Cont: Proof of Immortality VIII

Jabba said:
Since, in the case of complementary hypotheses, P(E) would be P(E|H)P(H) + P(E|~H)P(~H), I tried to estimate what-specific-hypotheses ~H would include (in this case) and how-probable-each-would-be multiplied by how-probable-each-would-be-to-result-in-E, and added up the products...
- Can you follow that, and does it make sense?
Click to expand...

I can follow it, and no it does not make sense. It doesn't make sense because you're trying to shoehorn one formulation of a solution into a different formulation of the problem. As I begin the list you're so terrified of: you don't know how to formulate a statistical inference.

The formulation you've chosen is "an hypothesis and its complement." And yes, numerous experts have confirmed that you're using the formula that reckons the conditional probability of the hypothesis given an event and incorporates the complement of the hypothesis, and that it's properly derived algebraically from Bayes' theorem. But you wrongly extend that endorsement to mean their approval as the right way to solve your problem. It's not.

If H were the hypothesis whose probability, given data, you were directly interested in, you might have a chance of convincing someone you were working the problem correctly. But that's not what you're doing. You've flipped it around -- the way nearly all fringe theorists err in doing -- and you're trying to establish your desired hypothesis by showing that some other hypothesis is so improbable on its own that it has to be discarded. You think you can do this by statistical legerdemain and formulating it as "a hypothesis and its complement," where "a hypothesis" is your competition and "its complement" is your theory.

But that's not what "a hypothesis and its complement" means, and you finally figured it out. If "a hypothesis" is materialism (which is the only hypothesis you've ever tested), then "its complement" is the set of all hypotheses that aren't materialism. And that set contains hypotheses that don't lead to immortality. That means the probability of immortality, given data, cannot just be 1-P(materialism), as you originally suggested. Your plan was never to compute the probability of immortality given data. It was instead to deduce it based on smack-talking the competition and disguising the exercise as math.

Once you finally figured out that you had to deal with a set of things that weren't materialism, you decided -- in your fumbling, pidgin-statistics way -- that you could fix your proof by enumerating all the hypotheses in the complement that weren't materialism and assigning priors and likelihoods to them in the same out-of-your-fanny way you used for all the other numbers. This is wrong not only because you're just pulling numbers out of your fanny -- thus proving nothing -- but because you don't understand why, fundamentally, you can't take that approach even if you could get actual data for that fan-out of hypotheses.

The reason "the complement" has a sort of special status in statistics is because it's allowed to include the notion of stuff you could never know and therefore never nail down. For that reason you can't -- and shouldn't try -- to exhaustively enumerate them all. You figured this out, which is why the last bit of your enumeration exercise was "whatever I haven't thought of." And you thought you could assign a prior and likelihood function to it. By definition if you can't know what it is, you can't start assigning numbers to it and have the resulting formulation mean anything. It certainly can't prove anything.

But more importantly, by doing what you've done, you've transformed the problem into a set of discrete hypotheses. The "hypothesis and its complement" formulation really doesn't help you there, even if you tried to gimmick it and try the union of them all and the complement of their union. Why? Because the union would contain both hypotheses that lead to immortality and those that don't. Computing one number of them would not discriminate between mortality and immortality.

If you have an incomplete set of discrete hypotheses and some data, what is the best way then to evaluate them statistically? I'll leave that to you, since it's your proof. I know the answer, and I've spelled it out to you several times in the posts you insist you can ignore because I'm so mean to you. But it's not the formulation you're using, I can say that. What I will say, however, is that you'll actually have to compute the probability of immortality given the observation of your current existence, and you won't be able to do it indirectly as one-minus-something.

Even statistics, which deals with uncertainty, doesn't let you cheat the way you've been planning to cheat. Every single statistician you've consulted has told you this.
 
Jabba said:
I need to go back and look up my last attempt to rate OOFLam
Click to expand...

"Rating" something seems to be today's code word for just guessing at all the numbers that pertain to it. You've been told by people you've approached as experts that if you simply make up all the answers, you cannot possibly posture anything that comes out of that exercise as proof of something.

"OOFLam" is an acronym you made up to confuse would-be critics as to what you really mean. None of your critics uses it, and they have asked you nicely to stop using it since it serves no purpose other than facilitating your equivocations from day to day. Your argument does not discuss simply the observation of a finite life, which is what you say the acronym means. We can say that because your argument does not deal with all the hypotheses that would result in observing only a finite life. In fact, it deals with only one such hypothesis -- materialism. Let me remind you there was a period in this debate when you admitted you were only trying to disprove materialism and accepted that that would not prove immortality. Your actual argument has not changed; you've simply walked back the admission of insufficiency and couched the rhetoric in a made-up concept that masks the ongoing equivocation. Your argument is a false dilemma between materialism and immortality.

But more importantly, the statistical behavior of materialism and the observation of your current existence has been explained. Your critics have carefully led you to the conclusions of what those behaviors must be under the rules of mathematics and under the criteria in your model. You agreed to them because you were made to see how they follow from the conditions of your proof. And you did so not realizing what the implications would be for your bottom line. That's the clincher. Your critics here have become more adept at leading you to see what must be, but in a way that doesn't let you realize it until it's too late.

What you're doing now is disturbing. Because the model gives you an answer that's not what you wanted, you want to go back and change all the premises you were carefully led to see must be. You are doing so without any regard for why they are what they are. You think that you can change them without such regard and that your proof remains sound. The only reason you want to change them is because the proof got the wrong answer, not because it's mathematically valid or acceptable to do so. Your critics could ask for no better proof of your intent to mislead and deceive using mathematics.

Those revelations are really why your proof fails. Other people see this right away about you -- you just want to pretend your preconceived beliefs are objectively valid. Well, we saw it too, but we honestly and wrongly believed we could talk you out of this fool's errand. Obviously we cannot, and obviously you're bound and determined to pretend math will give you what you want most. Is that proof of immortality? Nah, nothing so noble. When you're not haunting this forum you tend to be more candid about what you really want: proof that atheists really believe (or should believe) all your spiritual nonsense and that the atheists you encounter who challenge your beliefs are just putting on a show. You're trying to beat atheists at their own game, and you're being really quite silly in your efforts to convince yourself you've succeeded.

...but so far, I still think that what I've done does make sense...
Click to expand...

No. It's been proven mathematically that what you're doing does not make sense. What you think otherwise is irrelevant, except perhaps in proving that you're hopelessly entrenched and that no attempt to educate you can possibly succeed.
 
Jabba said:
- I had assumed that "that other thing" was my attempt to explain my calculation of P(E).
- Since, in the case of complementary hypotheses, P(E) would be P(E|H)P(H) + P(E|~H)P(~H), I tried to estimate what-specific-hypotheses ~H would include (in this case) and how-probable-each-would-be multiplied by how-probable-each-would-be-to-result-in-E, and added up the products...
- Can you follow that, and does it make sense?
Click to expand...

So, still no PDF?
 
Jabba said:
jt,
- I need to go back and look up my last attempt to rate OOFLam -- but so far, I still think that what I've done does make sense...
Click to expand...

Just serves to show that you don't actually read the objections of your critics. No matter what we say, you still think you're correct, despite the fact that you've been shown and proven wrong thousands of times over.

Thousands. Think about that.
 
jt512 said:
You've insisted that P(E)=1. Now that you've finally figured out that P(E) = P(E|H)P(H) + P(E|~H)P(~H), calculate P(E) from your "model." Do you finally see that you've contradicted yourself? "Can you follow that?" "Does it make sesne?"
Click to expand...

Jabba said:
jt,
- I need to go back and look up my last attempt to rate OOFLam -- but so far, I still think that what I've done does make sense...
Click to expand...

Jabba, you've insisted that P(E)=1, but the value of your denominator in Bayes' Theorem, which is P(E), is not 1. So you are claiming both that P(E) equals 1 and does not equal 1. Even an infant would understand that that is a contradiction.
 
Jabba said:
I still think that what I've done does make sense...
Click to expand...

JLkO4l.jpg
 
jt512 said:
Jabba, you've insisted that P(E)=1, but the value of your denominator in Bayes' Theorem, which is P(E), is not 1. So you are claiming both that P(E) equals 1 and does not equal 1. Even an infant would understand that that is a contradiction.
Click to expand...
jt,
- I'm no longer claiming that P(E) = 1.
 
Jabba said:
jt,
- I'm no longer claiming that P(E) = 1.
Click to expand...

How do you justify having claimed it for years? Perhaps it's finally time for you to end the charade that you have any competency in probability. You've never fooled anyone anyway.
 
jt512 said:
How do you justify having claimed it for years? Perhaps it's finally time for you to end the charade that you have any competency in probability. You've never fooled anyone anyway.
Click to expand...

"Darn, I didn't get the answer I wanted. I'll just change the inputs willy-nilly until the answer comes out the way I want. Yay! I proved something! In your face, atheists!"
 
jsfisher said:
So, still no PDF?
Click to expand...
- I still don't understand what you want re the PDF. My best guess was that you were referring to the distribution of P(E), and seems like my calculation provided the mean of that distribution...
 
Jabba said:
- I still don't understand what you want re the PDF. My best guess was that you were referring to the distribution of P(E), and seems like my calculation provided the mean of that distribution...
Click to expand...

You need to stop guessing.

The probability density function is for the fairness of your nickel. It is essential for determining P(H), the probability your coin is fair. (It also plays a role in determining P(~H), but that is unnecessary for the inference.) You just need the PDF, a little bit of combinatorics, and some Calculus to come up with P(H), P(E), and P(E|H). Then it is just plug-and-chug, as they say.
 
Jabba said:
- I still don't understand what you want re the PDF. My best guess was that you were referring to the distribution of P(E), and seems like my calculation provided the mean of that distribution...
Click to expand...

And then I asked you what else you'd need to know about the distribution besides its mean, and what effects those things would have on the model. And you didn't answer, I'm guessing because you don't know.
 
You must log in or register to reply here.

Back
Top Bottom