[EM] Replacement Candidates? - Where are you hiding?

[Bart Ingles]

Ron Tannenwald wrote:
> [snip]
>       I agree that runoff voting does not possess the shortcoming of
> non-elimination ranking schemes(i.e. lower ranked candidates can help defeat
> favorites). However STV is not particularly easy and heavily favors candidates
> with small but fervent support (e.g. single-issue constituencies) to the
> exclusion of more "acceptable" compromise candidates.Such "acceptable"
> candidates are very likely to be eliminated in the first rounds leaving a
> field
> of "extreme" choices none of whom is acceptable to the backers of the others.
> Just as in plurality voting (a much simpler method) a winner might be totally
> unacceptable to an overwhelming majority of the electorate.In addition STV
> is non-monotonic;it seems weird that you can help your favorite by not voting!
> No system is without its flaws. Perhaps my idea would be improved by allowing
> all candidates in a second election;eventually,though,something has to be done
> to prevent a reoccurence of strategic truncation.


A variant of Instant Runoff which has been discussed on this list
several times over the last few years, allows voters to give
equal-ranking to candidates.  In essence, the voter is encouraged to use
strategy when necessary; rather than voting a compromise over his true
favorite, he can simply vote for both as co-first choices.  These
equal-votes are then all counted as full votes, as in Approval voting. 
Only when both first choices are eliminated does the vote transfer to a
second-ranked choice.

The method appears in the archives for this list as Instant Runoff-1 or
IR-1 in various places.  Obviously not very promininent, since there
doesn't even seem to be a more formal name for it.  Interestingly,
though, it seemed fairly well received even among Condorcet supporters,
at least as an alternative to Instant Runoff.  The greatest criticism
seemed to be that it was not widely studied, and would require the
public to understand  both Instant Runoff and Approval in the same
system.

When looking into it myself, the method appeared to be better than
Instant Runoff at allowing a highly-rated Condocet winner to succeed,
while rejecting low-rated Condorcet winners almost as well as IRO and
Approval do separately.

Bart Ingles