4 ms·
I'd really like to see a Chi-square analysis on this by a smarter statistician than me. Power should be pretty low since you only have 16 bytes. Alternatively
by drewcrawford 17y ago
I'd really like to see a Chi-square analysis on this by a smarter statistician than me. Power should be pretty low since you only have 16 bytes.
Alternatively, in a series of sixteen uniformly-distributed bytes, what's the probability that the birthday paradox doesn't occur? That is, the probability that no two bytes are identical?
I can say as someone who's played around with English-language markov chains for years that this definitely isn't displaying the characteristics of English text, even with transformations like rot13 or multiplication. So the next logical question is, "Is it uniform?"
- roundsquare 17y agoHow would you do a chi-square on this? Doesn't seem to me to be the right sort of test. If you think its hex/ascii, you can try to do a hypothesis test on the max frequency of each digit/letter and see what happens there. But, for any test, you need an underlying model against which you test. What model are you thinking of?
- ghotli 17y agoCan someone explain why chi-square analysis would even be a useful thing to perform on this string?