[EM] CD, briefer, better-workng, and with Juho's clarifications

[Juho Laatu]

We seem to have similar understanding on how the CD should be defined. My problem was just that I couldn't see these features (multiple strategic groupins, that prefer C candidates above others, causing some cadidate outside C to win) fully reflected in the CD definition that you sent out earlier today.

Juho


On 11.1.2014, at 20.38, Michael Ossipoff <email9648742 at gmail.com> wrote:

> 
> 
> On Sat, Jan 11, 2014 at 11:03 AM, Juho Laatu <juho4880 at yahoo.co.uk> wrote:
> Some comments on the CD criterion itself.
> 
> I see CD as a situation where there is a group of voters (V) that has majority and whose members all prefer certain set of candidates (C) over all other candidates. With sincere votes one of the C candidates is guaranteed to win (it may be unclear which one of them). Some V voters can improve the expected outcome by voting some C candidates "lower". The Chicken Dilemma emerges when multiple members/subgroups of V (with different preferences within C) follow this strategy, and as a result some candidate outside C wins.
>  
>  
> Yes, that's the situation that CD is intended to test for.
>  
>  
> 
> This is not an exact definition yet, but you can see what I mean. Should we assume that in CD there are multiple strategic groups within V?
>  
> Yes. There might not always be, but there often will be. U.S. progressives are a terribly fractious group, with certain factions, which will remain un-named, angrily criticizing all progrssives who don't share their ideology.
>  
> With a chicken-dilemma-vulnerable voting system, the chicken dilemma would devastate the progressives. The progressives would mutually self-destruct.  ...unless they were real good at probabilistic partial-cooperation anti-defection strategy. Should we count on that? Better to just not have chicken dilemma, and thereby have the benefit of MMC and the Condorcet Criterion.
>  
>  
> Should we assume that one of the candidates outside C will win?
>  
>  
> Most definitely.
>  
> Michael Ossipoff
> ----
> Election-Methods mailing list - see http://electorama.com/em for list info

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20140111/864195aa/attachment-0004.htm>