Give A 9. Give B 5. Everyone else gets zero. Or if you think A and B have no risk of losing, give B 0 so your A v B vote counts more. If you think both candidates are going to lose to another, then vote B 9 so that it stands more of a chance.
In the instant runoff, you only have one option: A>B
As for encouraging a 2nd place candidate, I think I need to do a bit more research. I forget why, but perhaps I was thinking of another voting system.
There are reasons to vote other than 0 or 9. Say you prefer A>B>C. Give B a rating between 0 and 9 so that if A doesn't win, B will. Also, don't make it too close to A because if C loses to both A and B, you prefer A.
Very good point. I had to think about this example for a long time. In this case, the only way that range voting can work is with accurate polling.
Obama supporters will try to get Obama over Hillary, but they will also fear McCain and accept Hillary as a compromise. They will watch the polls to see the lowest at which they can rate Hillary such that McCain still loses. Hillary supporters do the same. However they don't do exactly the same. Each has a perception of the risk involved in possibly allowing McCain to win. Say Obama supporters think this is a greater risk than Hillary supporters. Then they will give more padding between Hillary v McCain at the expense of Obama v Hillary. Hillary will win because her supporters are more moderate and more willing to accept the risk of McCain winning.
At first glance, it seems odd that the more moderate supporters would be more polarized against a spoiler. I will have to think about this.
Perhaps we would still need parties to sort out this spoiler effect.
Specifically, looking at the polls, they will see
-Hillary 29%
-Obama 31%
-McCain 40%
Then they will adjust to
-Hillary 45%
-Obama 41%
-McCain 40%
because Obama supporters see more risk in McCain and would like more padding for Hillary v McCain.
Q. How do different voters coordinate their votes so that the democrats each get enough to overcome McCain?
A. They can vote in proportion to how much padding.
Q. What if they would like to vote a 2.4 but must choose between a 2 and 3?
A. Roll a 20-sided die, if the result is 8 or less, vote 3. If the result is 8 or more, vote 2.
As a side note, this strategy could be used for range voting with a range of 0 or 1. This is approval voting and could be done with checkboxes, which is simpler than numbers. But I don't think it is simple to roll dice, so maybe most people won't like this.
Also, I'm not sure if range voting would lead to the situation you described. I realize it is only an example, but there's the possibility that we won't come across this situation. For example, there might be a moderate candidate that all of the McCain, Obama, and Hillary supporters also support and who will get more votes because his supporters are less afraid of right and left sides and will put more of their vote towards him.
You have uncovered the mystical secret of the ancient pixie sticks. The chemical composition of which turns a citizen into a docile servant. Two sticks have been added to your satchel.
I say a voter should maximize his vote. This leads to a situation where people who don't like a popular candidate will vote for the most popular person who is not that candidate. This happens in our current elections. It is uncertain at the beginning which candidate we should fear will win and who stands the best chance at dethroning him. Polls are extremely influential in this first formation of consensus. The exploit is at this step, when people are trying to figure out what their neighbors are thinking.
I think we don't need to have this mob-mentality guessing game. We could all vote more sincerely by allowing more of a preference to be recorded.
So I think you should suggest a way that the system can be gamed.
Perhaps you are thinking of Arrows impossibility theorem which applies only to ranked voting systems. Range voting is not one of those. It allows ties and weightings. Arrows theorem does not apply.
Perhaps you could have each party use range voting during the primaries, and I am not sure how this would affect things.
We do need more than two parties, because otherwise voters only need to decide which one or two issues is most important, and debates will center on these few big things while leaving out smaller things. For instance, many people just vote on the issue of "Life v Choice" (TM).
I don't think the B supporters would do that. Every voter will maximize the influence of his own vote. I say the following is the smart way for each of the supporters to vote:
A gets 9 from his 10 supporters, who give 0 to all other candidates.
B gets 9 from his 10 supporters, but they give 0 to all other candidates.
C gets 9 from his 10 supporters, who give 5 to B and only 0 to A.
A gets 90
B gets 140
C gets 90
B has the support of C-voters and is the only candidate to cross party lines, so he wins.
Each voter has the power to maximize the influence of his own vote. It is an equal power.
Maryland, Illinois, Hawaii, and New Jersey have signed legislation to give their votes to the national popular vote. The legislation goes into effect when more than half of the electoral votes come from states that have this legislation.
http://www.nationalpopularvote.com/
Contact your state representative.
I think STV encourages candidates to endorse less popular parties. This way they put more distance between themselves and their competitors. A better system is range voting, which does not do this. It also allows more preference to be recorded. In short, you rate each candidate 0 to 9.
This reference does not consider non-ranked systems. Yes, all ranked systems fail Arrow's theorem, but there exist other systems. One such system is range voting. Give each candidate a 0 to 9 rating. This gives more ability to show preference than any ranked system. The candidates ratings are totaled and the highest total wins.
Because there is only one round of eliminations, there aren't weird effects where a candidate will endorse a less popular party to increase his chances of winning. Also, there isn't any disagreement about how to count preference.
A simpler way to do range voting would be a range of 0 to 1. This is approval voting and can be done on current ballots.
You are a smart one. "One man-one vote" ensures that to prevent one candidate from winning, a voter will vote for the most popular of the other candidates. There will always be two frontrunners, and the method with which they are chosen will be mob rule. Indeed, no third candidate can win. I don't feel any power when I cast my vote.
There is a chance that the two candidates will become too moderate and similar. This is because advancing on the middle ground gets the candidate converts. Then, when there is not much difference to the voter, he will be able to choose between the two and one third-party candidate. However, it is late in the election cycle and the third candidate doesn't have much chance.
Perhaps you already know this. In that case, what do you think is the best alternative voting system? I think approval voting is better than what we have now and is also better than IRV. I would enjoy the debate.
I challenge your IRV. Instant Runoff is a multi-stage voting system and so falls victim to strategic voting in sinister ways. A strong candidate will support a weaker candidate to draw votes from his more equal competitors. How do you refute this?
Also, have you considered range voting or approval voting. Approval voting is just an easier to implement version of range voting. Check wikipedia for descriptions of each. Basically, range voting is rating each candidate on a scale of 1 to 10 or some other range. Approval voting is a coarser version of range, with just a 0 or 1 for each candidate. I argue in favor of this because it is a cardinal system rather than ordinal (IRV is ordinal). All ordinal systems suffer from Arrow's impossibility theorem, although some argue this doesn't matter. Also, range voting shows the least Bayesian regret, as seen on wiki.
Ha, insightful... Baloney. Are they really wasting their votes? Admit it. They really do want to vote as they do. It may seem obvious that nobody really chose these two knuckleheads as the best future for our democracy-corporation, but you can't expect it to work that way. Once we have two frontrunners, they're hard to un-stick. Try to bring up a third, and there's a threat to whichever frontrunner they most resemble. People don't want to take the chance of letting down their defenses to who they view as the greater of two evils, and they settle for the lesser of the two evils. That's it. Game over. The third party needs more political wrangling to overcome this barrier.
P.S. Only if there really isn't much difference between the two big parties will there be a chance for a third to seem like a good gamble. Unfortunately, abortion rears its ugly head and becomes one of the only issues Americans really ever get to weigh in on.
Do farmers use their waste to fertilize their new crops? Isn't that what compost heaps are for? Wouldn't the farmers then need to buy petroleum-based fertilizers to supplement their shortage of compostable waste? Any farmers out there?
It isn't easy to start a political party. I would also like to meet the genius who implements this. I generally have confidence in people, but the dead vote is another issue.
What do slashdotters think about other methods of reform? For example, Olympic range voting (hot-or-not style) would prevent the two-party system from reigning supreme. Then, something like this freegovernment.org movement could take off. (Leaving us with a one-party system)
Why don't we have more parties? That's just how one vote per office works. It's a game, and there are other options, such as one vote per candidate - just a thumbs up or thumbs down. This wouldn't end corruption in one fell swoop, but it would allow new ideas from small parties to compete fairly. In the 70's lots of research was done on the game theory of voting. Range voting seems least susceptible to strategies and most sincere, but might be harder to implement than approval voting (thumbs up or thumbs down)
Not in a sincere way. We all vote strategically.
It's a problem that has a solution. Approval voting. Look to the Olympics to see people really voting for the best suspended ring champion, ribbon dancer, figure skater, etc. Rank your candidates on the hot-or-not scale, and I do believe that we will be able to vote sincerely.
It's hard to do so with all the other issues and only two people to choose from. I think approval voting would change the red v blue game, and some people are coming out on local levels to support alternative voting systems. Recently in Champaign County, home to the University of Illinois, there was a proposed referendum on instant runoff voting. While I oppose it, it might be better than our current one-vote-per-office system. Even better would be approval voting, which has one vote per candidate. Most people that I've talked to haven't thought of a different way to vote. They only know that there's electronic, paper, etc. Instant runoff voting has a big drawback with a strategy called "vote this guy for #2", but at least someone out there knows we aren't voting in the most sincere way.
It's only a two party system because the game is designed for it.
People do actually vote for who they want to vote for. They vote strategically. Nobody wants the greater of two evils, so they vote for the lesser of two evils. To have people vote sincerely, we need a different voting game.
We ought to look toward the Olympics this year and observe a voting system that really works. There is no spoiler effect. There aren't any parties. Each candidate gets one vote.
It might be a little complicated to count votes, so instead of a range of ratings from 0 to 10, voters could either give a 1 or 0. They would either approve or disapprove.
People could still vote for the lesser of two evil frontrunners, but they could also approve of another candidate that might have a wide appeal. If this candidate had more than 50% approval, he would win. Have a third party or independent candidate that you think a majority would like? Then vote your approval. The game is changed.
Slashdot Mirror
You say A>B>everyone.
Give A 9. Give B 5. Everyone else gets zero. Or if you think A and B have no risk of losing, give B 0 so your A v B vote counts more. If you think both candidates are going to lose to another, then vote B 9 so that it stands more of a chance.
In the instant runoff, you only have one option: A>B
As for encouraging a 2nd place candidate, I think I need to do a bit more research. I forget why, but perhaps I was thinking of another voting system.
Why vote 6 when you can vote 9?
There are reasons to vote other than 0 or 9. Say you prefer A>B>C. Give B a rating between 0 and 9 so that if A doesn't win, B will. Also, don't make it too close to A because if C loses to both A and B, you prefer A.
Another example from another person who asked:
http://games.slashdot.org/comments.pl?sid=969061&cid=25081019
Very good point. I had to think about this example for a long time. In this case, the only way that range voting can work is with accurate polling.
Obama supporters will try to get Obama over Hillary, but they will also fear McCain and accept Hillary as a compromise. They will watch the polls to see the lowest at which they can rate Hillary such that McCain still loses. Hillary supporters do the same. However they don't do exactly the same. Each has a perception of the risk involved in possibly allowing McCain to win. Say Obama supporters think this is a greater risk than Hillary supporters. Then they will give more padding between Hillary v McCain at the expense of Obama v Hillary. Hillary will win because her supporters are more moderate and more willing to accept the risk of McCain winning.
At first glance, it seems odd that the more moderate supporters would be more polarized against a spoiler. I will have to think about this.
Perhaps we would still need parties to sort out this spoiler effect.
Specifically, looking at the polls, they will see
-Hillary 29%
-Obama 31%
-McCain 40%
Then they will adjust to
-Hillary 45%
-Obama 41%
-McCain 40%
because Obama supporters see more risk in McCain and would like more padding for Hillary v McCain.
Q. How do different voters coordinate their votes so that the democrats each get enough to overcome McCain?
A. They can vote in proportion to how much padding.
Q. What if they would like to vote a 2.4 but must choose between a 2 and 3?
A. Roll a 20-sided die, if the result is 8 or less, vote 3. If the result is 8 or more, vote 2.
As a side note, this strategy could be used for range voting with a range of 0 or 1. This is approval voting and could be done with checkboxes, which is simpler than numbers. But I don't think it is simple to roll dice, so maybe most people won't like this.
Also, I'm not sure if range voting would lead to the situation you described. I realize it is only an example, but there's the possibility that we won't come across this situation. For example, there might be a moderate candidate that all of the McCain, Obama, and Hillary supporters also support and who will get more votes because his supporters are less afraid of right and left sides and will put more of their vote towards him.
You have uncovered the mystical secret of the ancient pixie sticks. The chemical composition of which turns a citizen into a docile servant. Two sticks have been added to your satchel.
I say a voter should maximize his vote. This leads to a situation where people who don't like a popular candidate will vote for the most popular person who is not that candidate. This happens in our current elections. It is uncertain at the beginning which candidate we should fear will win and who stands the best chance at dethroning him. Polls are extremely influential in this first formation of consensus. The exploit is at this step, when people are trying to figure out what their neighbors are thinking.
I think we don't need to have this mob-mentality guessing game. We could all vote more sincerely by allowing more of a preference to be recorded.
So I think you should suggest a way that the system can be gamed.
Perhaps you are thinking of Arrows impossibility theorem which applies only to ranked voting systems. Range voting is not one of those. It allows ties and weightings. Arrows theorem does not apply.
Perhaps you could have each party use range voting during the primaries, and I am not sure how this would affect things.
We do need more than two parties, because otherwise voters only need to decide which one or two issues is most important, and debates will center on these few big things while leaving out smaller things. For instance, many people just vote on the issue of "Life v Choice" (TM).
A gets 9 from his 10 supporters, who give 0 to all other candidates.
B gets 9 from his 10 supporters, but they give 0 to all other candidates.
C gets 9 from his 10 supporters, who give 5 to B and only 0 to A.
A gets 90
B gets 140
C gets 90
B has the support of C-voters and is the only candidate to cross party lines, so he wins.
Each voter has the power to maximize the influence of his own vote. It is an equal power.
Well, to be fair, Italian cops will take bribes. Where did the mafia come from?
Maryland, Illinois, Hawaii, and New Jersey have signed legislation to give their votes to the national popular vote. The legislation goes into effect when more than half of the electoral votes come from states that have this legislation. http://www.nationalpopularvote.com/ Contact your state representative.
better off to have range voting and prevent unproportional promotion of less popular parties with IRV.
Superior to the ones you have mentioned is range voting. It allows the voter to record even more preference, as it allows ties and weightings.
I think STV encourages candidates to endorse less popular parties. This way they put more distance between themselves and their competitors. A better system is range voting, which does not do this. It also allows more preference to be recorded. In short, you rate each candidate 0 to 9.
This reference does not consider non-ranked systems. Yes, all ranked systems fail Arrow's theorem, but there exist other systems. One such system is range voting. Give each candidate a 0 to 9 rating. This gives more ability to show preference than any ranked system. The candidates ratings are totaled and the highest total wins. Because there is only one round of eliminations, there aren't weird effects where a candidate will endorse a less popular party to increase his chances of winning. Also, there isn't any disagreement about how to count preference. A simpler way to do range voting would be a range of 0 to 1. This is approval voting and can be done on current ballots.
one vote per candidate would be a simple enough solution
You are a smart one. "One man-one vote" ensures that to prevent one candidate from winning, a voter will vote for the most popular of the other candidates. There will always be two frontrunners, and the method with which they are chosen will be mob rule. Indeed, no third candidate can win. I don't feel any power when I cast my vote.
There is a chance that the two candidates will become too moderate and similar. This is because advancing on the middle ground gets the candidate converts. Then, when there is not much difference to the voter, he will be able to choose between the two and one third-party candidate. However, it is late in the election cycle and the third candidate doesn't have much chance.
Perhaps you already know this. In that case, what do you think is the best alternative voting system? I think approval voting is better than what we have now and is also better than IRV. I would enjoy the debate.
I challenge your IRV. Instant Runoff is a multi-stage voting system and so falls victim to strategic voting in sinister ways. A strong candidate will support a weaker candidate to draw votes from his more equal competitors. How do you refute this?
Also, have you considered range voting or approval voting. Approval voting is just an easier to implement version of range voting. Check wikipedia for descriptions of each. Basically, range voting is rating each candidate on a scale of 1 to 10 or some other range. Approval voting is a coarser version of range, with just a 0 or 1 for each candidate. I argue in favor of this because it is a cardinal system rather than ordinal (IRV is ordinal). All ordinal systems suffer from Arrow's impossibility theorem, although some argue this doesn't matter. Also, range voting shows the least Bayesian regret, as seen on wiki.
Ha, insightful... Baloney. Are they really wasting their votes? Admit it. They really do want to vote as they do. It may seem obvious that nobody really chose these two knuckleheads as the best future for our democracy-corporation, but you can't expect it to work that way. Once we have two frontrunners, they're hard to un-stick. Try to bring up a third, and there's a threat to whichever frontrunner they most resemble. People don't want to take the chance of letting down their defenses to who they view as the greater of two evils, and they settle for the lesser of the two evils. That's it. Game over. The third party needs more political wrangling to overcome this barrier.
P.S. Only if there really isn't much difference between the two big parties will there be a chance for a third to seem like a good gamble. Unfortunately, abortion rears its ugly head and becomes one of the only issues Americans really ever get to weigh in on.
Ah yes, using the cellular network neutrally. Does it matter whether its voice or data or SMS? No. That is great.
Do farmers use their waste to fertilize their new crops? Isn't that what compost heaps are for? Wouldn't the farmers then need to buy petroleum-based fertilizers to supplement their shortage of compostable waste? Any farmers out there?
It isn't easy to start a political party. I would also like to meet the genius who implements this. I generally have confidence in people, but the dead vote is another issue.
What do slashdotters think about other methods of reform? For example, Olympic range voting (hot-or-not style) would prevent the two-party system from reigning supreme. Then, something like this freegovernment.org movement could take off. (Leaving us with a one-party system)
Why don't we have more parties? That's just how one vote per office works. It's a game, and there are other options, such as one vote per candidate - just a thumbs up or thumbs down. This wouldn't end corruption in one fell swoop, but it would allow new ideas from small parties to compete fairly. In the 70's lots of research was done on the game theory of voting. Range voting seems least susceptible to strategies and most sincere, but might be harder to implement than approval voting (thumbs up or thumbs down)
Not in a sincere way. We all vote strategically. It's a problem that has a solution. Approval voting. Look to the Olympics to see people really voting for the best suspended ring champion, ribbon dancer, figure skater, etc. Rank your candidates on the hot-or-not scale, and I do believe that we will be able to vote sincerely.
It's hard to do so with all the other issues and only two people to choose from. I think approval voting would change the red v blue game, and some people are coming out on local levels to support alternative voting systems. Recently in Champaign County, home to the University of Illinois, there was a proposed referendum on instant runoff voting. While I oppose it, it might be better than our current one-vote-per-office system. Even better would be approval voting, which has one vote per candidate. Most people that I've talked to haven't thought of a different way to vote. They only know that there's electronic, paper, etc. Instant runoff voting has a big drawback with a strategy called "vote this guy for #2", but at least someone out there knows we aren't voting in the most sincere way.
It's only a two party system because the game is designed for it. People do actually vote for who they want to vote for. They vote strategically. Nobody wants the greater of two evils, so they vote for the lesser of two evils. To have people vote sincerely, we need a different voting game. We ought to look toward the Olympics this year and observe a voting system that really works. There is no spoiler effect. There aren't any parties. Each candidate gets one vote. It might be a little complicated to count votes, so instead of a range of ratings from 0 to 10, voters could either give a 1 or 0. They would either approve or disapprove. People could still vote for the lesser of two evil frontrunners, but they could also approve of another candidate that might have a wide appeal. If this candidate had more than 50% approval, he would win. Have a third party or independent candidate that you think a majority would like? Then vote your approval. The game is changed.