Merged New telepathy test: which number did I write ?

[ComfySlippers]

As I mentioned earlier, I rather ignored this thread early on. I was just browsing all the way back on page 4 where the actual answers are discussed and I was taken aback to see that, in Michel's view, the protocol was that people should PM a secret vote in a short sentence along the lines of "My selected number is 1" and then openly post an MD5 hash of their PM plus a version of the same sentence with only the number changed and replaced with "XX".

I'm sure this must have been pointed out already, so sorry to re-hash it but surely that makes secret voting pointless. If I were obliged to use the identical form of words then I must have written one of;

"My selected number is 1"
"My selected number is 2"
"My selected number is 3"
"My selected number is 4"

It would only take a matter of moments to see what the MD5 hash is for those sentences and compare it to the posted one.

If you can make any sense out of this nonsense you win the prize.

 

[Dan O.]

Participant|PM to Agatha|blinded|Post MD5 hash|CR rating|Answer|Correct|CR * Correct
Hokulele|Y|Y|N|0|1|-1|0
Ladewig|Y|Y|N|-5|2|+1|-5
stanfr|Y|Y|Y|-5|1|-1|+5
Kid Eager|Y|Y|N|-5|4|-1|+5
fagin|N|Y|N|-4|?|0|--
femke|Y|Y|N|0|2|+1|0
gabeygoat|Y|Y|Y|0|4|-1|0
NaySayer|Y|N|N|?|?|0|--

Out of all the answers, he got one right. The other correct answer was ruled invalid by the credibility system. What does that tell you? Tells me he got lucky, and if one can't tell the difference between telepathy and chance, the test fails.

Your analysis fails to consider that the CRs for those three answers are negative so the correct analysis will show that Michel got 2 right, 1 wrong and 3 skipped.

 

[Nay_Sayer]

Out of all the answers, he got one right. The other correct answer was ruled invalid by the credibility system. What does that tell you? Tells me he got lucky, and if one can't tell the difference between telepathy and chance, the test fails.

Tells you michel is psychic, duh.


Dan won't recognize my better psychicness

 

[JimOfAllTrades]

I argue that a CR of 0 represents an answer has no credibility and is effectively excluded from the result, a positive CR is an assessment that the answer represents the truth and a negative CR is an assessment that the answer represents a falsehood.

I don’t recall Michel saying this. Is this actually his view, or just yours? If it’s just yours, then I’m afraid the only thing that comes to mind for me is “data mining after the fact”.

Even if it is his view, how do you get a “correct” answer for a wrong answer with a negative credibility score? It seems to me that at best you could say he correctly identified two people who replied with a falsehood (assuming your interpretation of Michel’s CR system is correct), and did not report the number they actually thought of. But this says nothing about what number they actually did think of, it could have been any of the other four possible answers. It’s not shown at all that they “heard” the correct number.

 

[kid meatball]

Participant|PM to Agatha|blinded|Post MD5 hash|CR rating|Answer|Correct|CR * Correct
Hokulele|Y|Y|N|0|1|-1|0
Ladewig|Y|Y|N|-5|2|+1|-5
stanfr|Y|Y|Y|-5|1|-1|+5
Kid Eager|Y|Y|N|-5|4|-1|+5
fagin|N|Y|N|-4|?|0|--
femke|Y|Y|N|0|2|+1|0
gabeygoat|Y|Y|Y|0|4|-1|0
NaySayer|Y|N|N|?|?|0|--

Your analysis fails to consider that the CRs for those three answers are negative so the correct analysis will show that Michel got 2 right, 1 wrong and 3 skipped.

You analysis fails to consider that there should have been three times as many incorrect guesses as correct guesses. In other words, Michel should have correctly assessed three times as many negative values than correct values. He got lucky on the one right answer with correct credibility and he did worse than chance on identifying incorrect guesses.

All your analysis contends that every test subject actually heard the broadcast and either lied or told the truth. That's a false dichotomy you've got going on there. There are further cases, such as Michel is not telepathic, or that some people are unable to hear the broadcast. You could go further and include couldn't hear but lied and said they did; couldn't hear but guessed anyway; or couldn't hear and told the truth. There are just too many possibilities with such a small scale test, none of which are ruled out at this point.

I'd like to see him repeat the test as is and see if he can repeat the results, especially regarding his CR. This test seems to indicate he is guessing who is legit and who is not.

In your equation, CR*correct, a CR of zero should be given a value of one, otherwise it is hard to distinguish correct*correct from incorrect*correct.

There really is too much wrong right from the beginning to make the results actually mean anything other that some people tried to guess a number between one and four.

 

[Dan O.]

Participant|PM to Agatha|blinded|Post MD5 hash|CR rating|Answer|Correct|Score
Hokulele|Y|Y|N|0|1|-1|0
Ladewig|Y|Y|N|-5|2|+1|-15
stanfr|Y|Y|Y|-5|1|-1|+5
Kid Eager|Y|Y|N|-5|4|-1|+5
fagin|N|Y|N|-4|?|0|--
femke|Y|Y|N|0|2|+1|0
gabeygoat|Y|Y|Y|0|4|-1|0
NaySayer|Y|N|N|?|?|0|--

You analysis fails to consider that there should have been three times as many incorrect guesses as correct guesses. In other words, Michel should have correctly assessed three times as many negative values than correct values. He got lucky on the one right answer with correct credibility and he did worse than chance on identifying incorrect guesses.

Congradulations. You win the gold star for mathematical excellence. This was one of the tests I mentioned multiple times. The proper scoring function has to account for the probability of getting that result. I tried to hint at that with the gambling analogy. If one were betting that a die would roll a specific number, tha payout would be 6 for 1. Conversely, if one were betting that a specific number would not come up the payout would be 6 for 5 (equivalent to betting on all 5 other numbers).

All your analysis contends that every test subject actually heard the broadcast and either lied or told the truth. That's a false dichotomy you've got going on there. There are further cases, such as Michel is not telepathic, or that some people are unable to hear the broadcast. You could go further and include couldn't hear but lied and said they did; couldn't hear but guessed anyway; or couldn't hear and told the truth. There are just too many possibilities with such a small scale test, none of which are ruled out at this point.

I make no such contention. It is Michel that contends that he can tell if people are getting the right answer, the wrong answer or no answer. The protocol tests that claim. If Michel cannot do that his average score will be about zero. If he can do that, his average score will be significantly above zero. Statistical analysis will tell us how far above zero Michel will need to score in order for the result to be statistically significant.

I'd like to see him repeat the test as is and see if he can repeat the results, especially regarding his CR. This test seems to indicate he is guessing who is legit and who is not.

The details of how Michel arrives at his CR. If he can consistently achieve a statistically significant score, he is doing something that is presumed to be impossible by current scientific standards.

In your equation, CR*correct, a CR of zero should be given a value of one, otherwise it is hard to distinguish correct*correct from incorrect*correct.

If Michel wants to use a CR value of one, he is free to do so. A CR of zero is the equivalent of placing no bet. Obviously, if no bets are ever placed, his ability will not be tested and he will fail.

There really is too much wrong right from the beginning to make the results actually mean anything other that some people tried to guess a number between one and four.

And there you are wrong. You need to give back that star.

 

[Nay_Sayer]

I am projecting....#6

 

[devnull]

I got 7. Damn it!

 

[AdMan]

Are we still guessing?

42 No, wait, 84. Doubly good.

 

[kid meatball]

And there you are wrong. You need to give back that star.

Boy, I sure will miss it.

There is clearly not enough data to say either way how well he performed. It was wrong from the beginning to expect any sort of result by performing the test once, with so few numbers, and with so few participants.

 

[fromdownunder]

There is clearly not enough data to say either way how well he performed. It was wrong from the beginning to expect any sort of result by performing the test once, with so few numbers, and with so few participants.

His three protocols here have not been an attempt to provide evidence for telepathy, as he assumes that he has this ability and that everybody hears the number he is sending out.

His tests have essentially been about his ability to provide evidence for who is telling the truth, and who is lying, which is an entirely different thing. He has not satisfied either claim, and has not even attempted to prove the telepathy claim.

Norm

 

[tsig]

tsigs' telepathy test

Ok let's put an end to this foolishness.

I am transmitting a number now you all PM me the number in your mind right now an I'll let you know if you got it right.

 

[Dan O.]

His tests have essentially been about his ability to provide evidence for who is telling the truth, and who is lying, which is an entirely different thing. He has not satisfied either claim, and has not even attempted to prove the telepathy claim.

Norm

If he can correctly identify at a rate above that of chance who is telling the truth about something that could only be communicated telepathically then he would be providing evidence of the existence of the telepathy.

There are a few outstanding issues with the protocol as it exists. The first issue is the number of tests that needs to be performed to be convincing for all sides. For me, I'll set a modest goal of p<.05 repeated three times. That's a cumulative 1 in 8000. If anyone can do that they are ready for the MDC. Before starting the test we would need to run the numbers on some hypothetical examples to determine how many trials and how many correct answers are required to reach that level. Similarly, we would need a commitment from Michel for what level of performance he would need to achieve to continue to believe that telepathy is occurring.

Then there is the issue of cold reading. Can Michel be picking up subliminal clues to the number someone is seeing from the text in their message. This in itself is an amazing ability but not paranormal. Is there some workaround that would block the subliminal clues but still allow the telepathy test? One thought is to take Michel out of the direct telepathy connection but still have him judge the connection between two other participants.

Michel, will that work? If "Bob" selects a random number, writes it down and posts a blinded message with the number replaced with xx and "Alice" reads Bob's message and thinks about the number that Bob wrote down and writes down the number that she thinks of and posts a blinded message about her number with the actuall number replaced by xx, will you be able to read the posts from Bob and Alice and determine if the numbers they wrote down do or do not match?

Will it make any difference if there are multiple communications going on at the same time? What if "Charlie", "Dick" and "Erica" join in and each take turns being the sender and the rest receiving.

There were some issues expressed about the commitment messages. These need to be generated to insure that there is no after the fact changes in the numbers people pick or see. Since there can be no changes to the CRs after tha actual numbers are revealed, the form of the commitment is not relevant as long as there is no ambiguity as to what number is being committed. Michel, if you have any problem with that, it needs to be addressed before the protocol is finalized.

The choice of the random number that the sender writes down needs to be insured random. If the starting number is truly random, the problems with uneven distribution of the numbers people see will be averaged out.

 

[carlitos]

Congradulations. You win the gold star for mathematical excellence..

Congratulations. You win the gold star for spelling excellence.

 

[Dan O.]

Congratulations. You win the gold star for spelling excellence.

Proove Webster was wright.

 

[ComfySlippers]

How an adult can waste time arguing about a silly test designed to prove something that doesn't exist is truly baffling.

 

[John Jones]

You demonstrate in your argument that you are ignorant of the fact that Michel has been seeing a real doctor. He doesnt need someone calling him names and applying labels to him. He is searching for answers and he needs someone that will honestly listen to him and help him find facts. Did you read that paper I linked to earlier about hearing voices? Are you still as ignorant about that subject as the medical field was until quite recently?

I am not playing games. I try to answer truthfully where I can. I also choose to ignore the majority of the off topic uncivil disruptive posts. Why do you insist on calling me a troll? Is it because I refuse to join the mob mentality that pervades this thread? Is your prejudice coloring your view of my posts?

I think it's more likely because you live under a bridge, and attempt to terrorize the three billy goats gruff.

But I've been wrong before.

 

[John Jones]

Proove Webster was wright.

Sea my sig,

 

[ComfySlippers]

John Jones had it right.
This is the worst troll ever.

Why do Trolls assume their victims are 3year olds?

 

[TheSapient]

You demonstrate in your argument that you are ignorant of the fact that Michel has been seeing a real doctor. He doesnt need someone calling him names and applying labels to him. He is searching for answers and he needs someone that will honestly listen to him and help him find facts. Did you read that paper I linked to earlier about hearing voices? Are you still as ignorant about that subject as the medical field was until quite recently?

I am not playing games. I try to answer truthfully where I can. I also choose to ignore the majority of the off topic uncivil disruptive posts. Why do you insist on calling me a troll? Is it because I refuse to join the mob mentality that pervades this thread? Is your prejudice coloring your view of my posts?

I think this is an interesting admission. So why is it that you never answer most civil, on-topic questions? Is it because you can't answer truthfully? What you prevent you from doing so?