The prestigious blog www.fivethirtyeight.com, has this article, “Might this be a good time for Instant Runoff Voting?” The article is a reaction to the May 22 special congressional election in Hawaii, at which a Republican was elected with less than 40% of the total vote cast. Thanks to Peter Gemma for the link.
IRV = THE method to elect Stalin/Hitler clones when the *moderate* Middle is divided.
34 H–M–S
33 S–M–H
16 M–H–S
16 M–S–H
99
IRV IGNORES most of the data in a Place Votes Table
H 34–16–49
S 33–16–50
M 32–67–0
Spare me and this list about any IRV use in a foreign nation.
The U.S.A. is now FULL of left/right Stalin/Hitler clone extremists — accumulated since 1929.
P.R. and nonpartisan App.V.
#1
Yawnnnnnnnn…….
If Hawaii had implemented a Top 2 Open Primary, then Hanabusa and Djou would have advanced to a runoff, where supporters of Ed Case would have been permitted to reconsider who they wanted as their representative. Or perhaps Case would have finished 2nd. Some voters may have been convinced that he was a “spoiler” and voted for Hanabusa.
I did like how the 538 column pointed out the ability to express preferences for as many candidates as there were running was “theoretical” given the actual practice in San Francisco, Oakland, Berkeley, San Rafael, Minneapolis, and Pierce County.
The thing is, knowing only that Hanabusa got more votes than Case, we can’t know that the electorate would actually be best-served by a Hanabusa/Djou runoff; it’s possible that Case would actually be a better choice (taking the aggregate opinion of the entire electorate into consideration) than either Hanabusa or Djou.
I think that in this case it is highly *likely* that Hanabusa would be preferred, but like Louisiana saw with the famous “lizard vs. wizard” runoff, there are times where the third-place choice from the initial vote *is* better than either of the top two.
(And I will throw in for approval and score voting, which have no such problems in finding the true best preference in a three-candidate race, unlike plurality, or top-two runoff, or instant runoff.)
Dale has no basis for suggesting that approval voting and range voting would have “found the best preference” in this race. Quite likely, backers of the three major candidates would have overwhelmingly cast “single shot” votes for their favorite candidate rather than cast an equally weighted vote for their “lesser of two evils” among the other two.
Some folks are better in math than psychology, it seems.
Rob,
It’s ironic for you to accuse Dale of misunderstanding voter psychology. Your “single shot” (aka “bullet voting”) argument is like saying that Nader supporters who cast a tactical vote for Gore back in the 2000 presidential race would have switched their vote from Gore to Nader if given unlimited votes, rather than still voting for Gore and also Nader. That is implausible, and you’ve offered absolutely no evidence to support it. It’s neither sincere nor tactical — it’s just harmful to the voter.
Because you and other IRV proponents so frequently make this unfounded claim, I created this page to cite:
http://www.electology.org/debate/BulletVoting
In short, say the two liberal candidates are L1 and L2, and the conservative candidate is C. If a voter knows who the two frontrunner candidates are, the simplest tactic is to just vote for his favorite frontrunner (like a tactical plurality voter) and then every candidate he likes better. If he doesn’t know who the frontrunners are, he should just approve every candidate he likes better than average. For instance, say he rates the candidates on a 0-10 scale as follows:
L1=10, L2=8, C=0
He prefers L1 and L2 to the average score of 6, so he should support both of them. You point out that if he supports L2, he may get L2 instead of L1 — which would decrease his satisfaction by 2 points. But I point out that if he bullet votes, he may get C instead of L2, which costs him eight points of satisfaction.