True random vs. Pseudo Random (can you pseudorandom true randomness)

Question

Ok, so this question involves a little bit ahead. Bear with me. There is this site random.org (and others like it) that pretend to use some kind of quantum process or another to produce true random numbers. If someone will request this site again and again and develop a massive log of true random numbers. This magazine is then rebuilt by the program to mix it as soon as possible. Is the result less random than when it was launched? How much? Any good / cheap further reading on this?

Answer 1

Reordering random numbers by fixed substitution does not change the degree of randomness.

So, if you have the perfect source of random numbers, then the same shuffled bits will be equally random. This will be true if the “shuffle” is a fixed reordering (for example, reversing all bits) or a shuffle generated by a pseudo random number generator (which is actually a very confusing way to determine a fixed reordering from some initial seed).

This is provable from the underlying mathematics - if you reorder a set of truly independent identically distributed random variables, then the resulting distribution will be the same as the one you started with. Therefore, it is equally random.

However, this does not work if the shuffling somehow depends on the values ​​of the random bits. If, for example, you sort the bits rather than rearrange them, then you will not have very good random output :-).

Answer 2

This will depend on how you reorder them. If you used a pseudo-random function for this, the results are likely to be less random. If you use true random chance to reorder, it will not be more random.

One thing that people forget is the reason for using a pseudo-random function over some true random numbers, and testing repeatedly. If you get unexpected results using a pseudo-random function, you can look at a possible problem.

Answer 3

If you have a process that needs N 'random' numbers, you can take N from the site and use them in that order, and everything will be fine. If you shuffle them, you will make them less random.

If you need a constant supply of random numbers, then the question is the relative quality of any pseudo-random juggling in comparison with what would happen if you had a true random sequence.

Since, however, linux and windows provide real random numbers using hardware entropy, why not just use them?