Re: How Cryptographers Assign their Probabilities?

From: Gianna Stefani (jxeql_at_jwhjdig.jqk)
Date: 03/08/05 

Date: Tue, 08 Mar 2005 12:37:18 +0100

Fab wrote:
>
> I'am gonna toss a coin. I say "the probability to get a head is 1/2":
> this is a definite numerical value of a probability, isn't it?
>

That is how to calculate the probability, yes. It *is* the probability
... it is not the *definite probability* ... by adding the word
*definite* you add confusion.

Generally it is agreed that a coin has two ways in which it can land
(two sides) and that it must land on any one of them ... thus a
probability of 1/2. (I have seen one land on its edge ... once.)

If a coin has n sides, the probability would be 1/n.
If the coin was representing the key to a cryprogram, the probability
would be 1/n.
If a cryptogram has n possible keys, the probability that the one you
just guessed at is the correct one is 1/n.

> But what does it really in Cryptography? How should I interpret such
> statement? This my question.
>

It works the same way. If there are n possibilities, in your example n
possible keys, the probability of any one of them occuring or being
selected is 1/n.

>
> Please, I do not follow you. What do you mean by "non-question"?
> Does it mean that my question is not well-posed, that it is a question
> that should not be asked or that it has a definite answer?
>

I mean that the question does not logically exist. The probability *is*
the probability. There is not one set of probabilities used by everyone
except cryptographers and a separate set for cryptography in which the
principle of 1/n does not apply. 

-- 
Gianna Stefani

Relevant Pages

  • Re: Sort benchmark and a classic paper by Gray
    ... Statistical complexity theory predates computer science, ... of the keys not being too evil, ... This is of course a hash table, with the additional benefit of knowing ... provably Owith probability one in the number of operations. ...
    (comp.arch)
  • Re: Analyses of Wangyong Mistakes
    ... cryptosystem does not impact on the probability of plaintext. ... The last part of your point i.e. "unknown but fixed ciphertext, ... to-one correspondence between all the plaintexts and keys for fixed ...
    (sci.crypt)
  • Re: SHA1 broken
    ... >> probability of occuring above a given threshold. ... When the diff occurs only a limited subset of keys are possible. ... for all p then the attack can't work. ...
    (sci.crypt)
  • Re: Confirmation of Shannons Mistake about Perfect Secrecy of One-time-pad
    ... For the above example, the probability ... plaintext changes when the ciphertext is fixed, ... It is unknown (it depends both on the fixed value of the cyphertext, ... all the keys are equally likely ... ...
    (sci.math)
  • Re: perfect security
    ... > rate of the message stream. ... For a system to be perfect secure ... If the number of keys is equal for each m \in M, ... Btw. the probability Pis the a-priori probability and Pis the ...
    (sci.crypt)
Quantcast