The Toy Single-winner System Laid Out on 2024-03-07 UTC
-
Collect Score-style ballots on the range {100, 99, 90, 50, 10, 1, 0}. Unmarked candidates count as zero.
My rationale for including 99 and 90 is that I think they can be effective for providing the appropriate amount of boost for a lesser evil, to ease people away from concern about candidates winning because of money support or fame. Think Nader and Gore, for example. Or Nader and Bush if you are a Republican. Maybe eventually, people will learn that evil should not be supported, even when it is lesser. But I expect this lesson to take time to sink in.
A whole class of systems could be described as using rounds of tallying wherein the goal of each round is to pinpoint and eliminate from further consideration, the "worst" candidate. The systems would differ in opinion as to the best criteria for determining "worst" (as well as obviously on the type of ballot). A drawback of probably all such systems is that they would generate a need to gather the ballots and tally them in one place, which is untenable with respect (or disrespect) to security. That is why this is a toy idea and there are grounds to avoid proposing it for use in political elections with thousands of voters.
If there is a Condorcet loser (Llull loser?), i. e. a candidate who is beaten pairwise in terms of preferences by every other candidate still in the running, that is the worst candidate.
Otherwise:
- construct a set of virtual votes by making a copy from the actual votes or from the previous set of virtual votes from a prior round, if any has been calculated.
- For the new set of virtual votes, remove the scorings of candidates who have already been eliminated from consideration by prior rounds.
- Linearly spread each virtual vote so that it uses the entire range (if it has no spread, it is exhausted and you can drop it). Note that this can result in virtual ballots rating candidates with just about any rational number from 0 to 100 inclusive, not just the numbers permitted for the real votes.
Then eliminate from consideration, the candidate having the lowest total score when the modified virtual votes are added up.
Last candidate standing wins -|
-
An evil doesn’t deserve 90% support. He deserves 0% support.
Approval easily allows that.
-
@michaelossipoff said in The Toy Single-winner System Laid Out on 2024-03-07 UTC:
An evil doesn’t deserve 90% support. He deserves 0% support.
That's true; however, depending, it may be justified for some voters to extend that support for strategic reasons.
Approval easily allows that.
Fact.
-
@jack-waugh said in The Toy Single-winner System Laid Out on 2024-03-07 UTC:
@michaelossipoff said in The Toy Single-winner System Laid Out on 2024-03-07 UTC:
An evil doesn’t deserve 90% support. He deserves 0% support.
That's true; however, depending, it may be justified for some voters to extend that >support for strategic reasons.
Of course strategy could be a reason for anything, & it depends on 1) One’s valuations; & 2) One’s probability-assessments.
- Valuations:
It’s so sad, what some people regard as acceptable. There was a book called, I’ve Been Down So Long It Seems Like Up To Me.
How well that describes lesser-evil voting !!
Even strategy can’t negate principle.
- Probability Assessment:
This is where people go wrong, BIGTIME.
We’re told that two unliked or even despised corrupt criminals, with the unpopular unwanted dismal same-old-same-old policies are “The Two Choices”.
…&, astoundingly, people believe it.
It’s a sucker-game. The old Good-Cop, Bad-Cop routine.
Hundreds of millions of suckers.
P.T. Barnum wasn’t kidding when he said that there’s a sucker born every minute.
-
I described above, a system with multiple rounds of tallying, like IRV. It is probably impractical to propose for office in government because in the first place, it is rather complex, and in the second place, it probably can't be tallied without bringing the entire set of ballots to a central location for the tally, which as Sass (and quite likely other writers as well) has pointed out, causes an unacceptably severe security risk to the integrity of the count. However, in private organizations, where there is less reason for suspicion, it could be practical. I described it to invite comment on whether it would make choosing a voting strategy or tactic more or less difficult as compared to Score, and/or how well it would do on whatever other criteria and constraints might seem reasonable to appeal to in connection with quality of a voting experience, outside of the integrity issue.
In response to your remarks about principle, I want to step away, at least temporarily, from the multiround tallying aspect of the above description. To discuss how principle applies, let's compare just these two systems to each other:
- Score{100, 0};
- Score{100, 99, 90, 50, 10, 1, 0}.
Now you argue that in system 1, the lesser evil deserves no support from any voter, and so the voters have, according to you, no grounds to use any of the score options that system 1 provides that system 0 does not provide. In other words, they should vote in Approval style, and never provide any support to evil.
Now I need to bring in the Prisoner's Dilemma (PD). This is a thought experiment in Game Theory and the way it is usually described, there are two prisoners, and their captor, who sets up a system of rewards and punishments such that the options of each prisoner when viewed from a Game Theoretic perspective clearly teach that she or he will do best to defect. The logic in support of such a conclusion consists of a two-by-two table that shows that no matter what prisoner B does, prisoner A is better off defecting than cooperating. The situation is symmetric and so the same argument would apply to prisoner B's evaluation of his choices.
But if we take a broader perspective on the PD setup, and evaluate the collective values of the possible collective outcomes, we see that the two prisoners are better off if both cooperate than if both defect.
Let's extend the thought experiment from just two prisoners to 300 million prisoners. I believe that the rewards and punishments can be set up so that if all the prisoners cooperate, they will reap grand rewards, but for each individual prisoner, the Game-Theoretic evaluation will indicate that it is wisest to defect.
Are you going to lecture the 300 million prisoners on what principle tells them to do, and expect them to follow principle? While the PD is in effect, it will continue to win.
Choose-one Plurality subjects the voters to the PD.
Perhaps you will say that Approval breaks the PD. In principle, it does. But voters accustomed to Choose-one Plurality may likely be worried about their voting power with respect to the greater-evil/lesser-evil portion of the contest, under conditions in which they doubt whether there is sufficient resolve among their fellow voters to assure that a moral candidate wins and all the evil candidates are defeated. I want to offer the 99, 90 options as a ladder out of the cesspool. A voter who has that concern can exercise power to do both: give full support to moral candidates, creating the possibility that with sufficient proportion of the other voters doing the same, the victory will come, but at the same time, exercising some influence against the greater-evil winning if there isn't sufficient resolve in the electorate as a whole to achieve total victory. So, I'm suggesting that system 1 as compared to system 0 will work a little better in the context of the sort of psychological motivations that tend to be well described by Game Theory, to break the adhesion between on the one hand, the muck, and on the other hand, the feet and calves of the voters.
-
@jack-waugh said in The Toy Single-winner System Laid Out on 2024-03-07 UTC:
I described above, a system with multiple rounds of tallying, like IRV. It is probably impractical to propose for office in government because in the first place, it is rather complex, and in the second place, it probably can't be tallied without bringing the entire set of ballots to a central location for the tally, which as Sass (and quite likely other writers as well) has pointed out, causes an unacceptably severe security risk to the integrity of the count. However, in private organizations, where there is less reason for suspicion, it could be practical.
Yes, though I propose Approval in public political elections, the disadvantages of ranked methods & Score don’t apply or manifest in polls. So I didn’t object to RCV in Tobin’s Free & Equal polls (though Condorcet would be much better), & Condorcet is my favorite for the current poll here, & there’s no reason to not do Score balloting too.
You brought up cooperation/defection dilemma to support a contention that there’s reason to support evils.
But what sort of cooperation would there be with an evil?
“Elect me & I won’t be as bad as Greater Evil” ?
No good. Reject his deal, because you can get something not evil at all, & needn’t bargain for a lesser one.
So, given that there’s no conceivable cooperation quantum evil, then there’s no way that a cooperation/defection dilemma could justify supporting a lesser evil.
[quote{
I described it to invite comment on whether it would make choosing a voting strategy or tactic more or less difficult as compared to Score, and/or how well it would do on whatever other criteria and constraints might seem reasonable to appeal to in connection with quality of a voting experience, outside of the integrity issue.In response to your remarks about principle, I want to step away, at least temporarily, from the multiround tallying aspect of the above description. To discuss how principle applies, let's compare just these two systems to each other:
- Score{100, 0};
- Score{100, 99, 90, 50, 10, 1, 0}.
Now you argue that in system 1, the lesser evil deserves no support from any voter, and so the voters have, according to you, no grounds to use any of the score options that system 1 provides that system 0 does not provide. In other words, they should vote in Approval style, and never provide any support to evil.
Now I need to bring in the Prisoner's Dilemma (PD). This is a thought experiment in Game Theory and the way it is usually described, there are two prisoners, and their captor, who sets up a system of rewards and punishments such that the options of each prisoner when viewed from a Game Theoretic perspective clearly teach that she or he will do best to defect. The logic in support of such a conclusion consists of a two-by-two table that shows that no matter what prisoner B does, prisoner A is better off defecting than cooperating. The situation is symmetric and so the same argument would apply to prisoner B's evaluation of his choices.
But if we take a broader perspective on the PD setup, and evaluate the collective values of the possible collective outcomes, we see that the two prisoners are better off if both cooperate than if both defect.
Let's extend the thought experiment from just two prisoners to 300 million prisoners. I believe that the rewards and punishments can be set up so that if all the prisoners cooperate, they will reap grand rewards, but for each individual prisoner, the Game-Theoretic evaluation will indicate that it is wisest to defect.
Are you going to lecture the 300 million prisoners on what principle tells them to do, and expect them to follow principle? While the PD is in effect, it will continue to win.
Choose-one Plurality subjects the voters to the PD.
Perhaps you will say that Approval breaks the PD. In principle, it does. But voters accustomed to Choose-one Plurality may likely be worried about their voting power with respect to the greater-evil/lesser-evil portion of the contest, under conditions in which they doubt whether there is sufficient resolve among their fellow voters to assure that a moral candidate wins and all the evil candidates are defeated. I want to offer the 99, 90 options as a ladder out of the cesspool. A voter who has that concern can exercise power to do both: give full support to moral candidates, creating the possibility that with sufficient proportion of the other voters doing the same, the victory will come, but at the same time, exercising some influence against the greater-evil winning if there isn't sufficient resolve in the electorate as a whole to achieve total victory. So, I'm suggesting that system 1 as compared to system 0 will work a little better in the context of the sort of psychological motivations that tend to be well described by Game Theory, to break the adhesion between on the one hand, the muck, and on the other hand, the feet and calves of the voters.
-
“Quantum evil” ?? Oh the joys of autocorrect !!
What I’d written was “cooperation with evil”.
Who knows why autocorrect relplaced that with “quantum evil.”
Disregard all of the obvious autocorrect errors.
-
@michaelossipoff said in The Toy Single-winner System Laid Out on 2024-03-07 UTC:
you can get something not evil at all
How can we convince the prisoners that they can get something not evil at all, when they are so accustomed to a system that denies them that possibility, and that does so by playing them off against each other?
-
The Chicken Dilemma is similar to the Prisoners’ Dilemma, but lesser-evil sucker-giveaway isn’t the Prisoners’ Dilemma.
In fact it’s no genuine dilemma. It’s just a matter of believing a hilariously preposterous media story or not. Two despicable despised evil aren’t the 2 choices. Do you really believe that most other people want an evil?
You don’t want split-vote? Then don’t split the vote by voting dishonestly for someone you have to hold your nose to vote for.
Vote honestly & find out how things could really be.
Please don’t be a lesser-evil giveaway sucker.
-
Plurality has a spli- vote problem, which is why we need a Progressive Primary. But, lacking that, the Green Party & their nominee are the obvious natural combining point for Progressive votes
…& any forum that includes voting strategy doesn’t need to avoid voting-strategy specifics.