Proof of Immortality, the IV league

[jt512]

If you want general criticism of the method itself, here's a paper from 2008
My favorite has always been along the lines of, "If I'm going to guess the priors anyhow, why not save time and just guess the posteriors too?" After all, any posterior is just the prior for a different calculation.

That's cute, and if you think it a serious objection then you should be glad about calculating Bayes factors, because they quantify the weight of the evidence independently (in case I haven't mentioned it) of the probabilities of the hypotheses.

Write it as often as you like. Heck, eventually it might actually apply to the point someone else is trying to make. It's actually another illustration of what I've been hammering on. You have a statement which is true in a certain context and to preserve that truth value, the context has to be managed to fit.

Flash: Guy on internet says Bayes' Theorem not really a theorem; only sometimes works.

 

[jt512]

Is this your peculiar version of Euler's "Sir, [a+bⁿ]/n=x, hence God exists—reply!"? Do you expect everyone to flee Russia, like Diderot?

Your ridiculous un-statistical anti-scientific glossolalia has gone too far!

Great argument!

 

[aleCcowaN]

Great argument!

It is convenient for you to notice it.

I don't know what you wrote before, but you are embarked in a large chain of poppycock posting lately.

 

[aleCcowaN]

The frequentist approach to immortality ... what a quackery!

 

[marplots]

Flash: Guy on internet says Bayes' Theorem not really a theorem; only sometimes works.

Aw shucks. And here I thought I had one of those teachable moments.

I've made this mistake before - tried to talk Fibonacci numbers with a quant. Maybe math makes people irritable.

 

[aleCcowaN]

We have two hypotheses, H1 and H2, and we have observed some data D. Each hypothesis makes a quantitative prediction about the probability of D given the model...

This is the kind of script a Hollywood writer uses in a class-C TV show to suggest something "scientific" is going on.

Ridiculous wording like "each hypothesis makes a quantitative prediction" shows the king is not only naked but more along the lines of the a king's fool.

 

[jt512]

Aw shucks. And here I thought I had one of those teachable moments.

I've made this mistake before - tried to talk Fibonacci numbers with a quant. Maybe math makes people irritable.

No. Math demands precise thinking, which in turn demands precise language. Maybe you think the statement, "You have a statement which is true in a certain context and to preserve that truth value, the context has to be managed to fit," is meaningful. But I look at it and ask myself, how can a mathematical theorem only be true in a certain context, after all, it's a theorem; what does he mean by "context;" anyway; what does he mean by "truth value"; and what does it mean to "manage a context"? And that's just one sentence.

 

[marplots]

No. Math demands precise thinking, which in turn demands precise language. Maybe you think the statement, "You have a statement which is true in a certain context and to preserve that truth value, the context has to be managed to fit," is meaningful. But I look at it and ask myself, how can a mathematical theorem only be true in a certain context, after all, it's a theorem; what does he mean by "context;" anyway; what does he mean by "truth value"; and what does it mean to "manage a context"? And that's just one sentence.

Here's a mathematical theorem which is true in one context but not another:
The sum of the interior angles of any triangle is 180 degrees.

That's a perfectly good and true theorem in Euclidean geometry.

But I wasn't really arguing that Bayes wasn't true so much as it isn't helpful in answering the question about evidence or the value of evidence. This despite the Bayes factor and the weight of the evidence being independent of the priors. The question is more about the meaning of evidence itself - which is not something Bayes addresses. Or does it? Is there an accepted meaning for the word in mathematics, and if there is, is it a meaning we can all agree on?

For me, the questions are philosophical in nature, but if mathematics has something useful to say, I certainly want to hear it.

 

[aleCcowaN]

No. Math demands precise thinking, which in turn demands precise language.

How ironic coming that from you! (besides it being another cinematographic phrase)

Have you the habit of using narrow formulas like this one of yours

to prove fantasies out of its scope?

 

[jt512]

Here's a mathematical theorem which is true in one context but not another:
The sum of the interior angles of any triangle is 180 degrees.

That's a perfectly good and true theorem in Euclidean geometry.

But that theorem doesn't even exist in systems of geometry in which it isn't true. Rather, the axioms of Euclidean geometry lead to that theorem. The axioms of other geometries do not.

But I wasn't really arguing that Bayes wasn't true so much as it isn't helpful in answering the question about evidence or the value of evidence. This despite the Bayes factor and the weight of the evidence being independent of the priors. The question is more about the meaning of evidence itself - which is not something Bayes addresses. Or does it? Is there an accepted meaning for the word in mathematics, and if there is, is it a meaning we can all agree on?

I am saying that X is evidence for Y vs Z if and only if observation of X increases the probability of Y relative to the probability of Z. If you can think of a context in which that is unreasonable, then please state the context in unambiguous terms and explain clearly why you think it is unreasonable.

 

[JoeMorgue]

And this is all missing the point.

You can't use probability to prove something that has or hasn't already happened.

I don't care how much Jabba or anyone could finagle the numbers to show X probability of this or Y probability of that it becomes meaningless when real world events happen.

People die. Arguing about the probability of it after it's already proven is pointless.

Trying to use probability to argue mortality is like arguing that Hillary Clinton won the election because all the polls before the election showed her in the lead.

Right now the death rate for humans stands at a steady 100% with roughly a jumbillion case studies.

I think we're good on our sample size.

 

[marplots]

I am saying that X is evidence for Y vs Z if and only if observation of X increases the probability of Y relative to the probability of Z. If you can think of a context in which that is unreasonable, then please state the context in unambiguous terms and explain clearly why you think it is unreasonable.

Several contexts:

1) X, Y and Z may be ill defined.
2) The word "observation" is ambiguous in some contexts. Specifically, we do not generally accept observations using methodologies like remote viewing, dream states, or from untrusted proxies.
3) In some contexts a mathematical increase in probability cannot be translated into frequentist terms, such as with binary outcomes. So, for example, there is no meaning or testable difference between the statements: Hilary has a 60% chance of winning the election tomorrow and Hilary has a 65% chance of winning the election tomorrow. (See next for subjective version instead of frequentist.)
4) In cases of belief (the closest to this thread), there is no value/meaning in saying I believe Y 20% but this new evidence makes me believe Y 25%. In many contexts we believe something or we do not in a binary manner.
5) The evidential value of an observation of X (increase or decrease probability) may be dependent on other observations. Since you wanted clear examples, I'll construct one:

a) The murder was committed with a gun (given)
b) The suspect has a gun (observation X)
c) The observation X has evidential value because it increases the probability of Y (he did it) over Z (he didn't do it).
d) The suspect is paralyzed from the neck down (observation W)

What happens to the evidential value of X? The fact that the suspect had a gun is suddenly no longer evidence at all, since the suspect is incapable of using a gun. We still observe X, but W has changed the meaning of X and X is no longer evidence of our proposition. W now has evidential value and influences our probabilities, but then along comes V, and so on.

The point is that X satisfies the conditions but in the real world, X comes with baggage, context. And the important thing is that we, the experimenters, often choose or limit the context.

 

[marplots]

People die. Arguing about the probability of it after it's already proven is pointless.

Trying to use probability to argue mortality is like arguing that Hillary Clinton won the election because all the polls before the election showed her in the lead.

Right now the death rate for humans stands at a steady 100% with roughly a jumbillion case studies.

I think we're good on our sample size.

But Jesus lives in my heart!

 

[aleCcowaN]

What a silly madness is going on in this thread. Using Bayes! Conveniently hiding what is essential for it to be valid: exhaustive events, mutually exclusive in pairs, with a probability that is not zero.

"I'm gonna depart from the probability of eternal life being different than 0, assign arbitrary values to all probabilities, dismiss everything that can point it isn't exhaustive and prove that pet concept to be a concrete value different than 0. My next venture is proving that masturbation can get you pregnant, for what I'll depart from a 22.37% probability of the Holly Spirit entering the subject"

 

[Mojo]

What a silly madness is going on in this thread. Using Bayes! Conveniently hiding what is essential for it to be valid: exhaustive events, mutually exclusive in pairs, with a probability that is not zero.

"I'm gonna depart from the probability of eternal life being different than 0, assign arbitrary values to all probabilities, dismiss everything that can point it isn't exhaustive and prove that pet concept to be a concrete value different than 0. My next venture is proving that masturbation can get you pregnant, for what I'll depart from a 22.37% probability of the Holly Spirit entering the subject"

Which would obviously result in Christmas.

 

[Slowvehicle]

What a silly madness is going on in this thread. Using Bayes! Conveniently hiding what is essential for it to be valid: exhaustive events, mutually exclusive in pairs, with a probability that is not zero.

"I'm gonna depart from the probability of eternal life being different than 0, assign arbitrary values to all probabilities, dismiss everything that can point it isn't exhaustive and prove that pet concept to be a concrete value different than 0. My next venture is proving that masturbation can get you pregnant, for what I'll depart from a 22.37% probability of the Holly Spirit entering the subject"

 

[Slowvehicle]

My Dear Mr. Savage:

Shall I appeal to a moderator to encourage you to respond to me?

 

[John Jones]

PM him and instruct him to answer on-board.

 

[Jabba]

Jabba, you can't prove the existence of immaterial souls using an argument that relies on assuming the existence of immaterial souls.

Mojo,
- Why do you think that's what I'm trying to do?

Because that's where you get the "infinity" in your "seven billion over infinity".

Mojo,
- You and I are 2 of Susan Blackmore's illusional selves. According to OOFLam, each of our particular illusional selves can exist only once -- if at all. There is nothing to say that you or I had to ever exist. How many potential illusional selves might there be?

 

[Jabba]

Fudbucker,
- Are you still around?