stats forum. Feel free to wheel out those derivations.

- Re P(E|~H), I'll have to get back to you on that...

No. You will not.

 

[Abaddon]

Not that I have any answers to Jabbas statements. But for curiosity:

What is the "hypothesis" H?

Apparently, whatever Jabba says at a given moment.

What is the data?

None.

What is the distribution for the prior?
and what is P(H)?

Moveable feast.

and where have you got that probability from?

Jabba wont tell.

What is the distribution used for the likelihood?

Not that either.

 

[Abaddon]

Let us be clear here, In Jabba wourls, the odds of us all being a simply a product of random evolution is 1 in 1 -100.

Au contraire. in the Jabbaverse, the likelyhood of random evolution plus "god" is far more likely than simple evolution.

I challenge anyone to work that out.

 

[Jabba]

Explain how exactly it is possible to have 0.5 of a life? or 0.63333 of a life? It is a binary condition. You have a life or you do not. Your OOFLam is blatantly a false premise...

Abbadon,
- When flipping a fair coin, what's the probability that it will land heads?

 

[Jabba]

Greta,
- H is that I have only one finite life (at most). Abbreviated as OOFLam.
- P(H) = .99
- That prior is an estimate based upon scientific conclusions. I suspect that I'm being generous.
- As to data and distributions, I don't really know the right answers... Maybe, these concepts are more relevant to frequentist inference than Bayesian inference? Maybe, "degrees of belief" replaces these concepts in Bayesian inference?
- Re P(E|H), the estimate is based upon the prior probability of each of the different events required to result in my current existence.
- Re P(E|~H), I'll have to get back to you on that...

Greta,
- I don't know, or remember, what's gone before that you should be aware of, but the specific "existence" that I'm referring to is my "self," and not my body.
- Re P(E|~H):
- I figure that ~H is made up of multiple specific options:
#1. My self exists an infinity of times.
#2. My self exists multiple times, but not an infinity of times.
#3. MY self exists infinitely, but on different "planes."
- If those are all the different specific possibilities, I need to estimate the prior probability of each (given that ~H is true), multiply those numbers by the likelihood of my current existence given that specific possibility, then add up the separate products. Hopefully, I got that right...
- Whatever, it ought to get us started.

 

[Jabba]

CowboyBear,
- I think I've responded to all of your questions and objections. I'll look again.
- Please tell me where you disagree.

 

[Abaddon]

Abbadon,
- When flipping a fair coin, what's the probability that it will land heads?

Potato.

Jabba you have tried this line before. It didn't work.

 

[Jabba]

Potato.

Jabba you have tried this line before. It didn't work.

Abaddon,
- Please explain why you say that.

 

[Abaddon]

Abaddon,
- Please explain why you say that.

Don't make me link the posts where your attempted "gotcha" was comprehensively destroyed. Unless you really want to. I will do it and you know I can do it. If you continue to abuse your hosts here and inflict your dumb ideas upon the good and learned folks here, then I have no choice but to share how your argument has been decimated elsewhere. Somehow, you think our hosts here are unaware that your argument fell flat on it's face. I am here to save our hosts some time.

For everyone else, Jabba tried this coin nonsense elsewhere and got nailed. Do not waste time with this nonsense.

 

[Abaddon]

And for the curious, we had five years over on ISF of the "Shroud of Turin is real" baloney. This is what you are getting into with this. Five years.

 

[Jabba]

Abaddon,
- As you know, I disagree.
- Please link whatever posts you want.
- I hope that someone over here is still listening.