Ganzfeld million dollar challange?

[Uncayimmy]

It boils down to this - and this has been said dozens of times before in this subforum: Doing what you claim to be able to do requires no odds.

It.
Requires.
You.
To.
Do.
What.
You.
Claim.
To.
Be.
Able.
To.
Do.

I disagree that it's that simple.
Claim: I can predict the outcome of a roll of a die.
Valid? No. Anybody can do that. Most only get it right one of out six times.

Claim: I can predict the outcome of a roll of a die with 90% accuracy.
Valid: Sounds paranormal to me. But how do we define 90%? 10 trials? 25? 100?

Now, you might argue that if I can predict something, why can't I do it 100% of the time. Well, GzuzKryzt, I can't even chew my food without biting my tongue or swallowing "wrong" every now and then. I have been reading for the majority of my 42 years, but just today I put in the "Bedtime Baby Signs" DVD when I intended to put in "Bath Time Baby Signs." I've been playing electric bass for 25 years, but not a gig goes by where I don't have a flub in some song I've played a few hundred times.

There are many ordinary skills that are not 100% reliable. Even Rick Berry missed free throws on a regular basis. So, how can any of the skills I mentioned be validated without some consideration of the odds?

 

[Uncayimmy]

Would it be possible to at least clarify if the preliminary test odds of say one in a thousand are carried over to the main test or does the main test start from scratch?

From the Challenge FAQ:

The protocol itself will not be changed, and neither will any of the documents you and the JREF have agreed upon. The final test may be longer, or require more conclusive results through more sets of the test to ensure that the preliminary test was not a fluke.

 

[Uncayimmy]

Why would the second time be more productive than the first?

Funny, you don't seem to have any trouble asking the same questions here more than once. Letters get lost in the mail. Sometimes they get lost in the office. Sometimes people just forget.

And why can't you or Unca Yimmy answer my very simple question about a Ganzfeld protocol?

Do you understand that:

• The applicant must first make a paranormal claim? Yes/No
  
• Protocols are negotiated between the JREF and the applicant? Yes/No
  
• You have not stated a claim? Yes/No
  
• Without a claim, there is no protocol? Yes/No
  
• You referred to "a" Ganzfeld protocol which correctly implies that there is no single specific protocol for anyone to even examine? Yes/No
  
• The JREF is under no obligation to respond to a vague inquiry that does not include a specific claim or a specific protocol? Yes/No
  
• I don't represent the JREF? Yes/No

Do you understand why no definitive answer can be forthcoming?

I have a general answer for you. The basic concept of isolating a couple of people and asking the two of them to best judge which out of four images was being "transmitted" is not irreparably flawed, but it's certainly a pain in the ass. The Zener cards are much more practical.

Taking 30 minutes per trial is impractical for a 1:4 guess. It would require an enormous amount of time to arrive at a statistically significant number. Now, if the guess were 1:1000, it might be doable in an afternoon.

Furthermore, the reported best results were getting it right 35% of the time instead of the expected 25%. That small of a difference could easily be explained by a flaw in my methodology/protocol. My ego is not so big that I think I could design any test where I would risk a million dollars that I didn't make some tiny mistake that gave a slight edge to the applicant.

This, I think, is at the heart of your question. It's also why each protocol and accuracy conclusion must be negotiated individually. The more variables in the experiment, the greater the risk of missing something, which in turn leads to a requirements for accuracy that must be safely beyond the boundaries where experimental error could be an explanation.

If you want to discuss Ganzfeld further, take it up in the General Skepticism forum. I'll be glad to engage you there. I think I have pretty much covered in a general, non-official sense why it's probably not a good candidate for a challenge. Of course, Honorton could certainly start an application himself. That would answer the question definitively.

 

[GzuzKryzt]

...
Claim: I can predict the outcome of a roll of a die with 90% accuracy.
Valid: Sounds paranormal to me. But how do we define 90%? 10 trials? 25? 100?
...

Now we are getting somewhere, since you made a specific claim.

Either 10 or googol rolls of a die: 90% remains 90%, right?

We would set a sample size of rolls big enough to eliminate a false positive and reduce the possibility of pure chance to an acceptable minimum - how about 20 rolls and the need for 18+ to be correct? - and then we'd have a basic outline of a protocol.

 

[fls]

Of course, Honorton could certainly start an application himself. That would answer the question definitively.

I think that if Honorton could start an application himself, that would certainly satisfy the Million Dollar Challenge.
He's dead.

Linda

 

[Uncayimmy]

I think that if Honorton could start an application himself, that would certainly satisfy the Million Dollar Challenge.
He's dead.

Linda

How would that work? I should send the JREF a letter asking them how the dead can apply for the MDC. Seems to me that just the act of applying should be enough to win.

 

[rjh01]

I can imagine the correspondence.
Applicant - I wish to apply for the MDC. Claim I am dead. Here is my death certificate. Here is evidence that I am the person named on the death certificate.
JREF - To what accuracy do you claim to be dead? How we do test for you still being dead?

 

[Uncayimmy]

Now we are getting somewhere, since you made a specific claim.

Either 10 or googol rolls of a die: 90% remains 90%, right?

Since I have no idea what you mean by "getting somewhere" I have to answer no, 90% does not remain 90%. It fluctuates. Take 10 rolls. There's no guarantee that I am going to miss one, thus my 90% accuracy may appear as 100%. Likewise, I may miss two in the first 10, making my accuracy 80% after 10.

Along those same lines, a 90% accuracy rate will be represented by different values depending upon when you start counting and when you take a peek. Suppose out of 100 trials I missed the first five and last five. Do the math as to what my 90% accuracy rate would look like after each roll.

We would set a sample size of rolls big enough to eliminate a false positive and reduce the possibility of pure chance to an acceptable minimum - how about 20 rolls and the need for 18+ to be correct? - and then we'd have a basic outline of a protocol.

I don't want to play semantics, but that's not really a protocol. The protocol is all the steps involved in rolling the die, getting my answer, and seeing if I'm right. It's something else entirely to determine what level of accuracy is considered success and what level of *confidence* is required in that value.

The MDC FAQ says that more trials may be required in the final test if someone passes the preliminary test.

If you told me where you were heading, perhaps I could give better answers. Right now I have no idea what you're driving at.

 

[Rodney]

If you want to discuss Ganzfeld further, take it up in the General Skepticism forum. I'll be glad to engage you there. I think I have pretty much covered in a general, non-official sense why it's probably not a good candidate for a challenge.

So why doesn't the JREF simply rule out Ganzfeld challenges?

Of course, Honorton could certainly start an application himself. That would answer the question definitively.

I can't disagree there.

 

[Uncayimmy]

So why doesn't the JREF simply rule out Ganzfeld challenges?

I'm not the JREF. Are you asking me to speculate? If so, I'd say it's for the same reasons that specifics about any type of challenge are not given in advance: There's no reason to.

Or maybe it's because Ganzfeld is not worthy of special treatment considering how poorly designed it was.

 

[GzuzKryzt]

Since I have no idea what you mean by "getting somewhere" I have to answer no, 90% does not remain 90%. It fluctuates. ...

Ah, fluc-tu-ates.

You said 90% accuracy. Now is it 90% accuracy or not?

 

[Uncayimmy]

Ah, fluc-tu-ates.

You said 90% accuracy. Now is it 90% accuracy or not?

You tell me. Here's a hypothetical run of 20 trials with a running percentage:

Trial Correct Accuracy
1 1 100%
2 0 50.0%
3 0 33.3%
4 1 50.0%
5 1 60.0%
6 1 66.7%
7 1 71.4%
8 1 75.0%
9 1 77.8%
10 1 80.0%
11 1 81.8%
12 1 83.3%
13 1 84.6%
14 1 85.7%
15 1 86.7%
16 1 87.5%
17 1 88.2%
18 1 88.9%
19 1 89.5%
20 1 90.0%

It looks like to me that my 90% accuracy rate was not revealed until the 20th trial. Thus, my statement that a 90% accuracy rate will exhibit fluctuations depending on when you look is accurate. Given enough trials (1,000, for example) it will be very close to 90% and remain there. Take any subset of the 1,000 trials and you will very likely see a number other than 90%. The smaller the subset, the greater the chances of seeing a number other than 90%.

I am not about to give you a lesson in statistics. I don't even know if you need one because I have no idea what point you're trying make other than trying to refute mine. My point is simple: There are many claims, paranormal and ordinary, that require statistical analysis.

If you have a point, please state it.

 

[Cuddles]

Then please enlighten me, at least as to the second issue: Are all protocols, including time-consuming ones such as Ganzfeld experiments, eligible for the Challenge?

As already pointed out, the rules and FAQ are very clear on this - no. Several kinds of claim are refused out of hand. Anything potentially causing harm to anyone, for example. In addition, certain claims like cloud-busting will not even be considered (although some that are now excluded may have been looked at in the past).

I really don't understand why you keep asking this as if it's a difficult question that people are deliberately avoiding. No, not all claims will be considered elligible for the challenge. The only way to find out is either to refer to the rules and FAQ or, if your claim is not included there, apply for the challenge and find out.

However, as I've noted before, it seems very unlikely that the JREF would accept a Ganzfeld experiment for the MDC. The main reason, other than the length of time, is that there's a big difference between a challenge and doing research. Take homeopathy, for example. You could claim to be able to tell the difference between two apparently identical solutions or to demonstrate in a randomised controlled trial of thousands of patients that one works better than the other at curing cancer. The JREF will be interested in the former, but not the latter. It's the same for any paranormal claims. Claim to be able to read minds, lift objects with your mind or whatever and they'll be happy to test you. Claim to be able to conduct a study over several years that may show some anomalous results at the end and they really won't care, at least as far as the challenge is concerned.

Of course, as always only the JREF can give an absolute answer. However, since they're not in the habit of giving absolute answers to hypotheticals, I wouldn't worry about it until someone actually tries applying.

 

[Rodney]

Of course, as always only the JREF can give an absolute answer. However, since they're not in the habit of giving absolute answers to hypotheticals, I wouldn't worry about it until someone actually tries applying.

Bear in mind that this thread was initiated by andy2001 asking: "Could Ganzfeld be used for the Randi challenge?" To which GzuzKryzt responded that, if andy2001 want a "definite and official answer", he could ask challenge@randi.org.

You now seem to be conceding that, if andy2001 does ask the JREF his question, he is not going to receive a response, any more than I did.

 

[GzuzKryzt]

You tell me.
...

Is this a Monty Python skit?



I made my point, UncaYimmy. The applicant has to state what he can do.

Since the OP is only a hypothetical scenario, I couldn't care less. Perhaps it would help you if you browsed the Challenge Application subforum. There are many similar claims and hence many similar JREF proposals.

 

[Bindamel]

I disagree that it's that simple.
Claim: I can predict the outcome of a roll of a die.
Valid? No. Anybody can do that. Most only get it right one of out six times.

Claim: I can predict the outcome of a roll of a die with 90% accuracy.
Valid: Sounds paranormal to me. But how do we define 90%? 10 trials? 25? 100?

You tell me. Here's a hypothetical run of 20 trials with a running percentage:

Trial Correct Accuracy
1 1 100%
2 0 50.0%
3 0 33.3%
4 1 50.0%
5 1 60.0%
6 1 66.7%
7 1 71.4%
8 1 75.0%
9 1 77.8%
10 1 80.0%
11 1 81.8%
12 1 83.3%
13 1 84.6%
14 1 85.7%
15 1 86.7%
16 1 87.5%
17 1 88.2%
18 1 88.9%
19 1 89.5%
20 1 90.0%

It looks like to me that my 90% accuracy rate was not revealed until the 20th trial. Thus, my statement that a 90% accuracy rate will exhibit fluctuations depending on when you look is accurate. Given enough trials (1,000, for example) it will be very close to 90% and remain there. Take any subset of the 1,000 trials and you will very likely see a number other than 90%. The smaller the subset, the greater the chances of seeing a number other than 90%.

Now this is probably more a result of a poor choice of example than anything, but I thought I'd clarify this.

My understanding is that the JREF's intent in setting successes/trials is to come up with a protocol where the likelihood of succeeding by chance in the preliminary trial is about 1/1000. For dice rolls, which carry a 1/6 individual probability, 1/1000 is 10 out of 20 trials, not 18.

A challenge applicant should be concerned with reducing the likelihood that he will fail purely by chance, given his ability. Increasing the sample size, while maintaing the 1/1000 chance success rate, decreases the likelihood that the claimant will fail by chance. In this particular case, if the claim is 90%, the odds of him failing 11 times out of 20 is less than one in a million.

At this point, we fall back into an argument that's been made here before, and seems to be at the root of all this discussion: if you're paranormal ability varies only slightly from chance (or requires a complex set of conditions), should JREF be required to spend their time and effort to test you? Given that it's their challenge, no (I do not, of course, presume to speak for them).


At that point, you should instead be asking yourself, "How can I profit from my ability?"

 

[Uncayimmy]

Now this is probably more a result of a poor choice of example than anything, but I thought I'd clarify this.

It wasn't a poor choice example at all. I chose it deliberately to demonstrate that performing "above chance" doesn't have to factor into it at all. Many claims are not 100%. Those that are not 100% will require some statistical analysis to determine a number of trials that will give sufficient confidence that someone is performing at their claimed percentage level.

Thus, to determine that I am performing at a 90% success rate one does not have to consider a 1:6 chance of getting it right by chance. One need only run sufficient trials to confirm 90% is correct. What are the odds of hitting a baseball going 85mph? I dunno. But if I say I can do it 40% of the time, we'll need to crunch some numbers to figure out how to validate that claim.

Where chance comes into play is determining whether a claim is paranormal or not. You'd better be able to kick chance's butt if you want it to be considered paranormal.

At this point, we fall back into an argument that's been made here before, and seems to be at the root of all this discussion: if you're paranormal ability varies only slightly from chance (or requires a complex set of conditions), should JREF be required to spend their time and effort to test you? Given that it's their challenge, no (I do not, of course, presume to speak for them).

I agree. I said as much already about Ganzfeld. I wouldn't put my money up in a bet where a minor mistake in protocol would be enough to make me lose.

At that point, you should instead be asking yourself, "How can I profit from my ability?"

Amen!

 

[Rasmus]

It wasn't a poor choice example at all. I chose it deliberately to demonstrate that performing "above chance" doesn't have to factor into it at all. Many claims are not 100%. Those that are not 100% will require some statistical analysis to determine a number of trials that will give sufficient confidence that someone is performing at their claimed percentage level.

Thus, to determine that I am performing at a 90% success rate one does not have to consider a 1:6 chance of getting it right by chance. One need only run sufficient trials to confirm 90% is correct. What are the odds of hitting a baseball going 85mph? I dunno. But if I say I can do it 40% of the time, we'll need to crunch some numbers to figure out how to validate that claim.

Where chance comes into play is determining whether a claim is paranormal or not. You'd better be able to kick chance's butt if you want it to be considered paranormal

I agree. But none of that is a problem of the JREF. The applicant has to know what it is they can do and make an according claim.

For that, they ought to know with what kind of reliability or confidence they can expect from themselves.

But that is something at least slightly different from the JREF announcing how big the likelihood is that someone without the claimed ability could pass the test by chance.

There is no need for an honest applicant to care let alone worry about that!

If the ability is less than 100% reliable - and I would always expect that - it would still be trivial for someone to pass the test that has the ability.

I can read.

This can easily be tested. It would be close to trivial to define a protocol for it. Part of the protocol would be that there have to be so-and-so many different words, randomly selected from a dictionary, to ensure that I am really reading rather than just guessing words.

If I can read reasonably well and am worried about errors I can easily have a test that is both lenient enough to allow for the odd mistake of the testee and strict enough to make it nearly impossible to pass by chance alone.

The english language has several hundred thousand words - so my chances of guessing just one of them are pretty small to begin with. (Not accounting for clues like lenght, or a partial ability to read, etc.) So it would be near enough impossibly to correctly guess 5 or 10 words, especially if they are the same lenght and contain the same vowels, say.

If I was now worried about mistakes: No problem, we can just double the amount of words and allow me a single mistake or some such thing - whichever makes me feel confident.

It's not going ot be a problem, because I can read.

 

[Uncayimmy]

It's not going ot be a problem, because I can read.

Right.

But if somebody is willing to put up a million bucks because he doesn't believe I can read when I know I can, then I am going to make triple-sure he's not trying to pull a fast one. I'm sure at least some of the sincerely mistaken and delusional are bright enough to at lest be wary of a trap.

This quote by you sums it up quite well: "If I can read reasonably well and am worried about errors I can easily have a test that is both lenient enough to allow for the odd mistake of the testee and strict enough to make it nearly impossible to pass by chance alone."

This seems to be at odds (pardon the pun) with the statement (not by you) to which I was responding: "Doing what you claim to be able to do requires no odds." Then again, I am reading "doing" as "proving" or "doing it to prove to others" rather than just "doing it in your basement."

For the record, I have no problems with the challenge as it is written. It's fair. I have no belief whatsoever that anybody is going to win the challenge because I don't believe the paranormal is real.

 

[GzuzKryzt]

Now this is probably more a result of a poor choice of example than anything, but I thought I'd clarify this.

My understanding is that the JREF's intent in setting successes/trials is to come up with a protocol where the likelihood of succeeding by chance in the preliminary trial is about 1/1000. For dice rolls, which carry a 1/6 individual probability, 1/1000 is 10 out of 20 trials, not 18.

A challenge applicant should be concerned with reducing the likelihood that he will fail purely by chance, given his ability. Increasing the sample size, while maintaing the 1/1000 chance success rate, decreases the likelihood that the claimant will fail by chance. In this particular case, if the claim is 90%, the odds of him failing 11 times out of 20 is less than one in a million.

At this point, we fall back into an argument that's been made here before, and seems to be at the root of all this discussion: if you're paranormal ability varies only slightly from chance (or requires a complex set of conditions), should JREF be required to spend their time and effort to test you? Given that it's their challenge, no (I do not, of course, presume to speak for them).


At that point, you should instead be asking yourself, "How can I profit from my ability?"

Thanks, Bindamel.



And again we see: It all depends on the claim.