Coombs' method: Difference between revisions

Added more criterion failures.
(Added info about how it differs from IRV)
(Added more criterion failures.)
Line 2:
'''Coombs' method''' (or the '''Coombs rule''')<ref>Grofman, Bernard, and Scott L. Feld (2004) [https://dx.doi.org/10.1016/j.electstud.2003.08.001 "If you like the alternative vote (a.k.a. the instant runoff), then you ought to know about the Coombs rule,"] ''Electoral Studies'' '''23''':641-59.</ref> is a [[ranked voting systems|ranked voting system]] created by [[wikipedia:Clyde Coombs|Clyde Coombs]] used for single-winner [[election]]s. Similarly to [[instant-runoff voting]], it uses candidate elimination and redistribution of votes cast for that candidate until one candidate has a majority of votes. Its difference from [[IRV]] lies in its elimination criterion: instead of eliminating the candidate ranked first by the fewest voters, it eliminates the candidate ranked last by the most.
 
==Example Properties==
 
Coombs' method frequently selectsfails the [[Condorcet winner criterion|Condorcet winner]]. However, thisthe does[[monotonicity notcriterion]], alwaysand happen.the For[[participation criterion]]. example:
 
This example, placed in [https://cs.angelo.edu/~rlegrand/rbvote/calc.html Rob LeGrand's voting calculator], shows that Coombs arrives at a different result than Condorcet. The examplefollowing isexamples pulledare fromdue ato Felsenthal and Tideman paper.<ref name="FT">{{Cite journal |last=Felsenthal |first=Dan |last2=Tideman |first2=Nicolaus |date=2013 |title=Varieties of failure of monotonicity and participation under five voting methods |url=https://www.researchgate.net/profile/Dan-Felsenthal/publication/257667897_Varieties_of_failure_of_monotonicity_and_participation_under_five_voting_methods/links/54aec0fb0cf21670b35870a6/Varieties-of-failure-of-monotonicity-and-participation-under-five-voting-methods.pdf?origin=publication_detail|journal=Theory and Decision |language=en |volume=75 |issue=1 |pages=59–77}}</ref> unless otherwise noted:
7:a>c>d>b
6:a>d>b>c
3:b>a>c>d
7:b>c>a>d
9:b>c>d>a
4:c>a>d>b
6:d>a>b>c
3:a>c>b>d
 
=== Condorcet criterion ===
This example, placed in [https://cs.angelo.edu/~rlegrand/rbvote/calc.html Rob LeGrand's voting calculator], shows that Coombs arrives at a different result than Condorcet. The example is pulled from a Felsenthal and Tideman paper.<ref>{{Cite journal |last=Felsenthal |first=Dan |last2=Tideman |first2=Nicolaus |date=2013 |title=Varieties of failure of monotonicity and participation under five voting methods |url=https://www.researchgate.net/profile/Dan-Felsenthal/publication/257667897_Varieties_of_failure_of_monotonicity_and_participation_under_five_voting_methods/links/54aec0fb0cf21670b35870a6/Varieties-of-failure-of-monotonicity-and-participation-under-five-voting-methods.pdf?origin=publication_detail|journal=Theory and Decision |language=en |volume=75 |issue=1 |pages=59–77}}</ref>
 
Even though Coombs' frequently selects the [[Condorcet winner criterion|Condorcet winner]], it sometimes fails to do so. For example:
{{ballots|
7:a A>cC>dD>bB
6:a A>dD>bB>cC
3:b B>aA>cC>dD
7:b B>cC>aA>dD
9:b B>cC>dD>aA
4:c C>aA>dD>bB
6:d D>aA>bB>cC
3:a A>cC>bB>dD
}}
 
This example, placed in [[Online_poll#Online polling sites|Rob LeGrand's voting calculator]], shows that Coombs arrives at a different result than Condorcet.
 
=== Monotonicity criterion ===
 
In the election
 
{{ballots|
1: A>B>C
10: A>C>B
11: B>A>C
11: B>C>A
10: C>A>B
2: C>B>A
}}
 
C wins, but if the 11 B>A>C voters raise C and vote B>C>A, then B wins.
 
=== Participation criterion ===
 
In the election
 
{{ballots|
7: A>C>D>B
6: A>D>B>C
3: B>A>C>D
7: B>C>A>D
9: B>C>D>A
4: C>A>D>B
6: D>A>B>C
}}
 
A is the CW and wins. But if three additional voters vote A>C>B>D then B wins.
 
==Links==
1,196

edits