This question already has an answer here:
Probability and statistics though related can be seen as separate topics. In my mind the former is more about thought experiments and understanding processes and the latter more about data and their description. They have clearly defined function sets that don't overlap too much. We have the distribution family, with Expectation and Probability on one side, and Mean, Skewness, LocationTest and what have you on the other. Catching them under the same tag would be similar to grouping topics such as animation and graphics.
I realize my position is weakened by the docs themselves which also group them quite often. There's even a Statistics and Probability guide page. And Mean c.s. work on distributions, so that could be confusing. I suggest we wait whether this will be a problem and if we see such a pattern emerging do something then and not now.