There was something of a "fad" in literary research a few years back (few on the archaeological time scale I work to) where people did immensely complex analyses of "old" works to solve burning questions like "did one person write all the Shakespeare plays" or "who was Shakespeare" etc. ad. infinitum.
Simple word counts don't tell a whole lot, but repetitive sequences can pretty well pin down "anonymous" authors, if one has a sufficient sample of their work.
As to analysing one's own work to determine whether some degree of repetitiveness or other banality is present, I would think that anyone needing "mechanical" means to detect weaknesses would probably not tend to see the meaning of the results. Conversation with one's peers, or the services of a respected mentor would be more likely to be productive.
Expecting any program, and especially a free-ware one, to contribute much to the quality of one's writing is akin to letting a Cosmopolitan instan quiz decide whether you have a satisfactory sex life.