6

I have code that pics a random number from 0 to 1. I am seeing that the number 1 is coming up far more times then the number 0 then I would think to be statistically possible.

This is my code:

int shipNumber = arc4random() % 2;

Should this code work? Am I just going crazy?

DOK
  • 32,337
  • 7
  • 60
  • 92
Mel
  • 2,055
  • 2
  • 26
  • 45

2 Answers2

5

That code should work.

What I suspect you're seeing is truly random (or, at least, sufficiently random) and your brain is trying to find patterns. (Everybody's brain tries to find patterns everywhere. That's how you're reading this. The issue is there are no patterns in randomness [that being pretty much the definition] for your brain to latch on to, so it invents some.)

If you really want to check your output for randomness, you'll need to do a statistical analysis of some kind or other.

Cajunluke
  • 3,103
  • 28
  • 28
0

You might be seeing modulo bias.

Community
  • 1
  • 1
spacehunt
  • 803
  • 5
  • 8
  • 2
    there is no modulo bias in this example because it is modulo 2 and there are exactly the same number of even and odd numbers in the range 0..2^32-1 – jjwchoy Jun 05 '11 at 14:58
  • From the arc4random manpage: "`arc4random_uniform()` will return a uniformly distributed random number less than `upper_bound`. `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". Since, in this case, `upper_bound` is a power of 2, modulo bias shouldn't be a problem. – splicer Jun 10 '12 at 19:44