There is a normal distribution of results in a dice pool, but it's only a factor in games where the variable is number of successes as opposed to target number or number of dice.
This is funny, but that's not what normal means.
I'm pretty sure the EV of xKy on a die of N sides is Sum{ni *x/N} where ni=N, N-1, N-2... and nn=y/(x/N)
If the only benefit of a system I am asked to evaluate is that it increases in immeasurable metric, I must evaluate it's use as minimal.