c - Uniformity of random numbers taken modulo N

Question

One common way of choosing a random number in [0, n) is to take the result of rand() modulo n: rand() % n. However, even if the results returned by the available rand() implementation are fully uniform, shouldn't there be a problem with the uniformity of the resulting [0, n) numbers when RAND_MAX + 1 does not divide evenly by n? E.g. suppose RAND_MAX is 2, and n is 2. Then out of 3 possible rand() outputs: 0, 1 and 2, we get 0, 1 and 0 respectively when we use them modulo n. Therefore the output will not be uniform at all.

Is this a real problem in practice? What is a better way of choosing random numbers in [0, n) uniformly deriving from rand() output, preferably without any floating point arithmetic?

Answer 1

You are correct, rand() % N is not precisely uniformly distributed. Precisely how much that matters depends on the range of numbers you want and the degree of randomness you want, but if you want enough randomness that you'd even care about it you don't want to use rand() anyway. Get a real random number generator.

That said, to get a real random distribution, mod to the next power of 2 and sample until you get one in the range you want (e.g. for 0-9, use while(n = rand()%0x10 > 10);).