The moral of Rob's example below is that PR should not be done by
sequential choice.
That's why PAV is superior to sequential PAV, for another example.
Since STV is a sequential choice method, it is essentially a bailing wire
and string kluge that may be good enough for government work, but should
be avoided in serious applications.
Forest
On Fri, 22 Aug 2003, Rob LeGrand wrote:
> John B. Hodges wrote:
> > CPO-STV is an awesome multiseat method, conceptually. I'm wondering
> > if there is a computationally efficient way of arriving at the same
> > "ideal" ensemble. My "For Dummies" guess is that the ideal ensemble
> > will never include a Condorcet loser and will always include a
> > Condorcet-winner if one exists. STV with Rob's "orphan" elimination
> > rule would (I guess) be sufficient to do that much.
>
> I'm not sure I agree that the ideal ensemble should always include a
> Condorcet winner and exclude a Condorcet loser. For example, if we're
> electing two winners from three candidates and the preferences are
>
> 45:Reagan>Anderson>Carter
> 20:Anderson>Carter>Reagan
> 35:Carter>Anderson>Reagan
>
> I would argue that the outcome {Reagan, Carter} is the most proportionally
> representative. In a sense, the Reagan-first voters' preference for
> Anderson over Carter *should* be ignored since they already got their first
> choice in Reagan. (Would anyone disagree with that? Is there a good
> argument for preferring the outcome {Carter, Anderson}?) So this "ideal
> ensemble" includes the Condorcet loser (Reagan) and excludes the Condorcet
> winner (Anderson).
>
> The moral here is that single-winner and multiwinner elections are quite
> different. While the best single-winner methods can concentrate on
> limiting insincere voting without worrying about proportionality, the best
> multiwinner methods must strike a balance between striving for good
> proportionality and discouraging insincerity.
>
> I think of a spectrum with single-winner systems at one end and something
> like Direct Representation (http://www.directrep.org/) at the other.
> Strewn in the middle are n-winner systems. Somewhat paradoxically, "vote
> for any number of candidates" (Approval) is best for the single-winner case
> and "vote for only one candidate" (SNTV) becomes best for the n-winner case
> as n increases. Approval is great at removing insincerity from strategy
> but gives poor proportionality (which is irrelevant in the single-winner
> case), while SNTV often encourages insincerity (though less and less as n
> increases) but gives good proportionality (which improves as n increases).
> So it's not obvious to me that the best multiwinner systems reduce in the
> single-winner case to the best single-winner systems.
>
> =====
> Rob LeGrand, psephologist
> [EMAIL PROTECTED]
> Citizens for Approval Voting
> http://www.approvalvoting.org/
>
> __________________________________
> Do you Yahoo!?
> Yahoo! SiteBuilder - Free, easy-to-use web site design software
> http://sitebuilder.yahoo.com
> ----
> Election-methods mailing list - see http://electorama.com/em for list info
>
----
Election-methods mailing list - see http://electorama.com/em for list info
