Well, I don't agree, but I hate combinatorics and probability so I might be wrong. Later I'll write a simulation that does a few million card deals and we'll see what the empirical results are.
I have to apologize, you are right, I still don't see why, but I'll have to have a think about it later ...
