Raymond Edward Powell protocol

[Gr8wight]

The identity of the receiver was a sticking point in the Paul Carey negotiation for a very good reason. It is one of the most important things in removing the applicant's strongest post hoc objection: that the JREF cheated by influencing the receivers. In my opinion, the receiver must be selected by/provided by the applicant, or the test should not happen. The purpose of the test protocol is to prevent any possibilty of the applicant cheating, regardless of what tricks he may use. If the JREF provides the receiver, I guarantee that the applicant will blame his failure on the receiver. Hell, even Achau Nguyen blamed his receiver, even though he provided his own.

 

[GzuzKryzt]

The identity of the receiver was a sticking point in the Paul Carey negotiation for a very good reason. It is one of the most important things in removing the applicant's strongest post hoc objection: that the JREF cheated by influencing the receivers. In my opinion, the receiver must be selected by/provided by the applicant, or the test should not happen. The purpose of the test protocol is to prevent any possibilty of the applicant cheating, regardless of what tricks he may use. If the JREF provides the receiver, I guarantee that the applicant will blame his failure on the receiver. Hell, even Achau Nguyen blamed his receiver, even though he provided his own.

Seconded on all counts, Gr8wight.

We simply have to assume the JREF knows what they're doing.

 

[Hauteden]

If this test is not going to be recorded by video tape. How will the the die roll/card selections be verified at a later time which cannot be refuted?

As an idea could all the die rolls be performed at one time recorded and signed by all applicable parties?

 

[Capt.Nemo]

Signatures...

It's a good idea to include sign-offs at every applicable opportunity. Not only to ensure that the applicant is indeed agreeing to the current step but also to have a clear record of the tester's actions.
Case in point: Recording each roll of the die is good, but not enough. Both the roll and the chosen card should be recorded and then sig'd by both tester and applicant to ensure that both sides are in agreement. This would be done for each new attempt.
The receiving end should do the same. The receiver writes/circles/crosses the result on the score-sheet and then signs it, as does the tester.
This way, neither applicant, receiver or testers can be accused of cheating and/or mistakes. None of those sneaky fly-specs changing the appearance of the die or misheard or mumbled numbers would make it into the excuses for why it did/didn't work. Or so I hope

 

[Cuddles]

As an idea could all the die rolls be performed at one time recorded and signed by all applicable parties?

No. The whole point of the test is that the applicant doesn't know what has been rolled. If he had to sign something saying he knew the result before trying to use his psychic powers to see the result the whole exercise would be rather pointless.

 

[brettDbass]

Is it worth considering the few seconds of delay between rolling the die on the minute and it coming to rest before Mr Powell can begin his "transmission"?
Particularly a problem if a 1 is rolled and a re-roll is required. He could be 15 seconds into his minute before he can begin to "transmit", by which time the receiver may have already believed he's received the image.

I would say roll the die on the minute, begin transmitting and receiving at 15 seconds past. That should accomodate any glitches, one would hope.

 

[DavidS]

I don't understand the use of a die at all. Random number tables are available, easy to use and take far less time than throwing a die.

One good reason is that the adversaries need to trust that the numbers are "truly random", and commit to that trust. Using tables opens all manner of questions about their source, generation, application, and custody; putting those issues to bed could be tedious.

It might be far easier and quicker to get folks to agree that a die is fair, after which the generation process (throwing the die) is transparent to all.

 

[Davide]

Is it worth considering the few seconds of delay between rolling the die on the minute and it coming to rest before Mr Powell can begin his "transmission"?
Particularly a problem if a 1 is rolled and a re-roll is required. He could be 15 seconds into his minute before he can begin to "transmit", by which time the receiver may have already believed he's received the image.

I would say roll the die on the minute, begin transmitting and receiving at 15 seconds past. That should accomodate any glitches, one would hope.

An additional way to cut the risk of protocol failure is to do as Cuddles suggested -- use ten cards. Only one roll of the die for each transmission is needed instead of an indeterminate number.

 

[petre]

I've had the thought that the option for the un-carded roll should be to send nothing at all. If the receiver is aware of that possiblity, it could be interesting. "Gee, I never got any images at all so I assumed he kept rolling 1's"

 

[RemieV]

An additional way to cut the risk of protocol failure is to do as Cuddles suggested -- use ten cards. Only one roll of the die for each transmission is needed instead of an indeterminate number.

We are using single-suit, no face cards, no ace. That's nine cards. Single-suit so it's not just a partial transmission (where the number is right, but the suit is wrong - could lead to an argument over whether or not it was a 'hit'), no face cards because Mr. Powell (and we) agreed that numbers are mistakable, no ace because it is identified in the same way as face cards (with a letter and a symbol rather than a number and a symbol).

 

[RemieV]

Is it worth considering the few seconds of delay between rolling the die on the minute and it coming to rest before Mr Powell can begin his "transmission"?
Particularly a problem if a 1 is rolled and a re-roll is required. He could be 15 seconds into his minute before he can begin to "transmit", by which time the receiver may have already believed he's received the image.

I would say roll the die on the minute, begin transmitting and receiving at 15 seconds past. That should accomodate any glitches, one would hope.

We are discussing a delay. However, it would not take fifteen seconds for someone to roll a die twice. I can't imagine it taking more than five seconds.

The receiver would be told that the later within the minute they write their response, the better. So even if it somehow did take fifteen seconds for Mr. Powell to start transmitting, that still leaves forty five more for the receiver to write down the card.

Remember that according to Mr. Powell, the transmission will be so strong and overcoming that everyone in the building will be able to see it. It is also immediate.

 

[sad_skinhead]

What about drawing lots, like they do for the football?

 

[SezMe]

One good reason is that the adversaries need to trust that the numbers are "truly random", and commit to that trust. Using tables opens all manner of questions about their source, generation, application, and custody; putting those issues to bed could be tedious.

I don't think this is a good reason. Consider the second sentence in Wiki's entry on random number tables:

Random number tables have been used in statistics for tasks such as selected random samples. This was much more effective than manually selecting the random samples (with dice, cards, etc.).

As for a trusted source, get the applicant and JREF to simply use this table.

Finally, using such a table eliminates the time delay issue completely.

 

[ben m]

There are various reasons you might use a random number generator. If you're doing a 1-D Monte Carlo integral you care about "uniformity" but don't give a hoot about long-range correlations. If you're seeding cryptographic keys you care only weakly about uniformity, but you want a long period.

What's the point of the random numbers for Mr. Powell? We care that his receiver can't exploit a flaw in the number-generators to "guess" what number Mr. Powell has been shown. For example: if the receiver knew that the numbers were being picked "randomly" by a human, he'd do well by guessing 7 and 3 often, 1 and 5 rarely. If the reciever a) knew that the numbers would be generated by a certain implementation of /dev/random, b) knew that the random-number seed would be the time of day in seconds, c) could guess within 1 minute what time the seed was drawn, and d) had a confederate or concealed computer feed him the 60 possible sequences, he'd have a 1 in 60 chance of winning.

Seriously, though, none of the random-vs-pseudorandom arguments above are at all relevant to this aspect of randomness. Since the guesser gets no feedback, we don't need to worry about correlations. Since we're only looking for 1e-6 odds, we don't care much about period. To avoid the guess-the-seed problem, at worst, we need only make sure that there are more than 1e6 bits of uncertainty in the seed---not good enough for cryptography by a long shot, but fine for us and easily achieved (for example: seed the RNG with a ten-digit hex number pulled from dice rolls.) We care about uniformity at the level of a few percent---as long as the generator doesn't land on one number twice as often as another, we're fine.

In other words: dice are fine. Cards are fine. Pseudorandom numbers are fine. Tables are fine, again, as long as the receiver can't guess (with 1e6 odds) where we're going to start in the table.

 

[Hauteden]

No. The whole point of the test is that the applicant doesn't know what has been rolled. If he had to sign something saying he knew the result before trying to use his psychic powers to see the result the whole exercise would be rather pointless.

Oh I guess I misunderstood. I thought the applicant was transmitting the information to some one else.

Hauteden

 

[Cuddles]

Oh I guess I misunderstood. I thought the applicant was transmitting the information to some one else.

Hauteden

My mistake. I must have got him confused with someone else. However, I would expect applicants to have to sign something after the test to confirm the outcome in which case signing every pick as well wouldn't be necessary. Could a JREF person comment on this?

 

[RemieV]

My mistake. I must have got him confused with someone else. However, I would expect applicants to have to sign something after the test to confirm the outcome in which case signing every pick as well wouldn't be necessary. Could a JREF person comment on this?

The test will more than likely be videotaped, rendering that unnecessary.

 

[Blippiwewe]

Do we have another test forthcoming?
Finally, a claimant says he can do 100% accuracy. No ands, ifs or butts.

The only thing striking me is of course that Mr. Powell does not bring his own "receiver". That was the breaking point e.g. in the Paul Carey test, since it would have provided him an "out". How is this different now?

However, if Mr. Powell has agreed to this, it should satisfy the need to meet the claimant's requirements, shouldn't it?

(Again keeping in our very cynical minds that Mr. Powell simply wants to take the test, fail and say "I would have succeeded but they wouldn't let me bring my own receiver." or any other excuse.)

I don't understand the need to accept the subject's 100% accuracy as a requirement. A result that is greater than chance is scientifically significant. Requiring anything more, even if the subject insists he can do it, only undercuts the credibility of JREF by seeming to require more than is necessary. It is a preliminary test and as such should require only that the subject do something more than chance would predict.

 

[steenkh]

It is a preliminary test and as such should require only that the subject do something more than chance would predict.

No. The MDC is a challenge to the claimants to do what they claim they can do. It is a courtesy that they offer him to perform at less than what he claims to do, but if he insists, that should be what he is tested for: that which he claims.

 

[Blippiwewe]

I disagree. The challenge, as I read it, is to produce evidence of paranormal or supernatural ability or phenomena. If you claim to be able to make a truck levitate but only manage to make a VW bug levitate would that satisfy the challenge? Not with your logic. Would it still be evidence of such phenomena? I'd say so.
My point is that requiring more than is scientifically necessary undermines the credibility of the organization and gives claimants an arguement they shouldn't be given.