Re: [EM] Declaration of Election-Method Experts and Enthusiasts: final stretch
Noise, but possibly worth a response. In writing about a Condorcet race the standard format seems to be AXY. For voting the ballot format seems to be to be able to assign rank numbers to as many of the candidates as the voter chooses. In reporting election results the n*n matrix has findable values for each pair of candidates. Robert calls the format he has seen for the matrix silly, and suggests another format. The reporting is a human readable copy of what is being computed - with the computing almost certainly done by computer if many candidates. Therefore a reporting format such as Robert's would be usable if humans could agree - or even have selectable choices of formats if enough desire. Dave Ketchum On Sep 7, 2011, at 1:12 AM, robert bristow-johnson wrote: still not sure of the efficacy of trying to persuade voters (or their elected representatives) to try out different ballot formats than ranked choice but... ... The n*n matrix used in Condorcet has information useful to those wanting to learn more about relationship of candidates. ... why, oh why, are all of you election method experts stuck on that silly n x n matrix geometry (where the main diagonal has no information you have to associate one number on the lower left with another number on the upper right, and it isn't obvious which number goes with which candidate) instead of grouping the pairwise totals *in* *pairs*??? like A 56 B 44 A 88 B 65 C 12 C 35 A 90 B 82 C 55 D 10 D 18 D 45 THAT format is where you have useful information about the relationships between candidates at a glance. if we're gonna tell people about Condorcet, why are we putting it in a stupid rectangular array where it is difficult to tell who beat who? it only makes it harder to sell this to skeptics. -- r b-j r...@audioimagination.com Imagination is more important than knowledge. 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] Declaration of Election-Method Experts and Enthusiasts: final stretch
Ok, where do I sign up? On 2011-09-05 23:13, Dave Ketchum wrote: I finally got around to a bit. I see both Judgment and Judgement - can one be a typo? Declaration of Election-Method Experts and Enthusiasts Contents When there is a list of items, some taking more than one line, something, such as indentation, should show start of each item. I see Enthusiasts here - Should also go with Experts below. Introduction It is time to change our voting system. We, the undersigned election-method experts and enthusiasts from around the world, unanimously denounce the use of plurality voting in elections in which there are more than two candidates. In this declaration we offer several ready-to-adopt replacement election methods that we agree will reliably produce much fairer results. Proper question is whether there MAY be more than two candidates: . There will never be more than two - so election method does not matter. . When there are more, voters can wish to vote against the worst by voting for more than one - impossible with plurality. . We cannot be bothered with this need - how bad this is depends on value of the election. Part of selling against plurality: . Wherever current experience is that runoffs are rarely needed and there is very little voting for other than the two main candidates, deciders may feel that there is no need for preparing for what has never happened to them. . Even with that normality, there can be times when voting for others happens in significant numbers. We need to alert deciders that this can happen in any district and this is what needs preparing for even if they are used to things staying simpler. Better ballots With better information from the voters, we can find better winners. Approval gives nothing but ability to vote for more than one. All the others provide for voters indicating which of the candidates they vote for are also their most preferred. Also, while Condorcet ranking unconditionally says that higher ranks are better than lower, there is nothing requiring or permitting saying how much higher. The other methods, depending on statements as to how much higher a ranked candidate may be, require that the voter indicate magnitude in the vote. Fairer counting methods Condorcet: . It is an approach to a tie that CAN result in those leading candidates needing some extra analysis to decide on a winner. . The n*n matrix used in Condorcet has information useful to those wanting to learn more about relationship of candidates. There are three Condorcet methods that identify the Condorcet winner (when there is one) without explicitly looking for the Condorcet winner, and they are, in alphabetical order: I claim that, if there is one, the CW should be found and, at our distance, we do not need to check on how the method goes about that. Even if there is no CW, the n*n matrix used to look for the CW is the obvious source for deciding on a winner - which points toward using n*n for this analysis. I have not chased down the innards of using IRV here, but wonder if, as used here, it is immune to the problems that afflicted IRV in Burlington. Anyway, I ask that IRV discussion stay out of the Condorcet discussion - seems like there were, earlier, better words about IRV than I see here. Also, seems like SODA should be kept away from Condorcet. In Using the fairer methods in organizations Private organizations are a great place to start voting reform. One particularly relevant example of a “private” election is the nomination process of a political party. It is true that our supported methods make this process less important, because, unlike plurality, they do not break down when more than one candidate from a party is running. Still, we expect that many parties would still want to have a formal nomination (“primary election”) process so as to focus their efforts on one or two candidates per office. We believe that any party using a superior voting system internally will see immediate benefits. A primary process with increased turnout, with fewer negative attacks, and with a more-democratic result will result in a stronger nominee who is better-prepared to win in the general election. This presumably is true in some states. In New York parties do not do elections. Primaries, done by government for the parties, handle both primary elections AND electing party officers. 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] Declaration of Election-Method Experts and Enthusiasts: final stretch
still not sure of the efficacy of trying to persuade voters (or their elected representatives) to try out different ballot formats than ranked choice but... ... The n*n matrix used in Condorcet has information useful to those wanting to learn more about relationship of candidates. ... why, oh why, are all of you election method experts stuck on that silly n x n matrix geometry (where the main diagonal has no information you have to associate one number on the lower left with another number on the upper right, and it isn't obvious which number goes with which candidate) instead of grouping the pairwise totals *in* *pairs*??? like A 56 B 44 A 88 B 65 C 12 C 35 A 90 B 82 C 55 D 10 D 18 D 45 THAT format is where you have useful information about the relationships between candidates at a glance. if we're gonna tell people about Condorcet, why are we putting it in a stupid rectangular array where it is difficult to tell who beat who? it only makes it harder to sell this to skeptics. -- 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] Declaration of Election-Method Experts and Enthusiasts: final stretch
I finally got around to a bit. I see both Judgment and Judgement - can one be a typo? Declaration of Election-Method Experts and Enthusiasts Contents When there is a list of items, some taking more than one line, something, such as indentation, should show start of each item. I see Enthusiasts here - Should also go with Experts below. Introduction It is time to change our voting system. We, the undersigned election-method experts and enthusiasts from around the world, unanimously denounce the use of plurality voting in elections in which there are more than two candidates. In this declaration we offer several ready-to-adopt replacement election methods that we agree will reliably produce much fairer results. Proper question is whether there MAY be more than two candidates: . There will never be more than two - so election method does not matter. . When there are more, voters can wish to vote against the worst by voting for more than one - impossible with plurality. . We cannot be bothered with this need - how bad this is depends on value of the election. Part of selling against plurality: . Wherever current experience is that runoffs are rarely needed and there is very little voting for other than the two main candidates, deciders may feel that there is no need for preparing for what has never happened to them. . Even with that normality, there can be times when voting for others happens in significant numbers. We need to alert deciders that this can happen in any district and this is what needs preparing for even if they are used to things staying simpler. Better ballots With better information from the voters, we can find better winners. Approval gives nothing but ability to vote for more than one. All the others provide for voters indicating which of the candidates they vote for are also their most preferred. Also, while Condorcet ranking unconditionally says that higher ranks are better than lower, there is nothing requiring or permitting saying how much higher. The other methods, depending on statements as to how much higher a ranked candidate may be, require that the voter indicate magnitude in the vote. Fairer counting methods Condorcet: . It is an approach to a tie that CAN result in those leading candidates needing some extra analysis to decide on a winner. . The n*n matrix used in Condorcet has information useful to those wanting to learn more about relationship of candidates. There are three Condorcet methods that identify the Condorcet winner (when there is one) without explicitly looking for the Condorcet winner, and they are, in alphabetical order: I claim that, if there is one, the CW should be found and, at our distance, we do not need to check on how the method goes about that. Even if there is no CW, the n*n matrix used to look for the CW is the obvious source for deciding on a winner - which points toward using n*n for this analysis. I have not chased down the innards of using IRV here, but wonder if, as used here, it is immune to the problems that afflicted IRV in Burlington. Anyway, I ask that IRV discussion stay out of the Condorcet discussion - seems like there were, earlier, better words about IRV than I see here. Also, seems like SODA should be kept away from Condorcet. In Using the fairer methods in organizations Private organizations are a great place to start voting reform. One particularly relevant example of a “private” election is the nomination process of a political party. It is true that our supported methods make this process less important, because, unlike plurality, they do not break down when more than one candidate from a party is running. Still, we expect that many parties would still want to have a formal nomination (“primary election”) process so as to focus their efforts on one or two candidates per office. We believe that any party using a superior voting system internally will see immediate benefits. A primary process with increased turnout, with fewer negative attacks, and with a more-democratic result will result in a stronger nominee who is better-prepared to win in the general election. This presumably is true in some states. In New York parties do not do elections. Primaries, done by government for the parties, handle both primary elections AND electing party officers. Election-Methods mailing list - see http://electorama.com/em for list info