[EM] Juho: Social Optimizations. The Sincere Ideal.

[Juho Laatu]

On 15.11.2012, at 18.00, Michael Ossipoff wrote:

> If I ranked all of the candidates sincerely, the Democrat and the
> Republican would be at the bottom of that ranking. Even if they're
> winnable.
> 
> So you can't say that not ranking unwinnable candidates allows you to
> vote a short ranking.

I said that not ranking unwinnable candidates does not cause very much harm. In your case I assume that your most preferred candiates can not win. Based on that you could leave them unranked, but you should rank at least one of the two winnable candidates.

That apprach will not change the result of the election (with high probability). This (focus on who wins) is the way we usually measure the performance of election methods and study the recommended voting practices for the voters. There could however be other resons why it would make sense to rank your favourite candidates. You could rank them at top to help them (or their party) to win at the next election. Or you could give them this way some encouragement and thumbs up, and you could increase their chances of becoming elected in some other important position. In order to get good information on the true preferences of the electorate (for statistics and studies) it would make sense for all voters to always rank as many candidates as possible. You may also have your personal reasons, like the feeling of ranking one of the frontrunners one but last (not first).

> Sure, there's a case for saying that people would enjoy indicating who
> is worst. I just don't think that the Democrats and Republicans
> deserve to be ranked at all. Not ranking them at all is better than
> dignifying them with rank positions--even last and 2nd-to-last.

That sounds like you are talking about implicit approval of all the ranked candidates. I prefer to see rankings as rankings, i.e. truncated vote A>B means A>B>C=D=E, not A>B>>>C=D=E, if we talk about traditional Condorcet methods that usually treat truncations that way. But I guess you are talking about methods that intentionally want to use implicit approval.

I believe many people would be happy to tell who is worst. But it is not a good idea to allow them to vote A>B>all_others>Y>Z since that could lead to unexpected and bad results. Allowing them to rank A>B>C>D>Y>Z(>all_others) is better since then they have to explicitly indicate that C and D are better than Y and Z.

Juho