Say I have a sequence that starts with:
H, H, H, H, H, H, H, H
Now each time I get another element one of two things can happen:
- it exhibits the same amount of randomness as the previous element T, T, T, T, T, T, T, T
- it exhibits more randomness than the previous element H, H, T, H, H, H, H, T
Assuming randomness is something that can be quantified, how do you analyze the sequence to determine the point (best approximation) at which it switches to completely random?
Note: in case this came out sounding like a homework problem (not sure it did) the context of this is using a pseudo random generator that seeds based on the current time in quick succession (very close time values) - they seem to generate sequences that are similar to each other and I wanted to test my theory (may very be wrong) that there's a minimum amount of time (greater than the smallest quantum in the target system) which must elapse from one run to the next in order to produce independently random results.