🚀 go-pugleaf

On 28/09/2010 4:46 PM, |-|ercules wrote:
> "Sylvia Else" <sylvia@not.here.invalid> wrote
>> On 28/09/2010 12:48 AM, |-|ercules wrote:
>>> As Sylvia is the only person to come up with genuine questions
>>> to test my psychic dominance, I offer her this challenge.
>>>
>>> Shuffle your original 6 cosmology questions and post the MD5
>>> of the file with a private keyword appended so I can't decrypt it.
>>>
>>> Read out the first question while filling out the first page at
>>> www.HercsHome.com/AskPsychic.html
>>>
>>> Fill in the second page using any 2 other questions from the list.
>>>
>>> Repeat for questions 2 to 6.
>>>
>>> If I can tell you the order of your 6 questions Sylvia puts in her
>>> sig line
>>> "Herc broke 729:1 odds when I tested his mind reading powers".
>>>
>>> If I get one question wrong - I put in my sig any of Sylvia's quotes of
>>> her choice, "I failed a psychic test of my own design" etc. etc.
>>>
>>> And the sigs stay on all posts and replies for 1 month. At least 1 post
>>> in aus.tv
>>> and sci.skeptic.
>>>
>>> Do you accept my psychic challenge Sylvia?
>>
>> Not as it stands, no. You would certainly renege on your obligation
>> regarding a sig line based on some claimed failure on my part to
>> comply with the rules. For example, you would claim that I hadn't
>> actually read out the questions at the appropriate times. Also, I'm
>> not going to take part in any test where my inputs are not in plain
>> view, so as to prevent you from disputing them later, which means no
>> web forms. It's Usenet or nothing. Thus all the steps I'm required to
>> take must be capable of being evidenced by a posting, either at the
>> time, or later by the revelation of a match for a previously posted
>> MD5 hash. That obviously implies no requirement that I not do certain
>> things.
>>
>> 6! is the number of possible orders, but each time I provide a list of
>> three questions on the second page, I'm telling you that one of those
>> three is the answer, and cannot appear as the answer in a subsequent
>> test. So the odds you've stated are wrong and higher than in reality.
>>
>
> Yeh I thought that too, weird stat to calculate.

You didn't mention it though. Hoping I wouldn't notice?
>
> What about a 1 (ONE) question trial, if I get that right we revise the
> protocol down to a micron so you're not wasting your time asking
> me 5 questions.
>
> Herc

And have you claim that I bailed out while you were winning, when we 
fail to agree on the revised protocol? I don't think so.

Calculating the odds is being complicated by in each case having only 
three possible questions out of the overall six.

If I shuffle the questions, and then create an MD5 for each question 
(plus random key) in turn, I can post a single list of six MD5s in the 
same order as the shuffled questions. I'd post an MD5 of the numbers of 
the shuffled questions, e.g. 634215, plus a random key, to show later 
that the posted individual MD5s are in the same order as the questions 
they represent. Your odds of getting the correctly ordered list by 
chance is then 1 in 6! as you observed earlier, because there is no 
information leakage on the way.

The problem for me is that it still leaves you wriggle room to claim 
that I created the MD5s of the questions first, and then shuffled the 
list of MD5s and posted those. You could then claim that I construed the 
order for the shuffled questions backwards from the MD5s.

Outcomes in the two cases are indistinguisable, but I think you'd claim 
it mattered if you got the order wrong.

I cannot see any way around the general problem that you're willing to 
claim that your power goes beyond extracting the secrets behind what I 
posted, and is affected by what I did, or thought. The latter would be 
problem even with volunteers being scrutinised by judges, because there 
will never be any objective evidence about what the volunteers thought.

Sylvia.