Markus Schulze wrote: >It is very problematic that RP and SSD are currently discussed as if RP >could be used only with margins and SSD could be used only with >winning-votes. (Actually, it seems to me that Adam Tarr really believes that.)
I recognize the distinction, I guess I just don't see the point of worrying about it. Is there any difference between "Plain Condorcet" as Mike calls it and Ranked Pairs computed with winning votes? One assumes all defeats are valid, and drops the weakest ones until there are no inconsistencies. The other starts adding victories from the strongest to the weakest until an inconsistency is found. Will they ever produce a different result? I see a small distinction between SSD and these methods, since it throws out the candidates outside the Smith Set first. But in the example I've been discussing, all candidates are in the Smith Set, so this is irrelevant as well. It also seems like I could design a method called "Smith Ranked Pairs" that drops the candidates outside the Smith Set, and then runs ranked pairs, which would be equivalent to SSD as far as I can see. So, I have equated SSD/PC with winning votes, and Ranked Pairs with margins, simply because that's the way each is presented in all the hypertext I have seen, and since I don't see any other distinctions between the two worth discussing. Since this seems to bother some people here, I can change to referring to winning votes and margins directly. But if there are other distinctions between Ranked Pairs and Plain Condorcet, I'd be interested to know what they are. -Adam
