• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Challenge applications

Pixel42 said:
Suppose you test dozens of people in this way and, no matter how convinced they are they that have this ability and that they will score 100% correct, none ever manage to get more than 28 right. What would you conclude?
Click to expand...

(A)
Your example is great on the methodology. You have provided an easier way to make my points.

I would ask, what are the odds of getting 28 hits out of 100 trials?

And what are the number of hits needed to get odds exceeding 1,000 per test?
And how do you calculate such odds given the hits are not 100%?

(B)
Go to http://www.internationalskeptics.com/forums/showthread.php?t=87775
(JREF Forum » General Topics » Education = My Zener card experiment)

The formula is BDF = C(N,M) * p^M * (1-p)^(N-M)

This is the scientifically accepted formula. Why not state so in the rules?

At http://saliu.com/theory-of-probability.html you will find that

C(N.M) = N! / (M! * (N-M)!)

((* is multiply, / is divide, ^ is power and ! is factorial.))
(( N is trials, M is hits, p is individual probability))
((Odds are the reciprocal of BDF.))

This is the same as the Excel BINOM.DIST function.
The Excel sheet is unreliable for doing 100 trials!!
It cannot compute the large numbers - why not put a warning in the rules?

(C)
The odds for 28 hits with 100 trials is 14 to 1 (a real failure)
The odds for 38 hits is 743, and for 39 hits it is 1,402.

Why not give your example, give the formula and give the required number of hits (39) for the witnessed test, the preliminary test, and for the final test?

I have done my own checks of my work, and I think I am fairly free of error.

Statistics is far from simple. I last did a course when I studied for a degree in business in 2007, so I had to search many websites to find the full and correct accepted formula.

(D)
What are the odds if the two tests are combined? 39+39 out of 100+100?

It is not 1,402 * 1,402 (= 1,965,604), but 226,395 using the formula.
To exceed 1 Million odds combined then one must get 41 for each test (or 40+41) (or 39+42).

Again I ask, are two tests of 1,402 okay, or is an overall combined result exceeding 1 million required?
 
Last edited:
Pixel42 said:
To win the MDC requires the applicant to pass just two tests: the preliminary and the final.
Click to expand...

Not really. To win it, you must be in it.

An application will not be considered without some supporting documentation and I cannot see that a letter from my clergyman saying I am a nice and gifted guy will do it. At least one successful test is required to "buy the ticket".

And I maintain the if Don wins, JREF will say he beat the test odds of 1 million, but Don CAN claim that he beat odds of 2 trillion by using my logic.
 
Last edited:
Sometimes my brain just keeping working on problems even when I am trying to sleep. So I got up.

I made an assumption that the self-test, the application test, and the preliminary test were the same protocol as the Final Test.

Bad assumption.

The rules say that the preliminary could uncover issues that would require re-negotiation of test protocol. This would mean that only the Final Test would count. And then all other tests are invalid as far as odds are concerned.

I presume that the Final Test must then beat odds of 1 million on one test. If so, a 100% success on 4 trials is not sufficient. It would require 100% success on 8 trials (in effect another preliminary added in).
 
PartSkeptic said:
Not really. To win it, you must be in it.

An application will not be considered without some supporting documentation and I cannot see that a letter from my clergyman saying I am a nice and gifted guy will do it. At least one successful test is required to "buy the ticket".

And I maintain the if Don wins, JREF will say he beat the test odds of 1 million, but Don CAN claim that he beat odds of 2 trillion by using my logic.
Click to expand...
It would only be a problem if he claimed he could do something with a probability close to chance, in which case he should not apply.

If the JREF was putting up a prize for people who could ride a bicycle, the rules might state that the applicant would have to ride 100 metres of level ground without touching the ground with his body at any point, while staying within the confines of a bicycle lane 1,5 metres wide. For an applicant who cannot ride a bicycle this would perhaps be a trillion to one that he could make it, but for a person who really can ride a bicycle, it should pose no problem at all.

Besides, there are no rules stating exactly what odds will be demanded by the JREF. It has actually happened that the JREF has lowered the odds for an applicant in order to make the test cheaper and shorter.
 
PartSkeptic said:
The rules say that the preliminary could uncover issues that would require re-negotiation of test protocol. This would mean that only the Final Test would count. And then all other tests are invalid as far as odds are concerned.
Click to expand...
That is possible, but it is difficult to say, because nobody has passed the preliminary test. Probably even the JREF will be unsure about what to do. But it has often been said informally by JREF staffers that the odds and the protocol would stay the same for the final test.

Presumably, the protocol would only be renegotiated if it became obvious that the applicant had found a way to circumvent the protocol, or if the rules turned out to be ambiguous.
 
PartSkeptic said:
Not really. To win it, you must be in it.

An application will not be considered without some supporting documentation and I cannot see that a letter from my clergyman saying I am a nice and gifted guy will do it. At least one successful test is required to "buy the ticket".

And I maintain the if Don wins, JREF will say he beat the test odds of 1 million, but Don CAN claim that he beat odds of 2 trillion by using my logic.
Click to expand...

And all this is irrelevant if he can do what he claims to be able to do. For any applicant who truly has the ability they claim to have, any discussion of odds is unnecessary, because they are not operating under random chance.
 
jsfisher said:
Minor nit. Odds = (1/probability) - 1.

E.g. A 25% probability translates to 3-to-1, not 4.
Click to expand...

Minor in some respect, but when I comes to math/stats definitions, there is the need to be precise. Thank you.
 
Gr8wight said:
And all this is irrelevant if he can do what he claims to be able to do. For any applicant who truly has the ability they claim to have, any discussion of odds is unnecessary, because they are not operating under random chance.
Click to expand...

My claim would be the that I can do the test proposed by Pixel42, and get enough right so that the majority of scientists would say the result is "significant enough to rule out chance".

Why would I not claim to be able to do 100%? Because if mental telepathy exists it is a weak ability, and the conscious mind starts guessing. A "receiver" under hypnosis might do 85 out of a 100, or even 100 out of a 100, but to find such a person is very hard. They are rare and probably young (about 10years old).

To get 85 out of 100 is so good it will impress most people. But what are the scientific odds. So great there are about 35 zeros, not the 6 zeros of a million to one.

Note: 80% right in 10 tests is not impressive. the odds are 2588 to 1.
The number of trials makes a huge difference.

JREF acknowledge this and give odds calculated by a formula. All I am saying is that JREF could be a little more open about odds and formula (some tests use different formulae, but there are probably only two or three)
 
steenkh said:
That is possible, but it is difficult to say, because nobody has passed the preliminary test. Probably even the JREF will be unsure about what to do. But it has often been said informally by JREF staffers that the odds and the protocol would stay the same for the final test.
Click to expand...

DowserDon needs 100% for all tests to beat the odds if he uses 4 trials per test. My initial point was that if he does 8 trials per test, he might need only 6 out of out 8 to give him a margin of error. In fact, JREF might insist on this protocol for the final.

DD's number of trials per test are low and the odds exceed 1,000 to 1. My other point was that if he can repeat these test over and over, it is hard to argue with his success rate, because then the odds are huge.
 
PartSkeptic said:
DowserDon needs 100% for all tests to beat the odds if he uses 4 trials per test.
Click to expand...
Which is why he should change his protocol so that every dowsing spot has a 50% chance of having a trench under it. Then he would need fewer walkways/spots than his existing protocol requires, and still be able to beat the required chance odds with a success rate a lot less than 100%.

PartSkeptic you still seem to be missing the point about the purpose of self testing and the setting of success criteria relative to chance success. Yes of course the odds against you getting a positive result if you don't have a paranormal ability increase the more testing you do, that's the whole point of establishing whether a result is repeatable: to eliminate the possibility of a false positive due to a statistical fluke. But someone who really can do what they claim to be able to do will get a better-than-chance result in every test, so the more testing they do the more obvious and undeniable their ability becomes. Someone who really does have a paranormal ability will find that easier, not harder, to prove with every test they do, as the odds against chance success accumulate in their favour.

As for why JREF don't go into success criteria and probability formulae in their rules, I can't speak for them but I would imagine it's partly because every claim - and therefore every agreed test protocol - is different so they need to be flexible (as you yourself pointed out with some claims this statistical approach can't even be applied), but mostly because such details would be totally incomprehensible to 99% of applicants. For them just saying something like "You need to find at least 7 of the 10 barrels of water buried in this piece of ground to pass the test" is perfectly understandable and reasonable.
 
Last edited:
I, for one, would prefer that PartSkeptic start a new thread. This is drifting from DowserDon's claim and moving closer to a Zener Card claim from PartSkeptic. I understand that PartSkeptic is trying to figure out how JREF develops protocols and figures odds by using DowserDon as an example, but I think it would be more helpful in a new thread.

One poster's opinion,

Ward
 
PartSkeptic said:
My claim would be the that I can do the test proposed by Pixel42, and get enough right so that the majority of scientists would say the result is "significant enough to rule out chance".
...
Click to expand...

Excellent. Now all you have to do is do it.

PartSkeptic said:
...
JREF acknowledge this and give odds calculated by a formula. All I am saying is that JREF could be a little more open about odds and formula (some tests use different formulae, but there are probably only two or three)
Click to expand...

All I am saying is that people like you could be a little more open about doing what they claim to be able to do.

(To the tune of "Give peace a chance".)
"All we are saying
Is do what you claim."
 
Upon re-reading my post, I may have been a little overzealous or impolite. What I meant to convey was, PartSkeptic, that I am greatly looking forward to results - any results - from an actual test or self-test. For the last years, please allow the brevity, we have seen a lot of this:

1. Forum member claims ability.

2. When asked to produce data to confirm ability, forum member produces words.

3. When told words do not equal satisfactory data, forum member produces more words.



Conclusion: Forum member has not produced data, but many, many, many words. Exceptions: Edge and connie sonne. That did not, and in edge's case does not, prevent them from producing many, many, many more words.

Since you're retired, take an afternoon and read both cases. Have chocolate or other helpers ready.
 
PartSkeptic said:
DowserDon needs 100% for all tests to beat the odds if he uses 4 trials per test. My initial point was that if he does 8 trials per test, he might need only 6 out of out 8 to give him a margin of error. In fact, JREF might insist on this protocol for the final.

DD's number of trials per test are low and the odds exceed 1,000 to 1. My other point was that if he can repeat these test over and over, it is hard to argue with his success rate, because then the odds are huge.
Click to expand...
Rather than get all caught up with the maths, I tend to refer to these handy tables that have been devised by the author (not me) with the MDC in mind.

http://www.automeasure.com/chance.html

They outline the number of trials and expected "success" due to random chance for "odds" of 1:100, 10,000 and 1,000,000.
 
PartSkeptic said:
Why would I not claim to be able to do 100%? Because if mental telepathy exists it is a weak ability, and the conscious mind starts guessing. A "receiver" under hypnosis might do 85 out of a 100, or even 100 out of a 100, but to find such a person is very hard. They are rare and probably young (about 10years old).
Click to expand...
Why do you not simply test yourself to find out exactly what odds you get right?
 
wardenclyffe said:
I, for one, would prefer that PartSkeptic start a new thread. This is drifting from DowserDon's claim and moving closer to a Zener Card claim from PartSkeptic. I understand that PartSkeptic is trying to figure out how JREF develops protocols and figures odds by using DowserDon as an example, but I think it would be more helpful in a new thread.

One poster's opinion,

Ward
Click to expand...

You mean two poster's opinons. :)
 
Thanks for all the comments, including the "lighter" ones.

I take the points made, and will now back down, and try not to sidetrack from DD.

When my son gets round to doing an experiment with me, I will give a short summary.
 
EHocking said:
Rather than get all caught up with the maths, I tend to refer to these handy tables that have been devised by the author (not me) with the MDC in mind.

http://www.automeasure.com/chance.html

They outline the number of trials and expected "success" due to random chance for "odds" of 1:100, 10,000 and 1,000,000.
Click to expand...

These are really good. Thanks.

I like to understand the underlying theory. Then I more fully understand the tables.
 
You must log in or register to reply here.

Back
Top Bottom