Randi's involvement in tests

[steenkh]

It is stated clearly in the challenge rules that Randi is not directly involved in the tests - unless he is specifically asked to participate by the claimant.

However, we have also read the story of the Russian girl who apparently could read through a blindfold. It was my impression from reading this account that Randi was present and was the one who had the goggles taped to her face. Was he specifically asked for in this case? I seem to remember other cases like it.

Although the rule that Randi should not be present is understandable, it also raises the possibility that someone will run off with the million, because Randi is undoubtedly one of the most experienced testers. How does Randi make sure that the testers employed are people who will not fall for cheap tricks? (Images of the Yellow Bamboo test again comes to mind).

[Carn]

Not letting Gary Schwartz conduct any tests?

Carn

[CFLarsen]

Quote:

Although others may become involved in the examination of claims and may add their reward money to the total in certain circumstances, James Randi (via the JREF) will carry out the implementation and management of the challenge.
Terms of the Challenge

[steenkh]

Hm, I think the rules have been changed lately. I do not remember having seen this rule before:

Quote:

4. No part of the testing procedure may be changed in any way without the agreement of all parties concerned. JR may be present at some preliminary or formal tests, but will not interact with the materials used.

Kramer states in this thread that:

(concerning independent investigators who lose interest in the task)

Quote:

We experience this a lot, and this most certainly leads many applicants to the conclusion that JREF is "jerking them around", forgetting that no JREF representative is involved in testing, and that tests are determined with Randi's approval, but without his direct involvement, in order to insure absolute impartiality in the testing procedure.

[Hellbound]

steen:

I think Kramar was referring to the preliminary tests with that statement; of course, I don't speak for him, so that's just my take on it

[KRAMER]

Randi does not conduct the test.

I think also that a good many instances in the past (wherein people underwent testing by Randi himself ) took place without an official JREF Challenge application having been filed. So, Randi often tests claims that have nothing to do with the actual JREF Challenge. When an application is filed, the Challenge rules are followed to the letter.

[steenkh]

OK, thanks, KRAMER!

Then the other part of my question remains: How does Randi ensure that the testers involved will not get fooled? What kind of qualifications do you demand of the testers?

[Stitch]

Quote:

Originally posted by steenkh
OK, thanks, KRAMER!

Then the other part of my question remains: How does Randi ensure that the testers involved will not get fooled? What kind of qualifications do you demand of the testers?

The testers are not actually that important IMHO. Devising a protocol that eliminates the possibility of cheating is however. Once you have that, a fellow beliver could perform the double blind testing. Hence why Randi no doubt oversees and has to agree the protocol with the claimant.

[KRAMER]

Quote:

Originally posted by Stitch
Devising a protocol that eliminates the possibility of cheating is however. Once you have that, a fellow beliver could perform the double blind testing. Hence why Randi no doubt oversees and has to agree the protocol with the claimant.

Precisely. Randi approves all test protocols and advises the investigators as to how to watch out for deception. If they comprehend the caveats Randi offers them, they are "qualified".

[Peter Morris]

Quote:

Originally posted by lackey KRAMER
Randi does not conduct the test.

Except in cases when he does. conduct the test.

Such as Mike G's dowsing test where Randi describes his own involvement.

And the psychic healer from Lithuania. And Randi's stated intention that he would conduct the test "this week."

And the aforementioned testing of the Russian x-ray vision girl, test performed by Randi.

And .... Well, you get the idea.

Actually, I'm hard pressed to find a single description of a million doller test that WASN'T conducted by Randi.

Perhaps you can cite a few examples?

Quote:

I think also that a good many instances in the past (wherein people underwent testing by Randi himself ) took place without an official JREF Challenge application having been filed.

The one's I've cited certainly official applications had been filed. The Lithuanians turned up unexpectedly, but Randi agreed to test them - AFTER an official application had been filed. Mike G's test was definitely official.

Quote:

So, Randi often tests claims that have nothing to do with the actual JREF Challenge. When an application is filed, the Challenge rules are followed to the letter.

Are you sure?

[Carn]

Quote:

Originally posted by Peter Morris


Actually, I'm hard pressed to find a single description of a million doller test that WASN'T conducted by Randi.

Perhaps you can cite a few examples?

Easy to do, if you accept german sites:

http://www.gwup.org/psitest/

There are several links to german newspapers and TV stations, that reported about the tests and all state that tests were conducted by german sceptics and Randi is only named as the one offering the 1 million.

On a report i saw on TV, it was also stated and showed on video, that german sceptics conducted the tests at a university and that Randi was not present, except for the retesting of one candidate, who scored good in the first test(34 of 50, while for passing he would have needed 40 of 50, chance expectation was 25 of 50).
German sceptics decided it was interesting enough to test again and Randi also thought it interesting enough to be present at the test, though he only observed and did not conduct anything. The second test result was 22 of 50, so likely candidate was only lucky on first try.(or it were Randis mystical anti-para powers working)

So all official prelimenary tests were done without Randi, the retesting was just interest of german sceptics.

Especially take a look at the fotos, no Randi to be seen.

Carn

[Peter Morris]

Quote:

Originally posted by Carn
Easy to do, if you accept german sites:

http://www.gwup.org/psitest/

I don't read German. Got any English ones?

NB Randi only accepts claim applications written in English, and rejects babelfish translations. Allow me the same.

Quote:

On a report i saw on TV, it was also stated and showed on video, that german sceptics conducted the tests at a university and that Randi was not present, except for the retesting of one candidate, who scored good in the first test(34 of 50, while for passing he would have needed 40 of 50, chance expectation was 25 of 50).
German sceptics decided it was interesting enough to test again and Randi also thought it interesting enough to be present at the test, though he only observed and did not conduct anything. The second test result was 22 of 50, so likely candidate was only lucky on first try.(or it were Randis mystical anti-para powers working)

Carn

Assuming that's an accurate desvcription, then I approve of that. It's something that Randi gets wrong. When a test subject scores higher than chance there are sevceral possibilities :

- fluke
- some flaw in the test design that skews the results
- deliberate cheating
- some real talent

When it happens in Randi's test, his response is to crow about it falling short of the target score, and deny any further test. This is totally wrong. These Germans, noting a higher than expected score tried the same guy again, to see if he could repeat the higher than expected score. That is the right way to do it.

[Carn]

Quote:

Originally posted by Peter Morris
I don't read German. Got any English ones?

NB Randi only accepts claim applications written in English, and rejects babelfish translations. Allow me the same.

Sorry, doesn't seem to exist in English and Randi does not tell anything in his commentary.

Quote:

Assuming that's an accurate desvcription, then I approve of that. It's something that Randi gets wrong. When a test subject scores higher than chance there are sevceral possibilities :

- fluke
- some flaw in the test design that skews the results
- deliberate cheating
- some real talent

When it happens in Randi's test, his response is to crow about it falling short of the target score, and deny any further test. This is totally wrong. These Germans, noting a higher than expected score tried the same guy again, to see if he could repeat the higher than expected score. That is the right way to do it.

I would not be certain about Randi "crowing" about applicants falling short of the target score, but i am at least certain that he does so with applicants who only score little better than chance and are still certain they have some ability.
As good scoring in tests does not happen often(after all its unlikely if its just chance), i do not know further examples, where a applicant scored good, but not good enough, so i do not know how Randi behaves in this cases, do you have links about such cases where applicant scored far better than chance?

Also in the report i saw and the newspaper articles, there was no hint that Randi in any way disproved the retesting and his coming to watch the retest, i think can only be interpreted as Randi wanting to know whether it was chance, cheating, real ability or test problems, so i cannot agree to your point, that Randi, as you implied, "always" makes fun of failed applicants, maybe he only picks those, where he also has other reasons to "crow" about.

But without lots of reports on prelim tests i can only guess and nothing more can you do, unless you know reports of tests where applicants scored good and where "crowed".

Also he does not deny any further test, though he sets the bar, that applicant has to wait 12 months and reapply(from official rules):
"There is no limit on the number of times an applicant may re-apply, but re-application can take place only after 12 months have elapsed since the preliminary test."

It seems to me that an applicant can even reapply with exactly the same claim and ask for exactly the same test protocol.

But i agree that waiting 12 months and reapply might demotivate people to retest, but that still different from "deny any further test".

BTW, i'm still waiting for your to prove through cites or something like that, that:

1."All scientist, except Einstein, ignored the(=all?) facts"

2."No scientist, except Einstein, did alter the theory to make it fit the facts."

at the thread http://forums.randi.org/showthread.p...0&pagenumber=3

I'm interested, because i think it might be possible to prove 2., but i think its impossible to prove 1.

Carn

Added: I insist so much on discussing those statements of you, that seem to have an "always" or "never" in it, because i think they are among those, that cause a lot of people to dislike you, first because, while everybody knows for example, that Randi can be very nasty, nobody thinks he always is and second its very hard to prove "always" and "never" statements and you do not seem to manage often and third disproving such statements is very easy, one conterexample is enough and people regularly try to show you one and you, as far as i have seen, never acknowledged, that somebody has found an counterexample or something thats close to counterexample.
I think it would be better, if you try to be careful with "always" and "never", i also got a bloody nose somewhere on this forum with that and had to change my position.

But maybe i just get the wrong picture.

[Peter Morris]

Quote:

Originally posted by Carn
As good scoring in tests does not happen often(after all its unlikely if its just chance), i do not know further examples, where a applicant scored good, but not good enough, so i do not know how Randi behaves in this cases, do you have links about such cases where applicant scored far better than chance?

Yeah, here's one.

Expected score : 10%
Actual score (for water dowsers) : 22%
Odds against this happening : approx 107 to 1

Randi's excuses :

1) To combine the results of three totally seperate tests. Different people were testing for different claims. They have nothing to do with each other. By this trick Randi makes the 22% figure vanish.

2) He crows about the figure falling below the target score of 80%. He makes loads of nasty remarks about how they thought they would get 80%, but they really only got 22%, much less than they thought. By this trick, Randi distracts from the fact that they scored substantally above chance.

Let me repeat. there are several possible explainations for this.

- fluke
- some flaw in the test skewing the results
- deliberate cheating
- actual dowsing skill

An honest person running getting this result would have said : Well, this sort of thing happens ocasionally if you run enigh tests

A competent person getting this result would have performed further tests, to see if the 22% success could be repeated.

Randi is neither competent nor honest. His response to this was to play with figures and deny that the score was above chance at all.

Quote:

Also he does not deny any further test, though he sets the bar, that applicant has to wait 12 months and reapply(from official rules):

That's not the same thing. I'm talking about additional trials as part of the same test. You are talking about a totally seperate test.

The Germans got it right. When they saw a result substantially above chance, they tried it again. Thus they were able to say the first test was just a fluke.

Randi got it wrong. When he saw a result substantially above chance, he simply denied it. And offered no further tests.

Quote:

BTW, i'm still waiting for your to prove through cites or something like that, that:

1."All scientist, except Einstein, ignored the(=all?) facts"

2."No scientist, except Einstein, did alter the theory to make it fit the facts.".

Einstein discovered relativity. Everyone else didn't. What more do you want?

Ask a silly question...

Quote:

Added: I insist so much on discussing those statements of you, that seem to have an "always" or "never" in it, because i think they are among those, that cause a lot of people to dislike you,

The reason people dislike me is that they want to belive Randi is some sort of hero, but I keep showing how dishonest he is. They really hate me for that.

Quote:

first because, while everybody knows for example, that Randi can be very nasty,

Except for Larsen who refuses to believe it.

Quote:

nobody thinks he always is and second its very hard to prove "always" and "never" statements and you do not seem to manage often and third disproving such statements is very easy, one conterexample is enough and people regularly try to show you one and you, as far as i have seen, never acknowledged, that somebody has found an counterexample or something thats close to counterexample.

Well, it is pretty rare. Randi fans are a strange bunch, and their ideas of what a counter-example is are peculiar to say the least. Very often the examples they cite prove the exact opposite.

[The Don]

Or the 22% could be the result of choosing a subset fo the tests (which I guess is a combination of sheer dumb luck and flaw in the design of the experiment).

[SpaceFluffer]

Quote:

Originally posted by Peter Morris
Yeah, here's one.

Expected score : 10%
Actual score (for water dowsers) : 22%
Odds against this happening : approx 107 to 1

It's not clear to me from this description how the percentage success values were arrived at. Anyone know?

How did you determine that the "Odds against this happening" are 107 to 1?

[Dr Adequate]

Quote:

Originally posted by Peter Morris
When it happens in Randi's test, his response is to crow about it falling short of the target score, and deny any further test. This is totally wrong. These Germans, noting a higher than expected score tried the same guy again, to see if he could repeat the higher than expected score. That is the right way to do it.

And is apparently what Randi did. Are you fawning on him, by any chance?

Quote:

Originally posted by Carn
On a report i saw on TV, it was also stated and showed on video, that german sceptics conducted the tests at a university and that Randi was not present, except for the retesting of one candidate, who scored good in the first test(34 of 50, while for passing he would have needed 40 of 50, chance expectation was 25 of 50).

German sceptics decided it was interesting enough to test again and Randi also thought it interesting enough to be present at the test, though he only observed and did not conduct anything. The second test result was 22 of 50, so likely candidate was only lucky on first try.(or it were Randis mystical anti-para powers working)

Of course Randi doesn't normally attend tests, but a hot chance like this is worth going out of your way to see, IMO.

[Carn]

Carn quote: BTW, i'm still waiting for your to prove through cites or something like that, that:

1."All scientist, except Einstein, ignored the(=all?) facts"

2."No scientist, except Einstein, did alter the theory to make it fit the facts.".

Peter Morris
"
Einstein discovered relativity. Everyone else didn't. What more do you want?

Ask a silly question...
"

Ok, Peter Morris, since 30 years science faces the problem, that the QCD standard model does not describe gravity in any satisfactory way. When you consider gravity, like its explained in General Relativity, QCD is just plain wrong.
It is certain that either QCD or GR or both are wrong at some point and that maybe a complete new idea is needed to get a consitent theory.
One try is super string theory. Unfortunately, its not yet testible, there are no big enough accelerators and also there are some reasons to believe its actually wrong.

Now consider in 40 years some genius shows up and solves the problem with a completely new theory.

Would you then say:

- All scientist between 1970 and 2040 ignored the fact that GR and QCD don't work together and no scientists did alter the theories, to make them fit the facts.

-Scientist knew there was something wrong, but all the ideas they had, how to deal with this problem failed, they were too stupid and the genius was needed to find a good theory.

I hope you see that the first statement is identical to the one you made for 19th century scientist, while the second statement is different, because it implies that scientist did not ignore the facts, though were too stupid to find a fitting theory, so both statements cannot be correct the same time.

Carn

[Peter Morris]

Quote:

Originally posted by Dr Adequate
And is apparently what Randi did. Are you fawning on him, by any chance?

Of course Randi doesn't normally attend tests, but a hot chance like this is worth going out of your way to see, IMO.

[Peter Morris]

Quote:

Originally posted by SpaceFluffer
It's not clear to me from this description how the percentage success values were arrived at. Anyone know?

How did you determine that the "Odds against this happening" are 107 to 1?

First of all I knocked up a quick computer simulation, ran it several hundred thousand times. This showed me the figure was slightly more than 100-1, maybe about 110-1.

Then I posted the figures on a maths forum and asked for an exact answer. Someone posted 0.935%. When I asked how this was calculated he replied "It's a binomial distribution, conveniently it's a function in excel."

So now we know.

[Peter Morris]

Quote:

Originally posted by Carn

I hope you see that the first statement is identical to the one you made for 19th century scientist, while the second statement is different, because it implies that scientist did not ignore the facts, though were too stupid to find a fitting theory, so both statements cannot be correct the same time.

Carn

Don't you think you're being a bit obsessive about this?

I can no longer even remember the point we were discussing when this first came up. Just get over it. We've moved onto other subjects.

[Zombified]

Quote:

Originally posted by Peter Morris
Don't you think you're being a bit obsessive about this?

Have you ever spent more than a couple posts on anything other than trying to discredit the challenge?

[CFLarsen]

Quote:

Originally posted by Peter Morris
First of all I knocked up a quick computer simulation, ran it several hundred thousand times. This showed me the figure was slightly more than 100-1, maybe about 110-1.

Can we see the source code for this?

[JimTheBrit]

Quote:

Originally posted by Peter Morris
When it happens in Randi's test, his response is to [...] deny any further test.

BS, see rule 5.

Quote:

This is totally wrong. These Germans, noting a higher than expected score tried the same guy again, to see if he could repeat the higher than expected score. That is the right way to do it.

The right way to do what? The million dollars is awarded for the successful demonstration of a claim. If the challengers can't demonstrate their abilities to the previously, mutually agreed on criteria for success, they fail. If others contend that the failure might yet have been successful enough to perhaps show evidence of paranormality, there's nothing to stop the claimant being tested by other parties. Neither is there anything to stop the claimant revising their claim and resubmitting it for testing at a later date.



*******************************************

That was their choice. If they choose to have me do
the tests -- and in
some cases, they INSIST on it -- I do so. That
applies to all aspects, to
suit the applicant. As for the escrow matter, all that
has been posted on my
web site.

James Randi

-----Original Message-----
From: Jim S [mailto:*EMAIL*ADDRESS*DELETED*]
Sent: Tuesday, March 16, 2004 4:02 PM
To: randi@randi.org
Subject: Concerns about your writings regarding the
challenge

Dear Mr Randi,

I’ve recently been reading through the commentary
archive and have noticed that a number of your
statements regarding the paranormal challenge seem
contradictory. I’ve listed below those that have me
puzzled, in the hope that you can address them and
restore my confidence in your award.

-Repeatedly, you state that the JREF is not involved
in the testing, which is completed by a mutually
agreed independent third party.

“The JREF does not involve itself in the testing
procedure, other than helping to design the protocol
and approving the conditions under which a test will
take place.� http://www.randi.org/research/index.html

Yet in the commentaries of 22nd Feb 2002
(http://www.randi.org/jr/022202.html) and 1st Oct 2000
(http://www.randi.org/jr/10-01-2000.html), you
describe your direct involvement in the testing of
Natalie Lulova and a healer from Lithuania (an
apparent contravention of rule 4 of the challenge
rules). Were you simply over-enthusiastic in these
cases or should I take it that testing is always done
independently from the JREF except in certain cases in
which the foundation has special interest (and then
only with the consent of the challenger)?

-*QUESTION*REGARDING*PROTOCOL*DESIGN*DELETED*

-*QUESTION*REGARDING*ESCROW*DELETED*

I look forward to your response and would be grateful
if you would grant me permission to post it in the
JREF forums where I have already voiced my concerns.

Regards,
Jim S., *LOCATION*DELETED*.

******************************************


From Application for status of claimant:
"NOTE: No special rules, exceptions, conditions, standards, or favors will be granted without the mutual agreement of those concerned — in advance."

[Peter Morris]

Quote:

Originally posted by Zombified
Have you ever spent more than a couple posts on anything other than trying to discredit the challenge?

Well, discussing Randi's scribblings, his tests, and general talk about skeptics is the main subject of the message board. Most of my posts have indeed been on this topic.

What would you want me to talk about?

[Peter Morris]

Quote:

Originally posted by CFLarsen
Can we see the source code for this?

Possibly, but a few potential problems.

1) I'm not sure how to post it here. As I understand it, we can't actually post content here, only links to it. That means that I'd have to set up a website, put my file on it, then provide a link to the file.

2) It was written in an old students edition of VB6. I believe that it will only run from within that particular software. The particular version I have does not permit creation of executables that would run on other computers.


3) I'm not sure if I even still have it. It was a simple program written many months ago for a single purpose. I had no real use for it afterwards.

[CFLarsen]

Quote:

Originally posted by Peter Morris
Possibly, but a few potential problems.

1) I'm not sure how to post it here. As I understand it, we can't actually post content here, only links to it. That means that I'd have to set up a website, put my file on it, then provide a link to the file.

2) It was written in an old students edition of VB6. I believe that it will only run from within that particular software. The particular version I have does not permit creation of executables that would run on other computers.

3) I'm not sure if I even still have it. It was a simple program written many months ago for a single purpose. I had no real use for it afterwards.

You can email it to me: editor@skepticreport.com

No "potential problems" whatsoever.

[Peter Morris]

Quote:

Originally posted by JimTheBrit
BS, see rule 5.

You misunderstand me. I've explained it above. But once again:

Someone makes a paranormal claim. Someone agrees to test that claim. They set up a test : Random chance would expect to score X in the test. A score of 8x is so unlikely that it would be strong evidence of paranormal powers. In fact the claimant only scores 5x, much higher than chance, much less than the target score.

Reaction of the Germans : Hey, that's interesting, he got higher than chance. Lets test him some more, and see if he keeps it up.


Reaction of Randi : You got less than the target, that's a fail. There will be no further tests done, but you can come back next year.

The point is, if Randi sets a target of 80%, then a score of 79.999999% is a failure, and proof that you're self-deluded.

Quote:

The right way to do what? The million dollars is awarded for the successful demonstration of a claim.

The right way to investigate a claim. It has nothing to do with a prize on offer.

Quote:

-Repeatedly, you state that the JREF is not involved in the testing, which is completed by a mutually
agreed independent third party.

good luck getting an answer. I'll predict he'll say something along the lines of "read the rules" and leave it at that.

[Peter Morris]

Quote:

Originally posted by CFLarsen
You can email it to me: editor@skepticreport.com

No "potential problems" whatsoever.

And what do you want it for, anyway?

[CFLarsen]

Quote:

Originally posted by Peter Morris
And what do you want it for, anyway?

So everyone can see how you got your result, of course.

[JimTheBrit]

Quote:

Originally posted by Peter Morris
Reaction of the Germans : Hey, that's interesting, he got higher than chance. Lets test him some more, and see if he keeps it up.
Reaction of Randi : You got less than the target, that's a fail. There will be no further tests done, but you can come back next year.[/i]

So the crux of your complaint against Randi seems to be the lapse of 12 months between tests. Randi isn't a competent investigator because he's not eager enough, is that what you're saying?

Quote:

good luck getting an answer. I'll predict he'll say something along the lines of "read the rules" and leave it at that.

Thanks. The reply was included in my post.

Edited for clarity.

[Peter Morris]

Quote:

Originally posted by JimTheBrit
So the crux of your complaint against Randi seems to be the lapse of 12 months between tests.

No. 12 years or 12 seconds makes no difference.

Quote:

Randi isn't a competent investigator because he's not eager enough, is that what you're saying?

No. Randi doesn't know how to conduct a test properly.


The point is, the German test one subject scored far higher than chance. So they extended the test. The additional stuff they did was PART OF THE SAME TEST.

In the same situation, Randi would not extend the test.

Quote:

Thanks. The reply was included in my post. [/b]

Ah, the trouble with top-posting. It makes conversatuions hard to follow.

[Prolix]

Originally posted by Peter Morris:
First of all I knocked up a quick computer simulation, ran it several hundred thousand times. This showed me the figure was slightly more than 100-1, maybe about 110-1.


Originally posted by CFLarsen:
Can we see the source code for this?

Peter, Your chance to avoid total loss of credibility.

Claus, I'll bet you can anticipate how Peter will try to further slither out of this.

[MESchlum]

Quoth a math nut:

Out of 50 attempts, you have 11 success. Since each attempt is binary (succeed/fail) you also have 39 failures. However, the order in which they take place does not matter, so long as you get the right number.

The exact mathematical furmula here is:

Let p be the chance of success, with n tests and k successes.

The chance is C(n,k)*p^k*(1-p)^(n-k)

C(n,k) is the number of possible combinations of exactly k successes among n tests, the rest is the chance of a specific order of successes and failures.

C(n,k) is n!/(k!*(n-k)!, where ! denotes the factorial, or the product of all integers up to that number. So 5! is 1*2*3*4*5 (=120), and 50! is BIG.

If you have Excel, or a decent pocket calculator, or know how to program, etc. you can compute this.

In the case of 50 attempts, 11 successes and a 10% chance of success, you get roughly 1/163 (to quote Excel -which is probably fairly accurate).

Approximating it by hand is more of a pain, but feasible, and gives the same size.

The answer is 1/163 for all practical purposes. Programs / rough calculations / instructions in use of software available on request, but this isn't the forum for it.

Should one try to write such a program for the merry folks of the forums? It's easy, really.

End of math rant.

[T'ai Chi]

I think he's getting it from:

Probability(of getting more 11 or more out of 50, given that the probability for each independent trial is 10%) =

SUM { nCr(n,r) * p^r * (1-p)^(n-r) }

(where nCr = n!/(r!(n-r)!)


Where n = 50, p = .1, and r runs from 11 to 50.

This yields .0093546.

[MESchlum]

Good point.

As a math note, I'd use 1-(Sum up to 10 inclusive), as it involves less work.

I get 1/69, roughly (or 0.014). The order of magnitude seems about right, anyway.

[MESchlum]

Here we go.

Just dug up an old program that computes Poisson tails, that is the odds of more than a certain number of events occuring when you know the number of expected events.

50 tests, 10% chance gives 5 tests expected, and 11 observed.

The Poisson tail is 0.01369. Not quite the exact value, but close enough for government work.

Using the same program, I can get estimates of how many tests you'd need to pass to have a one in a million chance (20), and so forth. It's small, it's dirty, it works. Please PM if interested.

[princhester]

Quote:

Originally posted by CFLarsen
Can we see the source code for this?

Hell, Claus. Get a clue. Scoring 22% on a 1 in 10 test is going to be fairly long odds. Does it matter exactly what they are? The important point has been made by JimtheBrit.

[princhester]

Quote:

Originally posted by Peter Morris
Well, discussing Randi's scribblings, his tests, and general talk about skeptics is the main subject of the message board. Most of my posts have indeed been on this topic.

Most of your posts on the Dope are also on the same topic. You're not getting off that easily.

[princhester]

Quote:

Originally posted by Peter Morris
No. 12 years or 12 seconds makes no difference.

The Australian tests were well over 12 months ago.

No one has come forward to apply saying they can get 22% on a 1 in 10 test. What conclusion do you draw, Peter?

Quote:

Originally posted by Peter Morris
Reaction of Randi : You got less than the target, that's a fail. There will be no further tests done, but you can come back next year.

A reasonable offer. Consider that testing a 22% claim requires recalculation of the proper setup (ie how many iterations are appropriate before the odds of a fluke are long enough to make it proper to award the prize).

What's wrong with saying "go away, think calmly about what we need to set up for next year then we'll have another go"

Sounds like good procedure to me.

Why is doing a hasty test for something different to what you set out to test in the first place redolent of good testing procedure?

I would have thought quite the opposite.