[Merged] Immortality & Bayesian Statistics

Status
Not open for further replies.
Toontown said:
So what now? Another 30 pages of irrelevant one-liners and pointless stonewalling?
Click to expand...

Why not? It's funnier than anything on TV. Somebody tries to prove that an impossibility like immortality exists by using a silly mathematical formula. What did you expect?
 
lenny said:
for the record:

what i'd like is an argument.

we find we have different, conflicting views/conclusions.
we work out a set of assumptions, and a framework.

we agree on some assumptions, disagree on others perhaps and clarify them.

we answer each others questions without insult with minimal diversion (hesitation or repetition : ), and determine whether or not we disagree fundamentally on any critical assumption. if so: we're done, and agree to disagree, and understand why.

if we more or less agree on the assumptions we follow the argument through toward the conclusion, and see exactly where along the way we diverge (as we must since we hold conflicting conclusions)

we learn that:
a) one of us made a mistake, and we're done
b) both of us made mistakes, and we start over. but for this thread we're done
c) an additional assumption on which we disagree, agree to disagree and we're done
d) our conclusions were not, in fact, in contradiction. and we're done

the key is to work together to get to the end, to learn something. which outcome is not as important as getting there.

that's what i like.
Click to expand...
Lenny,
- I totally agree. Have you read my discussions of debate on either of my websites?
 
Who is going to break the news to Jabba that maths has nothing to do with non-existent immortality?
 
Akhenaten said:
I'll give it a crack as soon as I've finished using Pythagoras' theorem to prove the Loch Ness monster is real and the cosine rule to plot the current location of Planet X.
Click to expand...

Silly Pharaoh:

You have to use string theory and translocation to locate Planet X...

Have the Universities in the Middle Kingdom fallen so low?
 
Jabba said:
Jay,
- Sorry about that.
Click to expand...


Instead of posting trite apologies for inconsequential typos, how about addressing the basic flaws in your "case".

Here's a post typical of those that you're wont to put aside in apparent hope that they'll be forgotten while you laughably declare yourself victorious in this shemozzle of a debate:


Jabba said:
godless dave said:
So where do you get those numbers? How did you calculate them?
Click to expand...


Dave,
- I'' start with the easy ones.
- P(SM|k) (the probability of the scientific model being true given existing evidence (before factoring in the implications of my own current existence).) I'm allowing for a 99% probability that the existing model is true -- given the existing considerations. Would you make the probability even larger?
- P(NSM|k) (the probability of the complement of the scientific model being true given existing evidence (before factoring in the implications of my own current existence).) Naturally, that would be 1.00 - 0.99, or 0.01.
- I'll have to get back to you re P(SM|me).
Click to expand...


Allowing for some arbitrary probability without even hinting at how it was calculated and "getting back to us" on another, despite these figures being absolutely crucial to your formula, with or without typos, is about as far from victory as it's possible to get.
 
Jabba said:
Dave,
- I'' start with the easy ones.
- P(SM|k) (the probability of the scientific model being true given existing evidence (before factoring in the implications of my own current existence).) I'm allowing for a 99% probability that the existing model is true -- given the existing considerations. Would you make the probability even larger?
Click to expand...
_

godless dave said:
I have no idea how you would even calculate such a probability.
Click to expand...
_

Dave,
- Based upon our own relevant evidence and intuition, we make our best guesses. Here, I'm simply allowing that the SM might, possibly, be wrong. In my own opinion, however, the prior probability that the SM is correct is much smaller than 99%, but even with a prior probability of 99%, the posterior probability turns out to be an unimaginably small number -- if not, simply, one over infinity.
_

Jabba said:
...- I'll have to get back to you re P(SM|me).
Click to expand...
_

godless dave said:
Considering that is one of the most important parts of your hypothesis, and has been a key sticking point since you started this thread, I would have thought you would have worked this out already.
Click to expand...
_

Dave,
- The likelihood of me currently existing -- given the current scientific model -- would (at best) depend upon one particular sperm cell from all humanity meeting up with one particular ovum from all humanity -- past, present and future. And that's only the tip of a bottomless iceberg! What about all the potential, but unrealized sperm cells and ova, and their potential, but unrealized offspring, generation after unrealized generation? See what I mean?
 
Jabba said:
_
Dave,
- The likelihood of me currently existing -- given the current scientific model -- would (at best) depend upon one particular sperm cell from all humanity meeting up with one particular ovum from all humanity -- past, present and future. And that's only the tip of a bottomless iceberg! What about all the potential, but unrealized sperm cells and ova, and their potential, but unrealized offspring, generation after unrealized generation? See what I mean?
Click to expand...

And how does that make you immortal? We are all the result of a sperm and and ovum meeting. If I catch the local train what are the chances that the first person I see on that train will be that person? See what I mean?
 
Jabba said:
_


_

Dave,
- Based upon our own relevant evidence and intuition, we make our best guesses.
Click to expand...

That's not how statistical analysis is done, not if you want meaningful results.


Jabba said:
Here, I'm simply allowing that the SM might, possibly, be wrong. In my own opinion, however, the prior probability that the SM is correct is much smaller than 99%, but even with a prior probability of 99%, the posterior probability turns out to be an unimaginably small number -- if not, simply, one over infinity.
Click to expand...

How did you calculate those probabilities?

Jabba said:
_

Dave,
- The likelihood of me currently existing -- given the current scientific model -- would (at best) depend upon one particular sperm cell from all humanity meeting up with one particular ovum from all humanity -- past, present and future. And that's only the tip of a bottomless iceberg! What about all the potential, but unrealized sperm cells and ova, and their potential, but unrealized offspring, generation after unrealized generation? See what I mean?
Click to expand...

No, I really don't. You could say that about absolutely any object in the universe. I don't see why it's significant.
 
Last edited:
What will happen to the immortals when the universe ends? Inquiring minds want to know. Could they save themselves by doing a few calculations?
 
Jabba said:
- Based upon our own relevant evidence and intuition, we make our best guesses. Here, I'm simply allowing that the SM might, possibly, be wrong. In my own opinion, however, the prior probability that the SM is correct is much smaller than 99%, but even with a prior probability of 99%, the posterior probability turns out to be an unimaginably small number -- if not, simply, one over infinity...
Click to expand...
godless dave said:
That's not how statistical analysis is done, not if you want meaningful results...
Click to expand...

Dave,
- Sorry about that. I'll remove "intuition."
- Otherwise, as I understand things, that is what is done in Bayes statistics. We may be far off in our judgement about the prior probability of a thesis, but speculation is accepted (and, for what it is) in Bayes statistics (according to my understanding).
- Are you certain that the scientific model is correct? If not, in your own evaluation of my claim, you are forced to speculate.
 
Hi guys,

I do not see much statistics on the last few pages, please let me know if I missed something. Regarding the chi-square discussion

Toontown said:
When are you going to stop fixating on irrelevancies and disprove the validity of the chi-square test? Should be easy. All you have to do is disprove the validity of the concept that an existing observation can have an "expected probability" associated with it. Even after it :boggled: already exists.
Click to expand...

The chi-squared test gives the (observed) observations each probability one. It aims to evaluate the probability of a null hypothesis, not the probability of the observations. So there is no "validity of the concept that an existing observation can have an "expected probability" associated with it” to disprove. The aim is to say something about the null, which in turn is often about the process that generated the obs, not the obs themselves.

If I have misunderstood what you intended, please point me to an example of what you mean in a stats text (any stats text).
 
Last edited:
Can we have a separate thread for all this stuff that has nothing to do with immortality?
 
lenny said:
Hi guys,

I do see much statistics on the last few pages, please let me know if I missed something. Regarding the chi-square discussion



The chi-squared test gives the (observed) observations each probability one. It aims to evaluate the probability of a null hypothesis, not the probability of the observations. So there is no "validity of the concept that an existing observation can have an "expected probability" associated with it” to disprove. The aim is to say something about the null, which in turn is often about the process that generated the obs, not the obs themselves.

If I have misunderstood what you intended, please point me to an example of what you mean in a stats text (any stats text).
Click to expand...

Give me a break. It was 40 years ago that I did this stuff. Is this some kind of test? If not, then you have apparently misunderstood everything.

What does

the sum of (Observed - Expected)2 / Expected

mean?

If, as you say, there is no "validity of the concept that an existing observation can have an "expected probability" associated with it” , then what do you suppose those "expected" values in the formula are?

How could you test the "null" hypothesis if you don't have any expected values? You already have the flies. Without the flies and the probability distribution, you can't do anything. Do do you toss out the expected probability distribution because the flies have already happened? I don't think so.

A fruit fly cross might have an expected ratio of 4:2:2:1
This ratio embodies the expected probability distribution for each offspring, under the heredity hypothesis being tested.

4/9 : 2/9 : 2/9 : 1/9 = 0.44 : 0.22 : 0.22 : 0.11

This is the probability distribution of the expected traits of any random fly in the offspring population. This is not the Copenhagen interpretation of quantum mechanics. Looking at the fly does not make the probability distribution go away. If the hypothesis is correct, then every fly has it's genesis associated with that probability distribution, because the probability distribution reflects the means by which the flies were formed.

If a fly is in the 0.11 class, then it "beat" the odds against being a member of any of the other classes. Think of those other classes as being non-viable, if that helps you understand what "beating the odds" means. In that case, you would be counting a lot of dead flies.

Multiplying these probabilities by the total number of offspring gives the expected totals of each class. The formula then compares these expected totals with the observed totals of existing flies in the prescribed manner, yielding the chi-square value, which is then converted to the probability of chance. The probability of chance is simply the probability that chance alone accounts for the variation between observed and expected totals.
 
Last edited:
Status
Not open for further replies.

Back
Top Bottom