Hacker News Reader

ln(52!)/ln(2)≈226 bits. You don't have to depend on a single random 64-bit number: https://en.wikipedia.org/wiki/Fisher%E2%80%93Yates_shuffle

View on HN

If the RNG has < 226 bits of state, such as most non cryptographic PRNGs, you cannot reach all shuffles even if you use multiple numbers.

I'm not totally sure that's true. I get where you are coming from, an 8 bit rng can only output 256 distinct values at most. 256 would be the longest cycle.

However if you combine the outputs of two, but don't step them at the same time, you can have more outputs.

Imagine you had one that outputs just 0 and 1 and then loops. You could have two of these updating at a different frequency and have four distinct outputs.

I think that makes sense.

Oh, yeah, there are ways to finesse it. I’m just wondering whether anyone involved in those early poker machines had made the effort.

In Nevada one was required to “prove the randomness” of random outcomes in video poker machines as well as other electronic gambling machines. So effort was made.

Hey, sorry to contact you via this method. I meant to reply to your comment under the thread on SCALE but it's locked. I'm expecting an offer from that company and would like to ask you about your experience there. If you don't mind, could you accept my connection request on LinkedIn and we could chat there? Thank you :)

Reply on news.ycombinator.com