Do you know how it then maps hashes to buckets? I'd expect the natural occurrence of integer values to be very non-uniform and heavily biased toward the small end.