The Grand Rascal
New member
- Joined
- May 24, 2021
- Messages
- 10
IMPORTANT: I am NOT a student, and this is NOT any form of, or in any way related to, any homework assignment. I am a private person, 62 years of age.
-----
I'm investigating the quality of the output of the "HotBits" random number server, at:
(URL removed)
HotBits derives its random bytes by measuring the interval between pairs of radioactive decay events. Supposedly, this guarantees randomness because their entropy source is governed directly by quantum mechanics.
I have gathered 168,656 (supposedly) random bytes from them, and wish now to test the quality of their randomness.
While I'm certain there must be many dozens of different tests for randomness, there is really only one that interests me: equal chance of selection. To be truly random, each possible value must have the same likelihood of occurrence.
So what I'd like to do, is to write a program in Commodore BASIC (I'm a C128 hobbyist) that will read all of the random bytes and test the frequency distribution of the byte values for linearity -- the more linear the distribution, the better!
My problem is that I don't know how to measure the linearity of a frequency distribution.
Is there some kind of simple formula, or a simple set of steps, that will provide me with this measurement?
Many thanks, in advance, for any information.
Range of possible values: 0 - 255, inclusive.
Number of values: 168,656.
--The Grand Rascal (a.k.a. "C128User")
-----
I'm investigating the quality of the output of the "HotBits" random number server, at:
(URL removed)
HotBits derives its random bytes by measuring the interval between pairs of radioactive decay events. Supposedly, this guarantees randomness because their entropy source is governed directly by quantum mechanics.
I have gathered 168,656 (supposedly) random bytes from them, and wish now to test the quality of their randomness.
While I'm certain there must be many dozens of different tests for randomness, there is really only one that interests me: equal chance of selection. To be truly random, each possible value must have the same likelihood of occurrence.
So what I'd like to do, is to write a program in Commodore BASIC (I'm a C128 hobbyist) that will read all of the random bytes and test the frequency distribution of the byte values for linearity -- the more linear the distribution, the better!
My problem is that I don't know how to measure the linearity of a frequency distribution.
Is there some kind of simple formula, or a simple set of steps, that will provide me with this measurement?
Many thanks, in advance, for any information.
Range of possible values: 0 - 255, inclusive.
Number of values: 168,656.
--The Grand Rascal (a.k.a. "C128User")