Algorithmic Asset Voting: Difference between revisions

Line 136:

1 C>A>B </blockquote>there is no strategically stable winner, because if any voter attempts Favorite Betrayal to make their preferred candidate the CW, then 2 other voters have incentive to alter their votes to make someone they prefer the CW, and so on.

== Semi-solid coalitions ==

A semi-solid coalition (a possibly incorrect idea based off of [https://en.wikipedia.org/wiki/Proportionality_for_Solid_Coalitions#Solid_coalitions solid coalitions]) is a group of voters who prefer a set of candidates over all others, but some voters in the group may prefer other candidates over the candidates the group prefers over all others. So, for example,

20 A

5 B>A

would be a semi-solid coalition of size 25 for {A, B}. A solid coalition would also technically be a semi-solid coalition, so any voting method that always elects proportionally from semi-solid coalitions should pass [https://en.wikipedia.org/wiki/Proportionality_for_Solid_Coalitions PSC].

It may be possible to make AAV always elect proportionally from Droop semi-solid coalitions. If so, then this may greatly speed up computation of the result, as only winner sets that satisfy all semi-solid coalitions need be considered; discovering semi-solid coalitions can be done by checking ballots and observing which candidates are ranked higher than others and by which voters. <ref>https://www.removeddit.com/r/EndFPTP/comments/euiup2/a_new_pr_concept_of_semisolid_coalitions/</ref> Semi-solid coalitions may overlap.

26 A>B

25 B

49 C

There is a Condorcet cycle here

== References ==