Strategy Done For The Voter
-
Let each election have two pollings.
At the first polling, the voters indicate their opinion about the popularity of the candidates with the other voters. The tally of this first polling makes some kind of average.
In the second polling, the voters provide their sincere scores for the candidates.
An algorithm carries out one of the best known strategies for each voter to decide the voter's vote in a Score election. These synthesized "votes" are then tallied as in Score to determine the winner.
Approval-ordered Llull (letter grades) [10], Score // Llull [9], Score, STAR, Approval, other rated Condorcet [8]; equal-ranked Condorcet [4]; strictly-ranked Condorcet [3]; everything else [0].
-
@jack-waugh said in Strategy Done For The Voter:
At the first polling, the voters indicate their opinion about the popularity of the candidates with the other voters.
Do they promise to do this honestly?
-
The problem is that, even with cardinal methods, you can't escape the Condorcet paradox with all the problems it implies.
Say you have three candidates and nine voters with the following preferences:
3: A > B > C
3: B > C > A
3: C > A > BThen when you consider for each group of voters the best strategy under approval voting and repeat that for some iterations always based on the previous poll, the winner will turn in cycles.
-
@rob said in Strategy Done For The Voter:
Do they promise to do this honestly?
It's only a suggestion to them. They are not required to promise.
I am under no illusion that I can escape Gibbard with this scheme, which must look as though invented by someone who thought he was being clever.
-
@casimir said in Strategy Done For The Voter:
3: A > B > C
3: B > C > A
3: C > A > BWhat evaluation does each faction put on its middle candidate, honestly?
