I think the clumping is more about how many consecutive same numbers there are in the array. For example, a naive me would generate: [3,4,5,1,2,4,2,1,2,6] wher

I think the clumping is more about how many consecutive same numbers there are in the array. For example, a naive me would generate: [3,4,5,1,2,4,2,1,2,6] whereas a true random distribution might generate: [3,4,4,4,4,1,2,2,2]. When you generate 6 million samples you would indeed have about 1 million per number. However you would still have subsequences of the same number.