NIST statistical tests

wxist

3/6/24, 2:54 PM

I'm having trouble testing a not-so-popular algorithm that I haven't found an implementation of, so I wrote it myself and now I'd like to test it with nist tests, but I have a suspicion that I'm doing something wrong.

I got encrypted file with ~10,000,000 bits. I tested my algorithm in the following wayю. I set the input parameter to 1,000,000, it will look like: /assess 1000000, then the amount of bitstreams i put 10: How many bitstreams? 10. Then i chose to test my file as binary, since every symbol in my file is a char, which contains 8 bits:

    [0] ASCII - A sequence of ASCII 0's and 1's
    [1] Binary - Each byte in data file contains 8 bits of data

   Select input mode:  1

I want to know if I'm doing everything right. And what will be the changes if I set the input parameter to 10,000 and the number of bitstreams to 1000?

Thank you for your help!

0 + 6

statistical-test

symmetric

nist

indistinguishability

Maarten Bodewes

3/6/24, 4:29 PM

In hindsight you're probably talking about an encryption algorithm? You haven't specified anything other than "encrypted file" when it comes to the algorithm so I initially tagged it for a PRNG, but that's probably wrong. Please add details & retag.

wxist

3/6/24, 5:07 PM

@MaartenBodewes I have an encryption algorithm. By using algorithm i try to encrypt file with a lot of texts. After the work of my algorithm, I get a file with ciphertext at the output. Since i want to test my algorithm for nist statistical tests i provide encrypted file

Paul Uszak

3/6/24, 9:08 PM

Yes, you're doing it right :-) and welcome wxist. So, if you're uncertain of the input, what about the output? Post it please...

Paul Uszak

3/6/24, 9:10 PM

1000 bitstreams will not tell you anything.

wxist

3/7/24, 10:10 AM

@PaulUszak i got another question about output of NIST statistical tests, it would be great if you check it out

fgrieu

3/7/24, 10:48 AM

I closed the question because the OP posted another a [more detailed version](https://crypto.stackexchange.com/q/105535/555)

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: NIST statistical tests

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.