Regarding the article's example #3, I'd like to add that, in fact: If candidate "Better" turned out to be more competitive than expected, with more support than expected, s/he could take the win away from "Classy" or from "Awful". The possibility of Better taking the win away from Awful is worth the possibility of Better taking the win away from Classy. ...further supporting & confirming the general suggestion that it's always optimal to approve (only) all of one's top-set. Michael Ossipoff