Juho wrote: > The "cooling off" type of rules is what I'd expect to see in place to > eliminate too radical and surprising changes. > > Not all changes in opinions are malicious (coups or something like that). > It is also quite possible and even typical that some representative changes > his opinion from A to B since he thinks alternatives A and B are about > equally good but he now sincerely thinks that B is now better. People that > have given their vote directly or indirectly to this representative may > however sincerely think that option A was good but B is disastrous > (probably based on some other criteria than what the representative based > his opinion change on). > > (The problem was thus one representative moving large masses of votes too > quickly for the voters to react and change their vote.)
Yes. It's a problem at any cusp of decision, where a single snapshot of the cascade has to be frozen as a "final result". But I think that manual ratification (1) is just as valid a solution, as the cooling off period (2). It's a given solution too. No government is going to accept the results of an open, continuous election to a public office, unless those results are ratified through its own, traditional electoral system. So the voters will take the assent which they have long expressed in the continuous election, and translate it to a "final result" at the polling station, one vote at a time. That slow, manual process will effectively filter out any sudden vote shifts that had occured at the last minute, without their actual assent. Suppose that the cascade has already gelled into a more-or-less stable shape, well before election day. Then the voters will already know the rough structure of their new government, and it's *that* they'll be giving their assent to on election day. But how it gets to that shape in the first place, passing through a transition - perhaps in mid-term, for a Mayor who is about to retire - I wish I had a better sense of the dynamics at that crisis point. My thought is turning to vote cycles. What happens if the Mayor tries to annoint a successor, simply by voting for him? The two would then be co-joined in a tight cycle, sharing the Mayor's electoral support. If that support held for a period of time, and especially if the two were actually sharing power, then it might decide the issue. On the other hand, what happens if the central delegates (all voting for the Mayor) suddenly form into a ring, voting for each other? They will catapult straight to the top of the standings, while the Mayor plummets to the second or third rank. If their ring can hold its electoral support - if its own combined support is deeper than the Mayor's - then the ground might be prepared for a new decision. This is the tricky part. The delegates must somehow transform the ring into a proper tree root, pointing to a single candidate. Maybe there's an orderly way to do that (thinking of Raph Frank's loop-to-tree algorithm) in which the weaker members drop out, or are ejected one by one? They could then turn their votes back into the ring, and freely shift them around till they settle on the next Mayor. Or maybe that kind of orderly procedure is impossible in a competive context, and the ring would just break apart into a free-for-all? -- Michael Allan Toronto, 647-436-4521 http://zelea.com/ ---- Election-Methods mailing list - see http://electorama.com/em for list info
