[EM] To Condorcetists:

[Juho Laatu]

On 19.5.2012, at 7.25, Michael Ossipoff wrote:

> You continue:
> 
> I mean that there could be need for further reforms. 
> 
> [endquote]
> 
> You like to speculate. Speculations aren't really answerable. To what needs
> are you referring, in particular? 

One key topic was the already discussed possible use of coalition governments instead of the single 50+% party governmnets of today. It wouldn't be anymore a pure "president's government".

> ... Certainly advocates of rank methods would want to propose them ...

Those questions are mainly not related to the Approval vs. Condorcet question but to the Plurality vs. compromise seeking single-winner method question.

> You said:
> .
> The problems of Approval with three or more potential winners might irritate
> people and change their voting behaviour. 
> 
> [endquote]
> 
> You like to refer to "problems" without specifying them.

As I said in my previous mail. The Approval problems that I refer to are the well known and well discussed problems of the Approval method. They are quite separate from the "impact on the two-party system" oriented questions (where Approval and Condorcet behave in quite similar way).

> In particular, you have never answered my question about what problem
> Approval will have that Plurality doesn't have.

Thanks, this is one concrete request that I can reply to. I tried to address also that question earlier but obviously my explanations were not good enough.

If we study Approval and Plurality as separate single-winner methods, then maybe you aleady know all the discussions, maybe even too well. My opinion is that the biggest problem of Approval is the difficulty of voters to find a working strategy when there are more than two poential winners (e.g. when there are two candidates from one wing and one from the other). Also Plurality has related problems with strategy, but in Plurality (in a two-party system) a good strategy is to vote for one of the two dominant parties. In Approval the voter may not have any such safe strategy option.

If we study Approval and Plurality as part of the proposed system where single-winner methods are used in single-member districts in a multi-winner election, then the answer is quite different. Then the key difference is that while Plurality is a key component that maintains the two-party structure, Approval aims at electing compromise winners that need not come from the two dominant parties. This may be good or bad, depending on one's point of view. Approval introduces a new non-classical and untested system, but that system may well be worth a try. Approval (and other compromise seeking methods) will have some specific features like the already discussed influences on the government stucture.

> No one can predict exactly what the changes would be.
> But they'd definitely be improvements.

Hmm. This sounds like the current system would be the worst of all possible systems. In that case all changes would of course be good.

> You said:
> 
> (or return back as in Burlington).
> 
> [endquote]
> 
> I get tired of asking you why people would want to go back to Plurality.
> This discussion isn't productive.

In Burlington people, or possibly only politicians, wanted to go back to the old system. It is obvious that at least within the current dominant parties there is some interst to maintain their current powerful position. Regular people may be less interested in going back. But I think going back to the old system is a risk (or why not sometimes a positive option) in every reform.

> I didn't observe any strong burial tendency in Burlington when I analyzed
> those votes. 
> 
> [endquot]
> 
> It was a municipal election.
> 
> Did your analysis include looking at the labels on the ballots that told
> what the voter really likes best?  Oh way, the ballots don't have such
> labels, do they :-)
> 
> As someone who is actually here, I've observed a strong favorite-burial
> tendency in a limited sample. But even in that sample, the consistency
> suggests that it won't be rare.

The Burlington votes are available and the election is a relatively large and certainly competitive political election. If there is a general tendency to bury, that tendecy should be visible in those votes, and there should be a large set of votes that have ranked minor candidates above some of the (three?) most potential candidates.

> You know what, I'm still only halfway through this post. Do I really have to
> wade through the rest of it? Not if the rest is anything like what I've been
> replying to so far. Maybe I'll resume later, but, if so,  I'll reply only to
> a few things that are relatively deserving of reply.

No need to comment all the lines. Short replies are better than long ones. I generally try to limit myself to few essential points + explicit questions and requests to me (+ correcting misunderstandings of what I said, if any). I also regularly read through my drafts to eliminate repetitive points and points with minor value to the discussion.

> You said:
> 
>> parties chosen by corporate media and big money contributors
> 
> I'd say this problem is for the most part independent of the used election
> method.
> 
> [endquote]
> 
> You'd say that because you aren't paying attention. Big corruption,
> disgusting to all,  here can keep winning only because of a Myerson-Weber
> equilibrium peculiar to Plurality. 
> 
> Two unliked  parties can keep winning forever, because the results (victory
> for one of them) seems to confirm people's belief that they're the only
> winnable parties. With Plurality, it goes on and on, self-confirming. 
> 
> No that is not independent of the election method. It depends on Plurality.

I agree that a two-party system has some properties that add risk of continuing poliices that people do not like. Relative certainlty of being back in power after one or two terms in opposition is one such factor. However multi-party systems are not free of this problem either. Problems that are related to "parties chosen by corporate media and big money contributors" can be addressed also e.g. by changing the financing structure of the political system (i.e. also by means that are not related to the used election method).

> You asked:
> 
> Could you point out a concrete example [of favorite-burial being optimal
> strategy in Condorcet] with sincere preferences and strategic votes, so I
> can check if it is likely to work in a real life election. (The one example
> that I analyzed in the other mail didn't seem to be a threat in real life
> elections.)
> 
> [endquote]
> 
> For one thing, the other posting didn't have a numerical example of that.
> 
> For another thing, the matter has been much discussed on EM, and it is well
> established that Condorcet fails FBC. 
> 
> It was established here that if you want to maximally help Compromise win
> instead of Worse, then you should rank Compromise alone at top.
> 
> And if Compromise is the only acceptable deemed likely to beat Worse, and
> Worse is an unacceptable, then you ranking Compromise alone at top is
> optimal.
> 
> If you want a numerical example, I'll post one when I get a chance, within
> the next few days probably.

I already commented one of your numerical examples in http://lists.electorama.com/pipermail/election-methods-electorama.com/2012-May/030400.html. But additional ones are welcome, especially concerning the optimality of the favorite-burial strategy.

Do you agree that with the given numbers the strategy that I commented is not really a viable strategy? Maybe with some other better chosen numbers?

Juho



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20120519/b770a8bf/attachment-0004.htm>