The difficulty of arguing

[William Waugh]

It would be easier to argue from a constraint than from a criterion, if there were a valid one to argue from.  For me, the goal is to convince people to advocate for a change, for single-winner cases, to one or more voting systems impervious to vote splitting.  Suppose Frohnmayer's (ongoing) contention were correct, for example, that for a system to be impervious to vote splitting, it is necessary and sufficient that it provide balance.  Then I wouldn't have to advocate for a particular voting system (with say a specific range) but I could just talk about how important invulnerability to vote splitting is, and say that the balance test suffices to check for it, and kind of lead people to the class of correct answers without having to be in a fight about which to prefer among the adequate.

However, Clay Shentrup argues in his recent talk that we need to look at the Bayesian Regret figures, and that anything else, even the balance constraint, is not looking at performance where it counts.  He compares this to looking at particular racecar characteristics rather than just racing the cars.  So, you experts don't seem to agree with Frohnmayer about the "necessary and sufficient".  But what does that leave me with to use in arguing, just a criterion, Bayesian Regret.  I can't get people on board for a voting system that will defeat the two-party system by saying "look at the Bayesian Regret figures and choose a voting system that rates higher on that criterion than most of the other voting systems do, to be sure that you are choosing a system that is impervious to vote splitting."  Listeners would not judge that I have a clear understanding what I am trying to convince them of.

What is the most effective kind of argument to make, to get people to spread an advocacy that every State in the US needs, for single-winner elections, a voting system belonging to a class such that you and I can feel assured, with valid grounds for feeling that way, that every voting system belonging to that class will be invulnerable to vote splitting?

Is it just, "the experts have looked carefully at a number of voting systems (for single-winner elections) and the only ones they find grounds to be sure of, as being impervious to vote splitting, are the variants of Score?"

[Rob Wilson]

Perhaps we could brand score voting variants as “accurate representation” since it most accurately measures the will of the voters, but that can seem a little ambiguous.

[Clay Shentrup]

I feel your pain William. It's too esoteric to explain to even very smart analytical people.

[William Waugh]

So is it true that a system is invulnerable to vote splitting if and only if it accords the voters equal power as determined by the balance constraint?

[Jameson Quinn]

Majority Judgment (and other modern Bucklin systems) are a counterexample. The obey only a weakened balance constraint; that is, given a set of pre-existing votes and a vote A, there is always a vote B which cancels vote A out; but it is not necessarily clear how to cancel out vote A if you don't know how others will be voting. Nonetheless, these systems resist vote-splitting at least as well as Score or Approval in theory, and perhaps a hair better in practice.

Condorcet(winning-votes) systems are similar; that is, nearly but not exactly "balanced" in Frohnmayer's sense, and resistant to vote splitting.

In general, rather than focusing on vote splitting only, I like to consider three scenarios. Vote-splitting; center squeeze; and chicken dilemma.

2015-04-17 6:23 GMT-04:00 William Waugh <2knuw...@snkmail.com>:

So is it true that a system is invulnerable to vote splitting if and only if it accords the voters equal power as determined by the balance constraint?

--
You received this message because you are subscribed to the Google Groups "The Center for Election Science" group.
To unsubscribe from this group and stop receiving emails from it, send an email to electionscien...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[William Waugh]

Does Range Voting meet all three constraints?

On Friday, April 17, 2015 at 9:24:14 AM UTC-4, Jameson Quinn wrote:

...

In general, rather than focusing on vote splitting only, I like to consider three scenarios. Vote-splitting; center squeeze; and chicken dilemma.

...

[William Waugh]

I have been trying to explain how advertising money can leverage unfair voting systems to prevent the general populace (those with little advertising money) from having significant political power.  The path of thought I have been chasing would suppose that the Prisoner's Dilemma (PD) is key to explaining the mechanism of exclusion. I suppose the voter wants to exercise what little power she has instead of none, similar to how the prisoner in the PD wants to serve less time rather than more. As the effects of the prisoner's choice are determined also by what the other prisoner does, the effects of the choice made by a single voter are determined by what the other voters do.  But when I try to construct the matrix of rewards and punishments that determines the voter's behavior under Plurality if the predictive power of Game Theory is to be believed in (which I think it is in the case of voting), I get stuck. In the PD, part of the problem is that the prisoners can't communicate. This would be worsened if there were 2 million prisoners instead of just two. I want to propose a mechanism of communication, as a temporary measure, to operate outside the official elections, to allow the electorate to inform itself about itself. Voters so armed with that knowledge could then cooperate in the plurality election, and get State legislators who will enact Score Voting. Can anyone fill the gap in my argument, and explain how it is that lack of knowledge of the other voters' ratings of the candidates prompts voters to vote for highly advertised candidates? Does it make sense to make the analogy with the PD?