I thought that this would look for unique words rather than just cycle through all words. You should have a default ignore word list that includes things like el, en, de, etc etc.