Re: How Random are Cards Really Dealt?
As I said before, using the pseudorandom generators is a practical concession to the annoyance of generating really random numbers. That said, if you have a pseudorandom number generator with such a huge range for seeds, what predictive power do you have? If the RNG is broken and can only make use of 128 bits of seed, you have a point, but if it can use 2000 bits of seed and actually produce different streams of numbers, then I don't see the problem.
There are posts from RGP back in the day where people took a sample of 50,000 or 100,000 hands and subjected them to analysis to see how often hand distribution, expected win percentages, etc. held up. Obviously that's not a huge sample, but the numbers were all pretty close to what you'd theoretically expect. At the very least, it's some evidence to corroborate the theory that everything is peachy with the RNG. It would be nice to see what the "shuffling audit" that you see places talk about occasionally actually consists of. I looked at Paradise, but all I saw was a piece of PriceWaterhouseCooper stationery that essentially said, "Yep, it's random!"
Until I see a similar amount of evidence pointing to flaws in the random number generator, I'm going to continue thinking it's okay.
|