In reply to rakesh reddy:
Your distribution constraint says for each randomization, the odds of choosing a 4 or 2 are 50%/50%. This is the same as flipping a coin 10 times in a row and expecting to see the same side exactly 5 times. The odd of that happening are quite low (10 choose 5)/2**10 = ~25%).