I haven't yet got the time to re-write them but merging the advantages of Approval, IRV and Ranked Pair was the goal of methods I), J) and K) described as candidates for the single-winner contest on the site http://groups.yahoo.com/group/Electoral_systems_designers/messages/113
Actually I) is Ranked-Pairs using relative margins as victory comparison criteria. It is the best criteria I can think of and I will still try to prove so to Mr. Ossipoff as our argument goes on... J) combines all three methods advantages I wanted: it uses ranked pair to rank candidates so to minimize the highest twarted (how do you write that?) majority, IRV's sequential dropping so you can gather support of different candidates with similar philosophies and residual approval weights so ranking unwanted candidates should not harm any of your preferred candidate. In addition, it produces weights as output so it can be used as a single-winner inner motor to produce a multiple-winners method with an exact proportional representation. K) shows how to generalize the popular IRV method to obtain weights as output... Unfortunately, the quality of my explanations could be enhanced... Maybe I should not have given examples with ties, this complicates the process. Would you Matt be interested in reading them and give me some feedback? It is the next job I have to do whatever. Feel free to join my group too... Steph. ---- For more information about this list (subscribe, unsubscribe, FAQ, etc), please see http://www.eskimo.com/~robla/em
