It is possible to inform the Test library that “exactly so-many tests will be performed.” If this has been done, you will be told if the actual counts differ. However, you are not under any obligation to do so.
My personal feeling, of course, is that you probably do want to avail yourself of this facility as much as possible, even if you calculate the expected count. This gives you a cross-check on the test itself (which, of course, will inevitably contain bugs, just as surely as the code that you are testing). When it makes sense to do this, it’s usually a good idea, IMHO. (The computer is always the best one to detect when you have mispeled something, or accidentally something out.) ;-)