[EM] Improved IRV

Forest Simmons forest.simmons21 at gmail.com
Sat Mar 25 00:17:57 PDT 2023


1. Let PL be the Pairwise Loser between the two candidates with the fewest
transferred  votes.

2. Eliminate every candidate that does not beat PL pairwise, including PL
itself.

3. While more than one candidate remains,' repeat steps 1 and 2.

Now Improved Coombs:

1. Let PL be the Pairwise Loser between the two candidates with the
greatest Bottom  counts among the remaining candidates.

2. Eliminate every candidate that does not beat PL pairwise, including PL
itself.

3. While more than one candidate remains,' repeat steps 1 and 2.

Improved  Borda:

1. Let PL be the Pairwise Loser between the two candidates with the
smallest  Borda Counts  among the remaining candidates.

2. Eliminate every candidate that does not beat PL pairwise, including PL
itself.

3. While more than one candidate remains,' repeat steps 1 and 2.

Improved Score

1. Let PL be the Pairwise Loser between the two candidates with the least
scores among the remaining candidates.

2. Eliminate every candidate that does not beat PL pairwise, including PL
itself.

3. While more than one candidate remains,' repeat steps 1 and 2.

Improved Academic Washout Elimination
1. Let PL be the Pairwise Loser between the two candidates with the least
GPA among the remaining candidates.

2. Eliminate every candidate that does not beat PL pairwise, including PL
itself.

3. While more than one candidate remains,' repeat steps 1 and 2.

Improved SPE
1. Let PL be the Pairwise Loser between the two candidates nearest the
worst end of the Elimination Agenda among the remaining candidates.

2. Eliminate every candidate that does not beat PL pairwise, including PL
itself.

3. While more than one candidate remains,' repeat steps 1 and 2.

Of all these improved methods this one is the only one that preserves
monotonicity ... which is why it is the one singled out for a progression
proposal.

But all of these methods are Banks efficient ... hence Landau efficient ...
which prevents the embarrassment of electing winner that is covered by a
loser.

None of the extant methods have that guarantee... not Rannked Pairs, nor
CSSD, nor Kemeny-Young , etc.

-Forest
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20230325/a117c693/attachment-0001.htm>


More information about the Election-Methods mailing list