Proof of Immortality III

[Mojo]

Mojo,
- I've already happened, but the implications of my happening have not been factored into the prior probabilities. That's how it's "new."

Ignoring it while making up your prior probabilities doesn't make it new; it just means that you have not considered it.

And you're still ignoring the point: whichever of the unimaginable number of possibilities resulted, you, or someone in your place, could point at it and say, "wow, dude, this is sooooo unlikely". But all they would have is a result from the set of possible results. It only becomes "special" if something makes it special: some factor that differentiates it from all the other possible results, such as having been specified in advance.

You didn't specify it until some time after 21st November 2012. Do you have any evidence that the event you have specified has happened since then?

 

[JayUtah]

But so far, I still think that I don't need to be a special case in this case.

Asked and answered. You're equivocating "special case."

See if this works.

And then you go on to present an example of reasoning by conditional probability as if you think the problem your critics have with your claim is that they don't understand conditional probability.

Let's see how your abstract analogy resembles your claim regarding immortalilty.

Hypothesis A has a prior probability of 55%; naturally enough, ~A has a prior probability of 45%.

In your example you don't say how you know this. It's assumed that P(A) is known. In your immortality argument you just made up a number and continually beg people to accept it or provide a different one.

We do an experiment and get a result that approaches zero likelihood if A is true.

In your example you don't give the conditional probabilities. You just declare that, having worked the problem, P(E|A) is now some dire value. Similarly in your argument you just make up the conditional probabilities that apply. In fact, in your argument P(E|~H) is undefinable because it is improperly formed as a disjunction.

You were told about this several times by several people, and you have simply ignored that fatal error.

You have to bet the farm on A or ~A -- it's up to you.
- Which would you bet your farm on?

No, we don't have to "bet the farm." You're trying to substitute probability for proof. When I first described the example of a Bayesian search I pointed out that the math doesn't tell where the object is, but rather where the object is most likely to be, given that searches in other areas have failed. It tells us what is prudent to do, not what the facts are. Hence Bayes is not proof of what is. Now you're trying to move the goalposts. You aren't proving whether an immortal soul exists or not. You're suggesting whether it's prudent to believe in an immortal soul given some set of assumptions. That is not proof.

This difference has been hammered home to you more times than I can count. You steadfastly ignore it. It is your prerogative to do so, but when you do then you can't translate your apathy into some accusation of imagined bias on the part of your critics. You don't get to call people closed-minded for not validating your beliefs when you willfully ignore their reasoned objections against it.

 

[JayUtah]

Mojo,
- I've already happened, but the implications of my happening have not been factored into the prior probabilities. That's how it's "new."

Straw man. The implications of your happening are wrongly factored into the likelihood ratio. That's how it's the Texas sharpshooter fallacy. Your error is common and straightforward enough that it has its own name. Several people here have told you this. The outside experts you consulted have told you this. When will it finally sink in?

 

[JayUtah]

But all they would have is a result from the set of possible results. It only becomes "special" if something makes it special: some factor that differentiates it from all the other possible results, such as having been specified in advance.

Yes, this.

Jabba is essentially saying that dice cannot possibly be fair because he proposed to roll a 3 and didn't make it. And he's trying to prove that by saying there was only a 1 in 6 chance of a fair die producing a 3. He's trying to squeak out of that by saying he could just as easily have predetermined a 4 and had the same odds, so his critics are wrong to question his math. The way to think about it is that if you roll a fair die, some number will come up. And 3 has just as much a chance of being produced as 4. Jabba insists that chance is involved in producing him, but he can't figure out that this is the case only if a desired outcome is predetermined. The chance of the roll of the die producing some number when rolled is 1 -- chance is not involved.

Jabba can't see the casino for the crap table, having shoved his head so far below the rails as to miss just how broad and sweeping his error is. The problem is not that he is predetermining the wrong outcome. The problem is that he is predetermining an outcome -- and he can't see how he is, so he denies that he is. This is why some people aren't cut out for statistical reasoning.

 

[Loss Leader]

Pixel,

- I do think that this is the weakest of my interim conclusions. But so far, I still think that I don't need to be a special case in this case.
- See if this works.

- Hypothesis A has a prior probability of 55%; naturally enough, ~A has a prior probability of 45%.
- We do an experiment and get a result that approaches zero likelihood if A is true.
- You have to bet the farm on A or ~A -- it's up to you.
- Which would you bet your farm on?

But you don't know these prior probabilities in your model. You just made some numbers up. If we live in a clockwork universe, the prior probability of you is 1. If we live in a small, cold universe without enough matter to form stars, the prior probability is 0.

We live in a universe where life is possible. That's somewhere between 0 and 1. And there is no way to determine what probability to assign. You are just making it up.

In any case, "bet the house" is not a standard of proof. If you saw 1 coin of unfamiliar design on some alien planet lying face up, there are two possibilities: this is a two-headed coin or that it has two different sides. If it has two heads, the likelihood of it landing on heads is 1, but if it has two sides, the likelihood drops to 1/2.

If you had to bet the farm, how would you bet on the next coin toss?

If your answer is, " I need more information before I bet anything," then you have admitted the weakness in your own argument.

 

[Mojo]

Jabba is essentially saying that dice cannot possibly be fair because he proposed to roll a 3 and didn't make it. And he's trying to prove that by saying there was only a 1 in 6 chance of a fair die producing a 3.

Not even that: he's thrown a die, and it has come up with a number between 1 and 6. He has then claimed that because there is only a one in six chance of this result if the die is fair, it is likely that the die is rigged. He never proposed that his roll of the die would produce any particular number.

And, of course, he is supposed to be trying to prove something entirely other than that the die is rigged.

 

[JayUtah]

If your answer is anything other than, " I need more information before I bet anything," then you have admitted the weakness in your own argument.

You yourself have said it so many times. There's a clear qualitative difference between the questions, "What is prudent to do?" and "What is factually true?" Probabilistic arguments can, at best, recommend an answer to the former, which is not at all the same as the latter.

 

[JayUtah]

Not even that: he's thrown a die, and it has come up with a number between 1 and 6. He has then claimed that because there is only a one in six chance of this result if the die is fair, it is likely that the die is rigged.

Yes, that's the point I was trying to make. You said it better.

And, of course, he is supposed to be trying to prove something entirely other than that the die is rigged.

He wants to style the scientific model as a random variable. And, as you say, he argues the chances of that random variable producing any particular number is so vastly small that the true system has to be rigged -- i.e., by a soul -- in order to have it produce the number. Or that's one way I see it. But yes, false dilemma either way.

 

[jt512]

Pixel,

- I do think that this is the weakest of my interim conclusions. But so far, I still think that I don't need to be a special case in this case.
- See if this works.

- Hypothesis A has a prior probability of 55%; naturally enough, ~A has a prior probability of 45%.
- We do an experiment and get a result that approaches zero likelihood if A is true.
- You have to bet the farm on A or ~A -- it's up to you.
- Which would you bet your farm on?

Jabba, your post suggests that you don't understand a fundamental principle of Bayesian inference, namely, that the amount of change in probability from prior to posterior depends on the relative probability of the evidence under each hypothesis. Your question is unanswerable because you did not say how likely the evidence is under ~A. Even though the evidence is improbable under A, it could be even less probable under ~A, making the posterior probability of ~A greater than that of A, and hence making A the better bet.

As for the you-being-special fallacy, it has been explained to you an uncountable number of times, including at least once by me. The "special" fallacy is easier to see if you apply your formula to someone other than yourself. Whether or not you think you're special, hopefully you have the humility to admit that you or no more or less special than any other individual. Therefore, you should arrive at the same posterior whether you apply your formula to you or someone else. Therefore, let's apply your formula to your nearest neighbor, the person living geographically closest to you, but not in your own home. Let's say this person is "Jack." Under the hypothesis that Jack was randomly produced the probability of his existence would be tiny (according to your "scientific" model), but the fallacy is that even if Jack had never been born, you would always have some nearest neighbor (unless you were literally the last person on earth), thus no matter who your nearest neighbor was, you would always observe a nearest neighbor. Therefore, your evidence is not that Jack is a nearest neighbor; it is that you have some nearest neighbor. That likelihood is basically 1 under both of your hypotheses. Therefore the likelihood ratio is 1, and the posterior probabilities of your hypotheses equal their priors. The evidence has no power to discriminate between the hypotheses.

It is confusing when you apply the formula to yourself because it is unclear what the alternative would be if you had not been born. But, when you attempt to apply the formula to someone else it is easy to see that if that specific person wasn't there, then you would simply apply the formula to someone else. There would always be someone to apply the formula to; hence the conditional probability of the evidence is 1 under any hypothesis about a world in which there is someone to ask the question.

It's even easier to understand the fallacy using a deck of cards. You shuffle a deck, deal out all the cards, and observe the sequence 2H, KC, AS, 5H, 5D, 7H, . . . . The probability of that particular sequence is 1 in 8 × 10⁶⁷, or in your terms "essentially zero." Yet if the shuffle hadn't produced that particular sequence it would have produced some other one. Do you think shuffle sequences have souls? I doubt it. Hopefully you can see the fallacy. You would have observed some sequence. Therefore, your actual evidence is simply that a sequence was observed, not a specific sequence. The conditional probability that a sequence is observed under the hypothesis that the shuffle was random is 1. Analogously, the conditional probability that you would observe some human being under the hypothesis that individuals are randomly produced is 1. It's just like the shuffle.

 

[caveman1917]

You yourself have said it so many times. There's a clear qualitative difference between the questions, "What is prudent to do?" and "What is factually true?" Probabilistic arguments can, at best, recommend an answer to the former, which is not at all the same as the latter.

Do you have an example of something which is "factually true"? Do you have an example of an argument which establishes something to be "factually true"?

 

[Mojo]

Even though the evidence is improbable under ~A, it could be even less probable under A...

Jabba's preferred hypothesis is even less probable than the one he is arguing against. The hypothesis he is arguing against requires his body to exist; his preferred hypothesis requires his body to exist, his immortal soul to exist, and his immortal soul to be associated with his body.

There is no way his preferred hypothesis can be more likely than the one he is arguing against.

 

[JoeMorgue]

Here's an interesting little math fact.

If I take a normal, everyday, unmodified, run of the mill, off the shelf deck of 52 playing cards and shuffle it well I have within statistical certainty created a sequence of 52 cards that has never yet appeared and never will appear in the universe. There are more ways to arrange a deck of 52 cards than they are atoms on Earth.

A well shuffled deck of cards is an amazingly unique event that would be for all practical purpose impossible to recreate through random chance.

Jabba. Does a well shuffled deck of cards have a soul?

*Reference: http://gizmodo.com/there-are-more-ways-to-arrange-a-deck-of-cards-than-ato-1553612843

 

[Mojo]

Hopefully you can see the fallacy.

I'm sure he can. I am also certain that something that Upton Sinclair said about men and their salaries applies to Jabba and his argument.

 

[Jrrarglblarg]

Wrong again, Bob.

http://yourbrainhealth.com.au/left-brain-right-brain-myth/

Don't lose the baby with that bathwater. From your link:

University of Utah neuroscientist Dr Jeff Anderson, lead author of the study, says,

It’s absolutely true that some brain functions occur in one or the other side of the brain. Language tends to be on the left, attention more on the right. But people don’t tend to have a stronger left- or right-sided brain network.

 

[Mojo]

But Jabba actually was claiming that some people have a stronger left- or right-sided brain network. Your quotation says that I Am The Scum was correct.

 

[jt512]

Even though the evidence is improbable under A, it could be even less probable under ~A,

Jabba's preferred hypothesis is even less probable than the one he is arguing against. The hypothesis he is arguing against requires his body to exist; his preferred hypothesis requires his body to exist, his immortal soul to exist, and his immortal soul to be associated with his body.

There is no way his preferred hypothesis can be more likely than the one he is arguing against.

The phrase of mine you quoted is discussing the probability the evidence given alternative hypotheses, but you're talking about the probability of the hypotheses themselves (either the prior or the posterior, it's not clear which).

 

[jt512]

You yourself have said it so many times. There's a clear qualitative difference between the questions, "What is prudent to do?" and "What is factually true?" Probabilistic arguments can, at best, recommend an answer to the former, which is not at all the same as the latter.

The conclusions of most scientific experiments are based on probabilistic arguments. Drug A works better than Drug B, p<.05. We discovered the Higgs boson, p<10^–6. These are claims of fact based on probability.

 

[Mojo]

The phrase of mine you quoted is discussing the probability the evidence given alternative hypotheses, but you're talking about the probability of the hypotheses themselves (either the prior or the posterior, it's not clear which).

Yes, I didn't put it very well. What I was trying to say was that the likelihood of Jabba's existence under his preferred hypothesis is even less than it is under the hypothesis he is arguing against, for the reason given, and that therefore by Jabba's argument his preferred hypothesis is less likely than the alternative.

I blame the beer.

 

[godless dave]

Mojo,
- I've already happened, but the implications of my happening have not been factored into the prior probabilities. That's how it's "new."

What implications?

 

[JayUtah]

The conclusions of most scientific experiments are based on probabilistic arguments. Drug A works better than Drug B, p<.05. We discovered the Higgs boson, p<10^–6. These are claims of fact based on probability.

That's correct, however in the former case the conclusion that Drug A works better than Drug B has a probabilistic component only because of our inability to measure accurately all the factors that apply to drug testing, the nature of sampling, and those attendant concerns. The actual proof of efficacy is based on observations which in turn come from a qualitative biochemical rationale. We observe abated symptoms. which are qualitative facts we measure -- but only inexactly, hence statistics.

Ditto the measurements in practical physics. The argument prior to its measurement was that the Higgs boson should exist based on the purely qualitative characteristics of a model that predicted it, and which otherwise served qualitatively as an adequate model of particle behavior. The measurements of the observations at the hadron collider -- being uncertain numbers -- have a statistical component. That is not the same as the model being statistical in nature or the proof being the conclusion of probability computation absent observation and measurement.

I view Jabba's exercise as inherently different because his conclusion is not a quantity derived by qualitatively supportable means, to which a statistical component of confidence is attached. It is instead a statement of belief based solely on a computation of probability that a certain thing should be the case.