[EM] Combating the Approval Burr dilemma

[Abd ul-Rahman Lomax]

At 06:13 PM 11/16/2006, Stephane Rouillon wrote:
>I definitively disagree with Abd, the Burr dilemna or any measure of the
>number of starategy
>a voter has to consider should be the main criteria to evaluate electoral
>systems.

The issue was about an agreement between candidates. I was not 
claiming that party affiliation should be irrelevant in all methods, 
but we are considering methods which provide no means of indicating 
party affiliation when voting. It is irrelevant to the method.

Now, if the addition of one candidate is likely to take votes away 
from another candidate, perhaps of the same party, or otherwise, we 
have a spoiler effect. Some methods are vulnerable to this and some 
are not. Approval does not need to have this problem, single-winner.

This thread was titled "the Approval Burr dilemma," and the example 
given presumes single-winner.

The alleged agreement between candidates, I'm claiming, is dicta, 
fluff, has nothing to do with how the method works. What candidates 
*can* do is to recommend that their supporters also vote for another 
candidate. They have fulfilled an agreement to do this when they make 
the recommendation, and that is a public thing. They have no control 
over how their supporters actually vote, in a secret ballot system.

Where is the dilemma? More precisely, who is faced with it?

The individual voter can simply vote for all candidates whom the 
voter considers acceptable, that is a basic Approval strategy. If 
being of the same party makes a candidate acceptable, fine. 
Presumably the recommendation of one's favorite candidate might have 
some effect.... But if a voter would rather see a candidate outside 
the party win, over seeing this less favored candidate win .... then 
the voter presumably is not going to vote for the less favored 
candidate from the party. Where is the dilemma?

Approval Voting does require some knowledge of the milieu, at least 
if one is going to use the frontrunner strategy. (Vote for your 
favorite among the frontrunners, plus any candidate you favor more than this.)

This strategy could easily lead you to not vote for the less favored 
candidate. And this is how the system is designed to work.

Approval is not the be-all and end-all of election methods. It is an 
astonishingly good improvement considering the cost, all that it 
takes is Count All the Votes. It is a step toward Range Voting, but 
IRV with overvoting allowed could also be interesting, and dumping 
the no-overvote rule would not be prejudicial against IRV.

But those who are stuck on preference as King, I can understand, are 
not going to like Approval. I just think that preference makes a poor 
king, one not so likely to unite the people behind a government. I 
much prefer Majority Consent, with supermajority consent being even 
better. And this is what Approval allows.