Talk:Status Quo Approval

From Electowiki
Jump to navigation Jump to search

I came up with that on my own. So it pretty much doesn't belong to wikipedia (yet:). Please tell me what criteria it passes and fails. It looks to me like it passes Majority, FBC, Condorcet Loser (sure), Monotonicity, Participation(?) and Consistency(?) . But that's too good to be true.R.H. 16:15, 14 Jun 2005 (PDT)

First we count who gets the most marks in the first and second column added together. That's the challenger. Then we look how often the challenger is ranked better than the status quo. Only if that happened more often than the "same as status quo" and "worse than status quo" ranking added together the challenger wins to the status quo.

If I understand this correctly, you simply elect the "approval winner" (the winner on the top two slots) if he gets the first slot on more than half the ballots. Otherwise the status quo wins, who is not an ordinary candidate. That makes it unusual; I guess we could say that the status quo candidate is just privileged, so that neutrality is failed. In that case:

Majority: No, because the challenger might not have a majority of first-slot ratings, and only the challenger and the status quo candidate can win. It could be that another candidate has a majority.
FBC: No. It could be that you need to put your favorite candidate in the last slot to let a compromise candidate beat the status quo which you don't like. That is, your favorite might be the approval winner, but not have a majority of first-slot ratings.
Condorcet Loser: Seems so. Actually, I don't see what keeps the status quo from winning and being the Condorcet Loser.
Monotonicity: Yes.
Participation: No, because adding your ballot can alter how much a "majority" is. You could cause the status quo to win, for example, because you didn't put the challenger in the first slot.
Consistency: Not sure. Obviously if the same challenger wins in two districts, he'll win when they're combined. But is it possible for the status quo to win in two districts, but lose to a challenger when the districts are combined? Probably not. So I guess "yes." Kevin Venzke 22:02, 14 Jun 2005 (PDT)

Thinking about it a little more, and still assuming I have understood the method correctly, there is a big strategy problem: Voters who like the status quo should give most candidates the middle slot, to try to turn a weak candidate into the challenger. Kevin Venzke 22:50, 14 Jun 2005 (PDT)

Sorry that I didn't answer earlier, that's because I am paralyzed by embarrassment. You probably understand the method better than myself. I wrote "If you were asked to put down an approval list on all candidates you regard as better than the status quo, unlike with real Approval Voting the Favorite Betrayal criterion would be violated", a problem that only happens to people who favor the status quo. Then I go on and fix these peoples' problem by making it a problem for everybody else.

Actually, I don't understand the text you quote. If you were asked to do what? And how would it fail FBC? Kevin Venzke 16:16, 15 Jun 2005 (PDT)

Majority: I thought I could get away with the "first votes on more than 50% of all ballots" definition. It looks to me like the definition you give is also failed by MCA?

Er, what definition? About MCA, if you believe that a three-slot method can satisfy Majority (this is controversial), then MCA satisfies it, because if there is at least one candidate with first-slot ratings from more than half the voters, then one of these candidates always wins. MCA also satisfies FBC. Kevin Venzke 16:16, 15 Jun 2005 (PDT)

Is there any way to use the data of such a ballot that makes sure nobody worse than the status quo gets elected without violating FBC? To me, "Approval" in the name implies first of all passing FBC and that is more important than participation or consistency. I hope that can somehow be solved if the challenger status isn't restricted to one. How about if the challenger loses we pretend he never existed and look who is the challenger then etc? Wait a second, that looks better but could it pay to put a compromise above the fav if the fav is the status quo and has no chance?R.H. 12:10, 15 Jun 2005 (PDT)

I agree that FBC is more important than Participation, and I don't think consistency is important at all. In answer to your last question, I think it definitely could. Voters who like the status quo cannot sincerely use the top slot! But they may need to, to keep a worse challenger from winning.
I'm not sure I understand what you want to do. You want to satisfy FBC and get no one worse than the status quo based on three-slot pairwise contests. Is that right? I think it may be possible. I'll think about it. Kevin Venzke 16:16, 15 Jun 2005 (PDT)
Damn, you are right. That change only gives different situations in which FBC is violated. How about that: First we look if more than 50% of the voters filled the first column to even consider changing the status quo. Then we look who has most approval (first + second column). Only if that candidate has more than 50% approval, he wins. Otherwise the status quo is kept. That should satisfy FBC. Unlike with the former rules, ranking the challenger equal to the status quo is counted in favor of the challenger. That would also mean a very unlikely chance that the majority ends up with something it considers worse than the status quo (I would prefer a zero chance of that though).R.H. 03:21, 16 Jun 2005 (PDT)
Well, anyone who has 50%+ first-slot ratings will have 50% approval. You should probably give up on candidates who don't have 50%+ approval.
I thought about the problem of satisfying FBC while not electing a candidate beaten pairwise by the status quo candidate (SQ). I have something that satisfies the former and most of the latter. First, disqualify any candidate X for whom the following is true: The number of voters ranking X above SQ, plus the number of voters ranking X and SQ together in the top slot (i.e., the SQ is a placeable candidate) is strictly less than the number of voters ranking SQ above X. (SQ is never disqualified.) Among remaining candidates, elect the approval winner (first+second slot ratings). What do you think? Kevin Venzke 08:27, 16 Jun 2005 (PDT)
Looks almost like the best possible. Thanks! But doesn't that mean it's possible that even when less than half the voters show a ranking of anything else over the status quo, the status quo could be changed? That's what I meant with "First we look if more than 50% of the voters filled the first column to even consider changing the status quo." In other words: Only when more than half the voters think they prefer something (not the same candidate for all) to the status quo, the status quo is challenged. Could something like that be added to your idea? I will rewrite the article soon (and the next time I have a half-baked idea I put it on my discussion list before starting an article).R.H. 12:10, 16 Jun 2005 (PDT)

Almost. You could change, above, is strictly less than the number of voters ranking SQ above X to is less than a majority. This preserves FBC, and almost requires a majority sentiment. I guess you could replace a majority with 60%, but you'd still not have a guarantee that it will be a majority. (For FBC purposes you have to count first-slot ties with the SQ as votes for change.)

I don't think it's promising to check to see whether a majority want something more than the status quo. It would be very easy to regret expressing that interest, if the eventual result is worse for you than the status quo. Kevin Venzke 13:08, 16 Jun 2005 (PDT)

Well, consider an election uses the >Q, Q, Q> ballot and the counting rule is: Marks in the "better than status quo" and "as good as status quo" columns are added together and regarded as "approval marks". Candidates that receive more "worse than status quo" than "approval" marks are disqualified. If no candidate qualifies, the status quo is kept. Otherwise, among the remaining candidates the one with most "approval" wins. This so far satisfies FBC, right? Now add First we look if more than 50% of the voters filled the first column to even consider changing the status quo. to the beginning of the counting rules. This means the status quo is stronger protected and if it still satisfies FBC, I guess I will use it as the final definition.R.H. 14:10, 16 Jun 2005 (PDT)
I think this doesn't satisfy FBC, since if you like X better than the SQ, and put X in the top slot, this is a vote for change, no matter whether it's X who can win. Kevin Venzke 14:24, 16 Jun 2005 (PDT)
So if one's true preference is X>SQ>everybody else, but one thinks X has no chance, putting X in the top spot is counterproductive. But one could still vote X=SQ>everybody else.R.H. 15:46, 16 Jun 2005 (PDT)
Looks like the last version can as well be done with a simple approval ballot. Only if the SQ has less than 50% approval and only if the candidate with most approval has more than 50% the SQ gets changed and the one with most approval is the winner.R.H. 05:37, 17 Jun 2005 (PDT)

Only if the status quo gets less than 50% approval a change is considered. That's status quo bias and there is more of that. We then look if the candidate with most approval has more than 50% and only if that requirement is also met the SQ gets changed and the one with most approval is the winner.

I do believe this satisfies FBC. Kevin Venzke 11:34, 17 Jun 2005 (PDT)