-
@ cee87b5a:b88898bd
2025-02-24 19:33:20
we generate set of random inputs as sample data.
we use 256 bit maximum value as input, which is sequence of 1s.
then we use the output of hash(input) as input for statistical prediction function, which should do best estimate for the correct input. optimal answer is exact same as input value, 256 bit sequence of 1s.
if any correlation can be found, we should see it if we run estimation enough times and find statistically that 1 occurs more often than 0. this result would indicate that estimate function is getting closer to the input value which is sequence of 1s.
example plots above is just single run. to find any statistical correlation, significant amount of iterations need to be run for candidate function.