• 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.

My Second Telepathy Test

Actually, assuming that the purportedly generated md5 hash in the op represents an actual string of 5 digits, people do not need to post incorrect guesses since they can generate an md 5 hash from their guess and compare with the one Loss Leader provided.

It's a bit of an interesting psychological experiment.
 
Wait...I'm getting something...I'm getting a T, or a C, or a P, or something like that. Anyone? Anyone know someone whose name begins with those letters? They want you to know that they miss you, and they are somewhere near a wood, or in a wood, or in something made of wood, and...and...Oh! 91100.
Wow. I love this telepathy thing.
 
Loss Leader said:
I ... don't know any of that. If anybody gets it right, I'll ask them whether they were 95% confident, I guess.
Click to expand...
:) that's uncomfortably close to a Michel H approach. If the idea is to show how this kind of test could be done 'properly', it really needs some minimal statistical protocol.

I used to do this kind of stuff (chi-squared tests, etc.) when I was in medical research, but that was over 30 years ago, so I don't remember the details now...

Having said that, I'll go 47395.
 
Last edited:
I just occurred to me to wonder whether guesses at a random digit sequence might have a skewed distribution of digits because that's what we tend to see in daily life due to Benford's Law.

I suppose the same kind of argument could be made for guesses at a random letter sequence (guesses might be biased by the frequency & position of letters in read text).
 
Norman Alexander said:
Since you haven't limited us to one guess per poster, these are my guesses:

00000
00001
00002
00003
00004
00005
00006
00007
00008
00009
00010
00011
...and so on up to...
99990
99991
99992
99993
99994
99995
99996
99997
99998
99999
Click to expand...

Daylightstar said:
Actually, assuming that the purportedly generated md5 hash in the op represents an actual string of 5 digits, people do not need to post incorrect guesses since they can generate an md 5 hash from their guess and compare with the one Loss Leader provided.

It's a bit of an interesting psychological experiment.
Click to expand...

Or a challenge on who can write code to brute force it in the least number of lines.
 
Last edited:
dlorde said:
I just occurred to me to wonder whether guesses at a random digit sequence might have a skewed distribution of digits because that's what we tend to see in daily life due to Benford's Law.
Click to expand...
I think it's a given that human attempts at randomness are not actually random, but I don't think LL's test is asking for random guesses (even if participants are trying, at times, to be such).


dlorde said:
I suppose the same kind of argument could be made for guesses at a random letter sequence (guesses might be biased by the frequency & position of letters in read text).
Click to expand...
I would be surprised if there isn't a corollary of Benford's Law for letters.
 
Daylightstar said:
Actually, assuming that the purportedly generated md5 hash in the op represents an actual string of 5 digits, people do not need to post incorrect guesses since they can generate an md 5 hash from their guess and compare with the one Loss Leader provided.

It's a bit of an interesting psychological experiment.
Click to expand...


The MD5 hash includes the numbers in a simple sentence. There is nothing you can know beforehand about the sentence. It makes brute-force close to impossible, I hope.

As regards my statistical methodology, I don't expect to need any. However, if anything interesting comes of this, I will be glad to consult with real stats people to design a better test.
 
Loss Leader said:
The MD5 hash includes the numbers in a simple sentence. There is nothing you can know beforehand about the sentence.
Click to expand...
umm.. it has to be in the format of #####, no? I guess you could have generated it from by spelling out the letters like "zero two one three four" and to complicate things you could separate the numbers with commas or semicolons or dashes or whatever...

It makes brute-force close to impossible, I hope.
Click to expand...

Holy crap, generating the hash is slow. I'll let you know in about a day or two
 
CynicalSkeptic said:
umm.. it has to be in the format of #####, no? I guess you could have generated it from by spelling out the letters like "zero two one three four" and to complicate things you could separate the numbers with commas or semicolons or dashes or whatever...
Click to expand...
OT and outside my expertise, but I have a layman's interest.

I don't think that's what LL meant when he said the numbers are "in a simple sentence." What I think he meant was that he created something like this:

2 horses ran 1 time around the number 7 track before being given 6 apples and 2 sugar cubes.
 
Garrette said:
I don't think that's what LL meant when he said the numbers are "in a simple sentence." What I think he meant was that he created something like this:

2 horses ran 1 time around the number 7 track before being given 6 apples and 2 sugar cubes.
Click to expand...

Yeah, thanks for the explanation. That makes sense.
 
You must log in or register to reply here.

Back
Top Bottom