# Is random.org truly reliable?

Can someone help me in gaining an understanding of the reliability of random.org being used by a University to select student applicants to extend invitations to?

• What exactly is your question? Jul 10 '20 at 11:48

Yes, using random.org to randomly select students is appropriate, since random.org creates truly random numbers from atmospheric noise.

But for cryptographic purposes you shouldn't rely on random.org as this post explains.

Random.org does provide high quality randomness. However, it is relatively easy to forge the results.

I'd recommend using your systems CSPRNG over the website, to compromise randomness for authenticity.

Random.org claims to generate truly random numbers from atmospheric noise.

It is just claiming cryptographic applicability, so don’t use it that way.

The details of how exactly how they do that is not given anywhere on the site or elsewhere. 3 radio receivers are mentioned somewhere along the way.

The main issue with pure randomness from natural processes is whether it needs post processing for use as a uniform source over whatever target set you need.

You would need to do some testing yourself, after specifying the parameters of what you need it for, which seems to be selecting $$M$$ students out of a candidate pool of size $$N$$, i.e., a universe of size $$\binom{N}{M},$$ over which you need some approximation of uniformity.

They also seem to have some limitations on how much randomness they supply, per request.