distributional assumptions

if you have a scenario where 6 balls are in a bag, numbered 1 to 6, and you take 3 out (with replacement) how would you find the probability distribution of the median of the numbers on the 3 balls,

i was able to do it for without replacement by considering the 20 different cases that arise but here i can see there would be 56 distinct cases, though not all equally likely, so i dont think that is the best way to go about it (by sheer computation)

any ideas guys

cheers

djx