I have a preference for shorter tag names when possible. That is why I created a tag for rayleigh, and not [rayleigh-distribution]
(but I didn't really think about it).
When there are enough questions categorized under a tag, if there is a demonstrated need, we can create additional tags that will further differentiate a category. As an example, I split up fisher when it seemed to become problematically heterogeneous (in my mind) and created new tags for fishersexact, fisher-information, fisher-transform, ronald-fisher. I do believe that having the right tags, and having threads tagged appropriately, is very important to the organization and usability of the site. However, I think think that this will always be a work in progress, requiring some continued maintenance as the site evolves; I don't think it is very possible or useful to try to predict what tags / differentiations will be needed in advance. For instance, you make a good point about the ambiguity of chi-squared, but I haven't noticed an actual problem. If someone else has, or if one emerges, we could split the tag up. Until then, I'm inclined to let it be.