On Friday, October 19, 2018 at 5:09:03 AM UTC-6, Martin Brown wrote:

On 18/10/2018 20:38, Gary Harnagel wrote:

On Thursday, October 18, 2018 at 11:14:28 AM UTC-6, Martin Brown wrote:

To use Bayesian statistics on these questions is scientific but you
are then stuck with the problem of dealing with maximally
uninformative "improper" prior probability distributions that are
not normalisable.

P(N, the number of deities in the universe) = 1/N

We should be based on the probability that there is one ancient
civilization that is God; i.e., N = 1.

You are a complete charlatan

No, I am not a charlatan. You are engaged in smearing.

with no understanding of statistics whatsoever.

And again you are wrong. I'm not an expert in statistics but I have
published papers in laser diode lifetesting and lifetime. You seem to
be venting out of some kind of frustration.

Your pathetic attempts at sophistry are at an end.

Just because I disagreed with you? You are being quite petty.

(the same prior applies to any scale factor)

P(x, does God exist 0=no, 1=yes) = 1/(x(1-x))

??? So P(x) = infinity if x = 0 or x = 1? How can you have
probabilities greater than unity? Even x = 0.5 gives P(x) = 4.

That is what makes un-normalisable prior probabilities so tricky.

It's no trick to see that 1/[x(1 - x)] is greater that one for ALL x
between 0 and 1. And I fail to see the meaning of your P(x) equation.