Re: [EM] [CES #4429] Looking at Condorcet
On Thu, Feb 2, 2012 at 10:09 PM, Richard Fobes electionmeth...@votefair.org wrote: On 2/2/2012 11:07 AM, Kristofer Munsterhjelm wrote: On 02/02/2012 05:28 AM, Jameson Quinn wrote: I honestly think that honest rating is easier than honest ranking. ... As a contrast, to me, ranking is easier than rating. ... I too find ranking easier than rating. I go back and forth on this, myself. Some thoughts: - If I had to rank more than ten candidates, I think it would be difficult unless I put them into three or four tiers first. Then, perhaps I would choose to rank the candidates within the tiers or perhaps I would leave them all tied if I didn't really care that much. Thus, for me, honest rating with just a few buckets is more basic than ranking. - If someone built a computer program that presented me pairs of candidates at a time as Kristofer suggested, that would make it somewhat easier. I think I would still prefer to divide them into tiers first, but if I divided them into tiers first, I might not need the pairwise comparison hand-holding. Also, suppose that I analyzed the candidates in three different policy dimensions that I consider equally important and I found that my policy preferences were: Foreign Policy: ABC Domestic Social Issues: BCA Domestic Economic Issues: CAB Now I prefer A to B, B to C, and C to A. A cycle among my own personal preferences when I compare them pairwise. Then my output ranking would depend on the order in which the pairwise questions were asked. ??!? - If I were trying to cast an honest Approval Ballot, then I would think about each candidate separately and decide whether I approve them or not. - If I were trying to cast a strategic Approval Ballot or a fully strategic Score Voting Ballot, then I would first rank all the candidates, then decide where to put my cutoff. So I can definitely see the argument of those who think that ranking is more fundamental than even approval voting. - If I were trying to cast an honest Score Voting Ballot, I would have to feel like there was an objective meaning for the various scores. Then I could consider each candidate separately and give them my honest scores. I probably wouldn't even normalize. If I were going to normalize, then I might as well go fully-strategic and vote approval-style. - If I were casting an MJ ballot, I think I would consider each candidate separately and vote completely honestly, knowing that my vote was doing everything it could to help any candidate where my score was higher than society's median and, similarly, doing everything it could to hurt any candidate where my score was lower than society's median. I realize that my vote would not be fully strategic if there were two frontrunners and I liked both of them or disliked both of them, but in that situation, who cares? - If a real election were being tabulated with Condorcet, I would vote honestly. - If a real election were being tabulated with IRV, I would warn people not to vote for minor candidates. Let me admit that a crucial point for me is that the only way to gain Independence of Irrelevant Alternatives is to tell the voters to evaluate each candidate independently and vote honestly, which may make me biased towards rating methods. FBC is very important to me and I'm still skeptical of the FBC-compliant ranked-ballot methods recently proposed. ~ Andy Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
It sounds to me as if, of all the methods you mentioned, you would prefer MJ. How would you vote with SODA? (go ahead and think of your answer before you read mine) I think I'd almost always just delegate to my favorite with SODA. If I don't like my favorite's delegation order, that would make me reconsider whether they're really my favorite. If I decide they still are, I would consider whether I thought the difference between my preferred order and their predeclared preferences would matter. If I decide it does, then look for the best candidate I think has a chance, and vote for them and everyone better. Chances of me ever getting to that last step would around one in 10, I reckon. Jameson 2012/2/3 Andy Jennings electi...@jenningsstory.com On Thu, Feb 2, 2012 at 10:09 PM, Richard Fobes electionmeth...@votefair.org wrote: On 2/2/2012 11:07 AM, Kristofer Munsterhjelm wrote: On 02/02/2012 05:28 AM, Jameson Quinn wrote: I honestly think that honest rating is easier than honest ranking. ... As a contrast, to me, ranking is easier than rating. ... I too find ranking easier than rating. I go back and forth on this, myself. Some thoughts: - If I had to rank more than ten candidates, I think it would be difficult unless I put them into three or four tiers first. Then, perhaps I would choose to rank the candidates within the tiers or perhaps I would leave them all tied if I didn't really care that much. Thus, for me, honest rating with just a few buckets is more basic than ranking. - If someone built a computer program that presented me pairs of candidates at a time as Kristofer suggested, that would make it somewhat easier. I think I would still prefer to divide them into tiers first, but if I divided them into tiers first, I might not need the pairwise comparison hand-holding. Also, suppose that I analyzed the candidates in three different policy dimensions that I consider equally important and I found that my policy preferences were: Foreign Policy: ABC Domestic Social Issues: BCA Domestic Economic Issues: CAB Now I prefer A to B, B to C, and C to A. A cycle among my own personal preferences when I compare them pairwise. Then my output ranking would depend on the order in which the pairwise questions were asked. ??!? - If I were trying to cast an honest Approval Ballot, then I would think about each candidate separately and decide whether I approve them or not. - If I were trying to cast a strategic Approval Ballot or a fully strategic Score Voting Ballot, then I would first rank all the candidates, then decide where to put my cutoff. So I can definitely see the argument of those who think that ranking is more fundamental than even approval voting. - If I were trying to cast an honest Score Voting Ballot, I would have to feel like there was an objective meaning for the various scores. Then I could consider each candidate separately and give them my honest scores. I probably wouldn't even normalize. If I were going to normalize, then I might as well go fully-strategic and vote approval-style. - If I were casting an MJ ballot, I think I would consider each candidate separately and vote completely honestly, knowing that my vote was doing everything it could to help any candidate where my score was higher than society's median and, similarly, doing everything it could to hurt any candidate where my score was lower than society's median. I realize that my vote would not be fully strategic if there were two frontrunners and I liked both of them or disliked both of them, but in that situation, who cares? - If a real election were being tabulated with Condorcet, I would vote honestly. - If a real election were being tabulated with IRV, I would warn people not to vote for minor candidates. Let me admit that a crucial point for me is that the only way to gain Independence of Irrelevant Alternatives is to tell the voters to evaluate each candidate independently and vote honestly, which may make me biased towards rating methods. FBC is very important to me and I'm still skeptical of the FBC-compliant ranked-ballot methods recently proposed. ~ Andy Election-Methods mailing list - see http://electorama.com/em for list info Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
On 3.2.2012, at 21.45, Andy Jennings wrote: - If someone built a computer program that presented me pairs of candidates at a time as Kristofer suggested, that would make it somewhat easier. I think I would still prefer to divide them into tiers first, but if I divided them into tiers first, I might not need the pairwise comparison hand-holding. Also, suppose that I analyzed the candidates in three different policy dimensions that I consider equally important and I found that my policy preferences were: Foreign Policy: ABC Domestic Social Issues: BCA Domestic Economic Issues: CAB Now I prefer A to B, B to C, and C to A. A cycle among my own personal preferences when I compare them pairwise. Then my output ranking would depend on the order in which the pairwise questions were asked. ??!? A typical assumption is that the rankings of individuals are transitive. I think this is pretty much based on the assumption that people rate the candidates (unconsciously) anyway. Your three opinions were maybe 987, 987 and 963 in terms of ratings. That means that your transitive preference order is CAB, and you do have valid (unconscious) ratings for the candidates. The alternative approach would be that you indeed have cyclic preferences. But that maybe means only that you may behave strangely if you base your judgements on limited information. If you consider only foreign policy and social issues, you may end up saying BC, but if you had remembered to think also about economic issues, you would quickly change your statement. - If I were trying to cast an honest Score Voting Ballot, I would have to feel like there was an objective meaning for the various scores. Then I could consider each candidate separately and give them my honest scores. I probably wouldn't even normalize. If I were going to normalize, then I might as well go fully-strategic and vote approval-style. I agree. Already normalization is strategic. (Or maybe you have been explicitly requested to give min points to the worst candiate, and max points to the best, in which case you could sincerely cast a sincere (re)scaled vote.) - If a real election were being tabulated with Condorcet, I would vote honestly. I agree. That is a good default strategy. (Strategic voting doesn't really make sense unless some expert that the voter trusts tells him to vote in some certain way. And also in this case the expert may well be wrong.) - If a real election were being tabulated with IRV, I would warn people not to vote for minor candidates. Also this approach makes sense. This is however not a complete strategy yet. If we look at Burlington, maybe also some supporters of a major (top three) candidate should not have voted their favourite. (Sometimes voting for the minor candidates is harmless , and also a useful tool to market the minor candidates for some secondary reasons.) FBC is very important to me Could one say that Condorcet methods are FBC compliant enough so that you can recommend people not to betray their favourite 1) as the defaut rule if they are not told by experts to do otherwise or 2) as the default rule that is in practice valid in all lagrge elections, where voters make independent decisions on how to vote, and where their opinions are not fixed but can change all the time? Juho Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
On 02/03/2012 08:45 PM, Andy Jennings wrote: - If someone built a computer program that presented me pairs of candidates at a time as Kristofer suggested, that would make it somewhat easier. I think I would still prefer to divide them into tiers first, but if I divided them into tiers first, I might not need the pairwise comparison hand-holding. Also, suppose that I analyzed the candidates in three different policy dimensions that I consider equally important and I found that my policy preferences were: Foreign Policy: ABC Domestic Social Issues: BCA Domestic Economic Issues: CAB Now I prefer A to B, B to C, and C to A. A cycle among my own personal preferences when I compare them pairwise. Then my output ranking would depend on the order in which the pairwise questions were asked. ??!? You could look at a (single-winner) voting method as a stand-in for a deliberative process. A ranked voting method tries to find the best common ranking given the data it has to go on, which are the votes themselves. In the ideal case, you'd just have a deliberative process instead of the voting. There would be some back-and-forth and then you'd reach a consensus. The problem is that it doesn't scale. But if a voting method is a stand-in for a deliberative process, then it makes sense that each voter's preference would be transitive. The voters would already have gone through an internal deliberative process to arrive at a ranking of the candidates they are considering. So if I'm right about that, then the voter would already know his own consensus ranking based on the foreign vs social vs economic tradeoffs and the relative weights they have to him. In practice, things aren't that clean, but I think it works to show, intuitively, that people would have transitive rankings and so wouldn't encounter the internal cycle problem. If a voter's internal ranking is transitive, then you would only need to ask him X better than Y? n lg n times for n candidates*, where lg is the base-2 logarithm. If not, you would have to ask him n^2 times. Condorcet-type methods could handle both cases - in the latter, n^2 case, a pairwise method would incorporate intra-individual cycles by the exact same logic as it'd handle inter-individual cycles. (As I have said before, I have been thinking about using Condorcet methods for getting a ranking out of preference comparisons where the individual may have internal cycles because the set is so large. Ranking pictures is a simple example of that, as there may be so many pictures that the person looks at different things when comparing X to Y than when comparing Y to Z. However, doing n^2 comparisons grows very quickly and becomes quite tiresome. Some amount of preprocessing may speed it up - like your tiers or an Approval first stage where the person is generous with the approvals but excludes that which he considers obviously uninteresting.) * The simplest algorithm that achieves this bound is, in essence, an insertion sort that uses a binary search for each insertion. Its constant factor is better than say, quicksort, too, since all we care about is the number of comparisons, not the time it takes to insert. Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
Ranking more than ten candidates? Condorcet does NOT require such. However, if too many are running, you need to look for sanity: . You may have preferences among those most likely to win - pick those you see as the best few of these. . Also pick among the few you would prefer, regardless of their chances. This voting will help them get encouraging vote counts even if there is no chance of their winning. . Do not waste your energy on others. Now do your ranking among these, hopefully having time to rank properly according to desirability, not caring, for the moment, as to winnability. Dave Ketchum On Feb 3, 2012, at 2:45 PM, Andy Jennings wrote: On Thu, Feb 2, 2012 at 10:09 PM, Richard Fobes electionmeth...@votefair.org wrote: On 2/2/2012 11:07 AM, Kristofer Munsterhjelm wrote: On 02/02/2012 05:28 AM, Jameson Quinn wrote: I honestly think that honest rating is easier than honest ranking. ... As a contrast, to me, ranking is easier than rating. ... I too find ranking easier than rating. As do I. I go back and forth on this, myself. Some thoughts: - If I had to rank more than ten candidates, I think it would be difficult unless I put them into three or four tiers first. Then, perhaps I would choose to rank the candidates within the tiers or perhaps I would leave them all tied if I didn't really care that much. Thus, for me, honest rating with just a few buckets is more basic than ranking. - If someone built a computer program that presented me pairs of candidates at a time as Kristofer suggested, that would make it somewhat easier. I think I would still prefer to divide them into tiers first, but if I divided them into tiers first, I might not need the pairwise comparison hand-holding. Also, suppose that I analyzed the candidates in three different policy dimensions that I consider equally important and I found that my policy preferences were: Foreign Policy: ABC Domestic Social Issues: BCA Domestic Economic Issues: CAB Now I prefer A to B, B to C, and C to A. A cycle among my own personal preferences when I compare them pairwise. Then my output ranking would depend on the order in which the pairwise questions were asked. ??!? ... - If a real election were being tabulated with Condorcet, I would vote honestly. - If a real election were being tabulated with IRV, I would warn people not to vote for minor candidates. There is no harm in minor candidates getting the few votes they deserve in IRV. However, if the vote counters, as they work, see the deserving winner as momentarily having the fewest votes, this candidate will have lost. Let me admit that a crucial point for me is that the only way to gain Independence of Irrelevant Alternatives is to tell the voters to evaluate each candidate independently and vote honestly, which may make me biased towards rating methods. FBC is very important to me and I'm still skeptical of the FBC-compliant ranked-ballot methods recently proposed. ~ Andy Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
On Fri, Feb 3, 2012 at 2:05 PM, Jameson Quinn jameson.qu...@gmail.comwrote: How would you vote with SODA? I would usually end up delegating to my favorite. I'd look at their ranking and if it was pretty good I'd delegate. Otherwise, I'd probably come up with my own ranking (perhaps based on theirs) and then choose a cutoff and vote approval-style. So my strategy would be pretty similar to yours, I think. (go ahead and think of your answer before you read mine) I think I'd almost always just delegate to my favorite with SODA. If I don't like my favorite's delegation order, that would make me reconsider whether they're really my favorite. If I decide they still are, I would consider whether I thought the difference between my preferred order and their predeclared preferences would matter. If I decide it does, then look for the best candidate I think has a chance, and vote for them and everyone better. Chances of me ever getting to that last step would around one in 10, I reckon. Jameson 2012/2/3 Andy Jennings electi...@jenningsstory.com On Thu, Feb 2, 2012 at 10:09 PM, Richard Fobes electionmeth...@votefair.org wrote: On 2/2/2012 11:07 AM, Kristofer Munsterhjelm wrote: On 02/02/2012 05:28 AM, Jameson Quinn wrote: I honestly think that honest rating is easier than honest ranking. ... As a contrast, to me, ranking is easier than rating. ... I too find ranking easier than rating. I go back and forth on this, myself. Some thoughts: - If I had to rank more than ten candidates, I think it would be difficult unless I put them into three or four tiers first. Then, perhaps I would choose to rank the candidates within the tiers or perhaps I would leave them all tied if I didn't really care that much. Thus, for me, honest rating with just a few buckets is more basic than ranking. - If someone built a computer program that presented me pairs of candidates at a time as Kristofer suggested, that would make it somewhat easier. I think I would still prefer to divide them into tiers first, but if I divided them into tiers first, I might not need the pairwise comparison hand-holding. Also, suppose that I analyzed the candidates in three different policy dimensions that I consider equally important and I found that my policy preferences were: Foreign Policy: ABC Domestic Social Issues: BCA Domestic Economic Issues: CAB Now I prefer A to B, B to C, and C to A. A cycle among my own personal preferences when I compare them pairwise. Then my output ranking would depend on the order in which the pairwise questions were asked. ??!? - If I were trying to cast an honest Approval Ballot, then I would think about each candidate separately and decide whether I approve them or not. - If I were trying to cast a strategic Approval Ballot or a fully strategic Score Voting Ballot, then I would first rank all the candidates, then decide where to put my cutoff. So I can definitely see the argument of those who think that ranking is more fundamental than even approval voting. - If I were trying to cast an honest Score Voting Ballot, I would have to feel like there was an objective meaning for the various scores. Then I could consider each candidate separately and give them my honest scores. I probably wouldn't even normalize. If I were going to normalize, then I might as well go fully-strategic and vote approval-style. - If I were casting an MJ ballot, I think I would consider each candidate separately and vote completely honestly, knowing that my vote was doing everything it could to help any candidate where my score was higher than society's median and, similarly, doing everything it could to hurt any candidate where my score was lower than society's median. I realize that my vote would not be fully strategic if there were two frontrunners and I liked both of them or disliked both of them, but in that situation, who cares? - If a real election were being tabulated with Condorcet, I would vote honestly. - If a real election were being tabulated with IRV, I would warn people not to vote for minor candidates. Let me admit that a crucial point for me is that the only way to gain Independence of Irrelevant Alternatives is to tell the voters to evaluate each candidate independently and vote honestly, which may make me biased towards rating methods. FBC is very important to me and I'm still skeptical of the FBC-compliant ranked-ballot methods recently proposed. ~ Andy Election-Methods mailing list - see http://electorama.com/em for list info Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
I'm going to continue to take a devil's advocate anti-Condorcet position here. Of course I still believe that Condorcet systems are good overall, and much much better than plurality or IRV. But I honestly think that MJ and SODA are better. 2012/2/1 robert bristow-johnson r...@audioimagination.com On 2/1/12 11:28 PM, Jameson Quinn wrote: Dave gives good reasons for Condorcet. I'd like to present the other side. Condorcet systems have many advantages. So what's wrong with Condorcet? It comes in a bewildering array of forms, thus reducing the unity of its supporters. But that's not the real problem. It admits both betrayal and burial strategy, thus encouraging dangerous, negative-sum strategizing from its voters. And that could be significant. But I think that voters will realize that they will almost never have the information and unity to pull off a successful strategy, so that's not the real problem. It is complicated to understand, and impossible to easily visualize, how it works. i disagree with that. i spelled that out (how you would spell it out to the average voter) in my just previous post. Condorcet is *simple* to understand. unless there is a cycle, no one should be disputing the CW outcome. the weak-CW with few 1st-choice votes is not a strong case. you elect the CW, because of the inverse consideration. if you elect someone other than the CW (as we did in Burlington 2009), you are electing a candidate when *more* of us voters marked explicitly on our ballots that we preferred someone else. *not* merely someone else in general (the anybody but Jack vote), but we voter said specifically we want Jill instead. how can it be a democratic decision when more of us choose Jill elected to office over Jack than those who choose Jack over Jill, yet Jack is elected despite the mandate from the voters? i have never seen an adequate answer to that. in a simple 2-person race, even if the vote margin is close, even if by *one* vote, if more of us want Jill than those of us who want Jack, then Jill is elected. would you have it any other way? if you would not have it any other way, you cannot make a consistent argument against electing the CW if there is one. but we can argue about what to do about cycles. But that's not the real problem. As a ranked system, it is hopelessly caught in the contradictions of Arrow's theorem. sure, and that's the case for any system. Arrow's Impossibility theorem does not apply to rated systems. Gibbard-Satterthwaite does; and there are certainly things that are impossible, even for rated systems. but it's less of a problem than the demonstrated problems (not mere theoretical issues with Arrow) of either IRV (as demonstrated in Burlington 2009) or FPTP (myriad times when there are spoiler candidates, which might happen in Burlington in 1 month). the *only* problem (a la Arrow) that i see with Condorcet is the potential of a cycle. but it won't happen often I agree that an honest cycle will be quite rare. But I believe that strategic (false) cycles are a problem. More on that later. and only when the three top candidate all have roughly equal support. it's the same kind of problem as a tie, and you create rules to deal with that difficult situation in some kind of sense that makes sense (and i'm not saying that there is a clear winner in which cycle-resolving method makes the most sense, but since a cycle is even less likely to involve more than three, it's really a moot question). But that's not the real problem. Some voters will mistakenly imagine that it's Borda. But that's not the real problem. (They'll imagine that MJ is Range, too. I don't see how they'd significantly misapprehend SODA, though.) The real problem is that I think that people just don't want to do that much work to vote. Yes, I know, you can just vote approval-style if you want to, but most people would feel guilty about not really doing the whole job then. what people don't want to do is to agonize over how to vote to serve their political interest when there are multiple outcomes, only one of which is the voter's hearts desire. there often is another outcome which is tolerable and another that is intolerable. what should the voter do to be counted among those against the intolerable outcome, yet still support his/her sincere favorite candidate? Condorcet has at least as much of a problem with strategy as MJ, and more than SODA. The basic strategic issue with MJ is the chicken dilemma. And this strategic dilemma applies also to Condorcet: 35: AB 25: BA 40: C A wins 35: AB 14: BA 11: BC 40: C Now B wins. But if the A voters had used a similar defensive strategy, then the B voters' strategy would make C win. Yes, you can definitely argue that this is less of a slippery slope in Condorcet than it is under Approval or Range. But you can make that same kind of argument for MJ, and SODA completely
Re: [EM] [CES #4429] Looking at Condorcet
On Thu, Feb 2, 2012 at 3:22 AM, Dave Ketchum da...@clarityconnect.com wrote: Voter can vote as in: . FPTP, ranking the single candidate liked best, and treating all others as equally liked less or disliked. . Approval, ranking those equally liked best, and treating all others as equally liked less or disliked. . IRV, giving each voted for a different rank, with higher ranks for those liked best, and realizing that IRV vote counters would read only as many of the higher rankings as needed to make their decisions. . Condorcet, ranking the one or more liked, using higher ranks for those liked best, and ranking equally when more than one are liked equally. You can combine all of those methods (though not IRV) into a super-ballot. I think this was suggested on this list at some point. Basically, you give each candidate a rating, but fractional rankings are allowed. You then construct the condorcet matrix. If a voter ranks A as 1 and B as 1.5, then that counts as half a vote for A over B. However, if the voter votes A as 1 and B as 5, then that only counts as 1 vote for A over B, since each voter gets a maximum of 1 vote. Ranked candidates are considered preferred by a full vote over unranked. This allows the voters to decide which method to use. Condorcet - just rank the candidates in order of your choice, equals allowed Approval - rank approved candidates as 1 Range/Scorevoting - rank all candidates from 0 to 1 (0 = favorite) Each voter could decide, without one group having much more power than others. Abstains aren't handled that well. Scorevoting assumes that they should have no effect. In theory, the rule could be that if a candidate is not ranked, then no preference ordering is assumed. The ballot would have a zero for all comparisons relative to that candidate. However, that is a lot of hassle, maybe there could be a box to indicate how you want unranked candidates handled. Do you want them equal lowest rank, or abstain. Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
On 02/02/2012 05:28 AM, Jameson Quinn wrote: I honestly think that honest rating is easier than honest ranking. (How's that for honesty per square word?) MJ is the only system which allows honest rating to be full-strength in practice; and SODA is the only good system which allows anything easier. (And no, approval is not easier than MJ, because approval forces some amount of strategizing.) As a contrast, to me, ranking is easier than rating. When I'm set to rate, I tend to think about whether I rated the candidate just right or not - did I rate him too high, too low? - but if I rank, I don't have to care about that. All I have to do is get a general idea of the order of preference, and then ask do I like X better than Y or vice versa. Maybe I'm uncommon, but I thought I would say it. I've heard the claim that rating is easier than ranking before, and maybe it still is -- to most people. I'll also note that many of the ranked voting methods can be also be applied even if the only information you can get from the voters or the system is is X better than Y for pairs {X,Y}. Thus, these can be used to determine winners in actual one-on-one contests (e.g. chess matches, kittenwar-style preference elicitation) where it would be hard to use ratings. Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
On 2/2/12 2:07 PM, Kristofer Munsterhjelm wrote: On 02/02/2012 05:28 AM, Jameson Quinn wrote: I honestly think that honest rating is easier than honest ranking. (How's that for honesty per square word?) MJ is the only system which allows honest rating to be full-strength in practice; and SODA is the only good system which allows anything easier. (And no, approval is not easier than MJ, because approval forces some amount of strategizing.) As a contrast, to me, ranking is easier than rating. When I'm set to rate, I tend to think about whether I rated the candidate just right or not - did I rate him too high, too low? precisely. that's something that an Olympic judge needs to worry about, but not a voter. and then the other issue is, when we are in the voting booth, we are not just judges. we are *partisans*. suppose it's a Score ballot and two candidates. even if i think that both Candidates A and B are okay, but i decide i like A better, would you expect me to rate A a 10 and B a 9? NO! i will not attenuate my vote: A gets 10 and B gets 0. once i decide i like A better, i want to exercise my entire franchise to help A defeat B, even if i wouldn't be so disappointed if B was elected. and then, with 3 or more candidates, the tactical problem is: how much do you score your 2nd-choice given two competing goals? you don't want to help your 2nd choice beat your 1st choice, but you also *do* want to help your 2nd choice beat your last choice. oh me oh my, oh me oh my! what to do, what to do?!! Approval has the same problem. - but if I rank, I don't have to care about that. All I have to do is get a general idea of the order of preference, and then ask do I like X better than Y or vice versa. and that's all you have to worry about in a simple-majority, 2-candidate, one-person-one-vote election. except for that there are more candidates, it should be no different for multiple candidates. Maybe I'm uncommon, no. but I thought I would say it. I've heard the claim that rating is easier than ranking before, and maybe it still is -- to most people. i don't believe it. -- r b-j r...@audioimagination.com Imagination is more important than knowledge. Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
2012/2/2 Juho Laatu juho4...@yahoo.co.uk On 2.2.2012, at 21.07, Kristofer Munsterhjelm wrote: On 02/02/2012 05:28 AM, Jameson Quinn wrote: I honestly think that honest rating is easier than honest ranking. (How's that for honesty per square word?) MJ is the only system which allows honest rating to be full-strength in practice; and SODA is the only good system which allows anything easier. (And no, approval is not easier than MJ, because approval forces some amount of strategizing.) As a contrast, to me, ranking is easier than rating. When I'm set to rate, I tend to think about whether I rated the candidate just right or not - did I rate him too high, too low? - but if I rank, I don't have to care about that. All I have to do is get a general idea of the order of preference, and then ask do I like X better than Y or vice versa. Maybe I'm uncommon, but I thought I would say it. I've heard the claim that rating is easier than ranking before, and maybe it still is -- to most people. I'll also note that many of the ranked voting methods can be also be applied even if the only information you can get from the voters or the system is is X better than Y for pairs {X,Y}. Thus, these can be used to determine winners in actual one-on-one contests (e.g. chess matches, kittenwar-style preference elicitation) where it would be hard to use ratings. I agree that it is very difficult to claim that rating would be easier than ranking. Let's see what I can do. Attempt 1: It is difficult to write something like abc on the ballot paper, or to push buttons of the voting machine so that all the candidates will be in the correct order. Answer 1: Don't use such procedures. If you want to be sure that ranking at least as easy as rating, use same ballots as with rating. You can derive rankings from them. This is a perfectly satisfactory answer (as long as the election method does not reward dishonest strategy). But in my experience, it is used more to dismiss than to answer the question; and for that, it does not serve. Attempt 2: Methods that do not allow equal ranking can not use rating style ballots. Answer 2: Use better methods or use rating style ballots and split the vote in two parts (or use random order). Attempt 3: If there are very many candidates, it is easier and faster to rate them individually, one by one, rather than compare every candidate pairwise to others. Answer 3: You can do this with rankings too if you are not interested in determining the preference order of those candidates that are almost equally good. Fast rating is also inaccurate in the sense that one may give more points to A than B although A is worse than B. Again, this is much the same as answer 1, and my response is the same. Attempt 4: People have used numbers and ratings in schools. Answer 4: Think that you are still in the school and just rate the candidates (ratings will be derived from those ratings). Balinski and Laraki make much of the fact that Majority Judgment uses ratings which have independent, absolute meaning, rather than being solely a relative scale. I think there is something to this argument. All the arguments are actually based on the fact that rankings can be derived from ratings. In the case of rankings the voter need not care about the scale of numbers that one uses (1,2,3 is as good as 1,49,50). Juho Election-Methods mailing list - see http://electorama.com/em for list info Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
On 2/2/2012 11:07 AM, Kristofer Munsterhjelm wrote: On 02/02/2012 05:28 AM, Jameson Quinn wrote: I honestly think that honest rating is easier than honest ranking. ... As a contrast, to me, ranking is easier than rating. ... I too find ranking easier than rating. This seems to be a pattern. Those of us who prefer ranking tend to prefer Condorcet methods, and those who prefer rating tend to prefer Range or Majority Judgement. I will add that in a survey I prefer rating because I want to convey relative preferences that cross between the different questions. But in elections, I do not care about expressing preferences across the different contests. Richard Fobes Election-Methods mailing list - see http://electorama.com/em for list info
Re: [EM] [CES #4429] Looking at Condorcet
Dave gives good reasons for Condorcet. I'd like to present the other side. Condorcet systems have many advantages. So what's wrong with Condorcet? It comes in a bewildering array of forms, thus reducing the unity of its supporters. But that's not the real problem. It admits both betrayal and burial strategy, thus encouraging dangerous, negative-sum strategizing from its voters. And that could be significant. But I think that voters will realize that they will almost never have the information and unity to pull off a successful strategy, so that's not the real problem. It is complicated to understand, and impossible to easily visualize, how it works. But that's not the real problem. As a ranked system, it is hopelessly caught in the contradictions of Arrow's theorem. But that's not the real problem. Some voters will mistakenly imagine that it's Borda. But that's not the real problem. (They'll imagine that MJ is Range, too. I don't see how they'd significantly misapprehend SODA, though.) The real problem is that I think that people just don't want to do that much work to vote. Yes, I know, you can just vote approval-style if you want to, but most people would feel guilty about not really doing the whole job then. I honestly think that honest rating is easier than honest ranking. (How's that for honesty per square word?) MJ is the only system which allows honest rating to be full-strength in practice; and SODA is the only good system which allows anything easier. (And no, approval is not easier than MJ, because approval forces some amount of strategizing.) Most voters are lazy. And they'll resent any system which rubs their nose in that fact. Which Condorcet does. (SODA, on the other hand, brings lazy voters together, and gives their representative as much negotiating power as possible without diluting the winner's leadership mandate.) Jameson 2012/2/1 Dave Ketchum da...@clarityconnect.com Mike offers serious thinking about Approval. I step up to Condorcet as being better and nearly as simple for the voter. Voter can vote as in: . FPTP, ranking the single candidate liked best, and treating all others as equally liked less or disliked. . Approval, ranking those equally liked best, and treating all others as equally liked less or disliked. . IRV, giving each voted for a different rank, with higher ranks for those liked best, and realizing that IRV vote counters would read only as many of the higher rankings as needed to make their decisions. . Condorcet, ranking the one or more liked, using higher ranks for those liked best, and ranking equally when more than one are liked equally. Condorcet is little, if any, more difficult for voters than FPTP and Approval. . For many elections, voting as with them is good and as easy. .. When a voter likes A and B but prefers A - Approval cannot say this, but it is trivial to vote with Condorcet's ranking. In Condorcet the counters consider each pair of candidates as competing with each other. Usually one candidate, being best liked, proves this by winning in every one of its pairs. Unlike IRV (which requires going back to the ballots as part of the counting), counting here can be done in multiple batches of votes, and the data from the batches summed into one summary batch for analysis. There can be cycles in Condorcet, such as AB, BC, and CA, with these winning against all others. This requires a closer look to decide on the true winner, normally one of the cycle members. . Here the counters see the cycle, rather than a CW - and how to pick a winner from a cycle is a reason for the dispute as to what is best. Range/score ratings have their own way of showing more/less desire. Truly more power than Condorcet ranking - AND more difficult to decide on rating values to best interact with what other voters may do. Write-ins? Some would do away with such. I say they should be allowed for the cases in which something needs doing too late to attend to with normal nominations. True that voters may do some write-ins when there is no real need - and I have no sympathy for such voters - this needs thought. Dave Ketchum On Jan 28, 2012, at 3:13 PM, MIKE OSSIPOFF wrote Re: [EM] Propose plain Approval first. Option enhancements can be later proposals.: The enhancement consisting of voting options in an Approval election should only be mentioned when there’s plenty of time to talk, and when talking to someone who is patient or interested enough to hear that much. And the enhancements should only be mentioned as possibilities, when speaking to someone to whom the whole notion of voting-system reform is new. Maybe that goes for SODA as well. Don’t propose too much change, when talking to someone new to the subject. So the method to propose first is ordinary Approval. If, in some particular community, there is a committee of people interested in working on a voting-system
Re: [EM] [CES #4429] Looking at Condorcet
On 2/1/12 11:28 PM, Jameson Quinn wrote: Dave gives good reasons for Condorcet. I'd like to present the other side. Condorcet systems have many advantages. So what's wrong with Condorcet? It comes in a bewildering array of forms, thus reducing the unity of its supporters. But that's not the real problem. It admits both betrayal and burial strategy, thus encouraging dangerous, negative-sum strategizing from its voters. And that could be significant. But I think that voters will realize that they will almost never have the information and unity to pull off a successful strategy, so that's not the real problem. It is complicated to understand, and impossible to easily visualize, how it works. i disagree with that. i spelled that out (how you would spell it out to the average voter) in my just previous post. Condorcet is *simple* to understand. unless there is a cycle, no one should be disputing the CW outcome. the weak-CW with few 1st-choice votes is not a strong case. you elect the CW, because of the inverse consideration. if you elect someone other than the CW (as we did in Burlington 2009), you are electing a candidate when *more* of us voters marked explicitly on our ballots that we preferred someone else. *not* merely someone else in general (the anybody but Jack vote), but we voter said specifically we want Jill instead. how can it be a democratic decision when more of us choose Jill elected to office over Jack than those who choose Jack over Jill, yet Jack is elected despite the mandate from the voters? i have never seen an adequate answer to that. in a simple 2-person race, even if the vote margin is close, even if by *one* vote, if more of us want Jill than those of us who want Jack, then Jill is elected. would you have it any other way? if you would not have it any other way, you cannot make a consistent argument against electing the CW if there is one. but we can argue about what to do about cycles. But that's not the real problem. As a ranked system, it is hopelessly caught in the contradictions of Arrow's theorem. sure, and that's the case for any system. but it's less of a problem than the demonstrated problems (not mere theoretical issues with Arrow) of either IRV (as demonstrated in Burlington 2009) or FPTP (myriad times when there are spoiler candidates, which might happen in Burlington in 1 month). the *only* problem (a la Arrow) that i see with Condorcet is the potential of a cycle. but it won't happen often and only when the three top candidate all have roughly equal support. it's the same kind of problem as a tie, and you create rules to deal with that difficult situation in some kind of sense that makes sense (and i'm not saying that there is a clear winner in which cycle-resolving method makes the most sense, but since a cycle is even less likely to involve more than three, it's really a moot question). But that's not the real problem. Some voters will mistakenly imagine that it's Borda. But that's not the real problem. (They'll imagine that MJ is Range, too. I don't see how they'd significantly misapprehend SODA, though.) The real problem is that I think that people just don't want to do that much work to vote. Yes, I know, you can just vote approval-style if you want to, but most people would feel guilty about not really doing the whole job then. what people don't want to do is to agonize over how to vote to serve their political interest when there are multiple outcomes, only one of which is the voter's hearts desire. there often is another outcome which is tolerable and another that is intolerable. what should the voter do to be counted among those against the intolerable outcome, yet still support his/her sincere favorite candidate? that is the real problem, Jameson. i don't think feeling guilty is a problem. but voter regret is, especially after helping elect someone like George W Bush to office because one voted for Ralph Nader. that, in a nutshell, is the problem. I honestly think that honest rating is easier than honest ranking. (How's that for honesty per square word?) MJ is the only system which allows honest rating to be full-strength in practice; and SODA is the only good system which allows anything easier. (And no, approval is not easier than MJ, because approval forces some amount of strategizing.) Most voters are lazy. And they'll resent any system which rubs their nose in that fact. Which Condorcet does. i don't see that. it's no worse than IRV, which is as easy as 1-2-3! :-) ranking is not hard. Condorcet does not ask anything more from the voter, lazy or not, than does IRV. all it requires is for the voters to make up their minds about the candidates by Election Day. how is that an unreasonable expectation of voters? (SODA, on the other hand, brings lazy voters together, and gives their representative as much negotiating power as
Re: [EM] [CES #4429] Looking at Condorcet
On 2.2.2012, at 6.28, Jameson Quinn wrote: Dave gives good reasons for Condorcet. I'd like to present the other side. Condorcet systems have many advantages. So what's wrong with Condorcet? It comes in a bewildering array of forms, thus reducing the unity of its supporters. But that's not the real problem. It admits both betrayal and burial strategy, thus encouraging dangerous, negative-sum strategizing from its voters. And that could be significant. But I think that voters will realize that they will almost never have the information and unity to pull off a successful strategy, so that's not the real problem. It is complicated to understand, and impossible to easily visualize, how it works. But that's not the real problem. Visualization is not really a problem if the method is simple and straight forward enough. If the method measures the number of required additional votes to beat all others, then a simple histogram can be used to show how far each canidate is from that position (or how far ahead the CW is). Juho P.S. If you want more information, maybe multiple columns to show distance of one candidate to all other candidates could be useful somewhere. P.P.S. Debian visualizes their Condorcet results in a more complex way that may be cryptic to people who don't understand the method fully. But their figures at least have lots of information and they are visually interesting. There is also some interesting additional information in the form of a None Of The Above box. The Debian approach is a bit complicated, but at least interesting. http://www.debian.org/vote/2010/vote_001 As a ranked system, it is hopelessly caught in the contradictions of Arrow's theorem. But that's not the real problem. Some voters will mistakenly imagine that it's Borda. But that's not the real problem. (They'll imagine that MJ is Range, too. I don't see how they'd significantly misapprehend SODA, though.) The real problem is that I think that people just don't want to do that much work to vote. Yes, I know, you can just vote approval-style if you want to, but most people would feel guilty about not really doing the whole job then. I honestly think that honest rating is easier than honest ranking. (How's that for honesty per square word?) MJ is the only system which allows honest rating to be full-strength in practice; and SODA is the only good system which allows anything easier. (And no, approval is not easier than MJ, because approval forces some amount of strategizing.) Most voters are lazy. And they'll resent any system which rubs their nose in that fact. Which Condorcet does. (SODA, on the other hand, brings lazy voters together, and gives their representative as much negotiating power as possible without diluting the winner's leadership mandate.) Jameson 2012/2/1 Dave Ketchum da...@clarityconnect.com Mike offers serious thinking about Approval. I step up to Condorcet as being better and nearly as simple for the voter. Voter can vote as in: . FPTP, ranking the single candidate liked best, and treating all others as equally liked less or disliked. . Approval, ranking those equally liked best, and treating all others as equally liked less or disliked. . IRV, giving each voted for a different rank, with higher ranks for those liked best, and realizing that IRV vote counters would read only as many of the higher rankings as needed to make their decisions. . Condorcet, ranking the one or more liked, using higher ranks for those liked best, and ranking equally when more than one are liked equally. Condorcet is little, if any, more difficult for voters than FPTP and Approval. . For many elections, voting as with them is good and as easy. .. When a voter likes A and B but prefers A - Approval cannot say this, but it is trivial to vote with Condorcet's ranking. In Condorcet the counters consider each pair of candidates as competing with each other. Usually one candidate, being best liked, proves this by winning in every one of its pairs. Unlike IRV (which requires going back to the ballots as part of the counting), counting here can be done in multiple batches of votes, and the data from the batches summed into one summary batch for analysis. There can be cycles in Condorcet, such as AB, BC, and CA, with these winning against all others. This requires a closer look to decide on the true winner, normally one of the cycle members. . Here the counters see the cycle, rather than a CW - and how to pick a winner from a cycle is a reason for the dispute as to what is best. Range/score ratings have their own way of showing more/less desire. Truly more power than Condorcet ranking - AND more difficult to decide on rating values to best interact with what other voters may do. Write-ins? Some would do away with such. I say they should be allowed for the cases