3 ms·

Can the bias not be transformed out? Naively, I'd assume that if you give me a biased RNG with a known distribution, I'd think I could just unroll the distribu

• MauranKilom 6y ago
  Curious how you would do the lookup if you have an RNG with probabilities 2/7, 2/7, 3/7 (or the same but with significantly larger denominators).
  • tape_measure 6y ago
    How about this? In rare circumstances, it may loop forever. def uneven_random(): # return 0 with 2/7 probability, 1 with 2/7 probability, and 2 with 3/7 probability n = random.random() if n < 2.0/7: return 0 if n < 4.0/7: return 1 return 2 def transformed_random(): # return 0 or 1 with 50% probability N = uneven_random() while N == 2 : N = uneven_random() return N
• IncRnd 6y ago
  Yes, as you rightly point out, the bias doesn't need to be introduced. This is a well known issue with divisions(modulo) in generating random numbers. This is literally why there is an injunction for people not to create their own crypto. Of course, the meeaning isn't that people aren't to work on crypto but that people shouldn't create RNGs like this thinking they are better than the prior-art, when they actually regress with already solved probems. This is the exact different between arc4random() and arc4random_uniform(). arc4random_uniform() is recommended over constructions like arc4random() % upper_bound as it avoids "modulo bias" when the upper bound is not a power of two. Unfortunately when I google for ios random number The very first return says We can use the C function rand() for this: int i = rand()%10+1;