Voting methods

[Fmr. Gov. NickG]

I've come across an interesting website that surveys different voting methods and analyzes them according to several sets of "ideal criteria".  

http://www.electionmethods.org

Obviously no method meets every criteria, but the site seems to strongly advocate a method they call "Cordorcet with Schwartz Sequential Dropping".   This method may be too complicated to be implemented realistically, but it is an interesting read.

[Niles Caulder]

Fascinating article, thank you Govorner!

I do take exception to the author's assertion that more of Ross Perot's supporters were inclined to vote for Bush over Clinton in '92...simply untrue.  Exit polls demonstrated Clinton's popular margin of victory would have been unaffected, and an examination of the turnout and results in '96 are in complete accordance with this assessment of the broad base of Perot's support.

[Nym90]

Fascinating article, thank you Govorner!

I do take exception to the author's assertion that more of Ross Perot's supporters were inclined to vote for Bush over Clinton in '92...simply untrue.  Exit polls demonstrated Clinton's popular margin of victory would have been uneffected, and an examination of the turnout and results in '96 are in complete accordance with this assessment of the broad base of Perot's support.

Indeed. In 1992, Perot took 17% of the Republican vote, and 13% of Democrats. In 1996, he took 6% of Republicans and 5% of Democrats. In both elections, he took the same percentage of the liberal vote as he did of the conservative vote. Perot probably hurt the Republicans a little more than he hurt Clinton in both races, but he hardly made much difference.

[minionofmidas]

Yes, very interesting.
I propose we take some of the old election results and see if there would've been any changes if this method had been used.

[Fmr. Gov. NickG]

My impression of this system is that it really helps centrist, compromise candidates, whether or not they have an independent base of support.  A candidate can win in this system without being anyone's first choice.  Instant Run-off (preferential) voting helps centrist candidates some, but only in they have a great enough base of support to make it to the final run-off.
Plurality vote (America's RL system) it at the other extreme, rewarding only candidates with a strong base, giving no value to compromise candidates.

Under Preferential voting and Cordorcet, Gore would have won Florida in 2000.  There would be no difference because the 3rd place candidate was on the extreme.

The difference can be seen in 1992, where Clinton would have (probably) won under Preferential voting, but Perot may have won under Condorcet voting.  Because Perot was ideological in the center between Bush and Clinton, he may have won a head-to-head match-up against both, making him the Condorcet "ideal democratic winner".  But under IRV, he would have been eliminated before the final run-off occured.

The big question to ask is: In a crowded field where no one has a majority, should a candidate who is everyone's second choice be the winner?  This candidate is Condorcet preferred to everyone other candidate, but he would be the first one cut off under IRV.

[minionofmidas]

Actually, I'd been thinking of forum elections...we're on the fantasy board, remember?
But yes, you're right. Condorcet helps centre candidates big time. It's thus a good idea for electing presidents, chairmen etc who are supposed to be acceptable to everybody, but not parliaments. Another thing is that it would not solve our forum elections problem, which is that it may benefit candidates if their supporters ignore the election system (Migrendel beat Umengus because one of his voters didn't cast a second preference. The same thing would have been true under Condorcet.)

[Fmr. Gov. NickG]

I'm trying to apply this to Akno's original Senate election (the one that created all the controversy), but I can't find any information on the site about how Condorcet voting deals with pairwise ties.  Does anyone familiar with this method have an answer?

[Akno21]

I'm trying to apply this to Akno's original Senate election (the one that created all the controversy), but I can't find any information on the site about how Condorcet voting deals with pairwise ties.  Does anyone familiar with this method have an answer?

I wish I did, but I don't really understand the Condorcet system well enough. Also, I don't know what all the controversy was about, although some people did complain, they had always been against IRV voting.

[minionofmidas]

I'm trying to apply this to Akno's original Senate election (the one that created all the controversy), but I can't find any information on the site about how Condorcet voting deals with pairwise ties.  Does anyone familiar with this method have an answer?

I don't think there was information on what to do with exact ties (and, of course, an exact tie is an exact tie is an exact tie. I'm not sure if they were exactly tied, though. Maybe dig up the vote thread and lemme see). You might try and understand that "cloneproof" bit. That went right over my head.

[Fritz]

Here are the results of that controversial election:

Voting is now closed.  

Final results:

Round 1: Keystone Phil 6, Akno21 4, NickG 4, Bullmoose 2.
Round 2: Keystone Phil 6, Akno21 5, NickG 4.
Round 3: Akno21 8, Keystone Phil 6.

Akno21 WINS!

No more posting in this thread, please.

Bullmoose was eliminated first, one of his voters had Akno as second preference.

NickG was eliminated next, three of his voters had Akno as second preference.

[Fmr. Gov. NickG]

If Cordorcet voting were applied to this election, an interesting thing happens.

Because Keystone Phil loses to every other candidate head-to-head, he is not even part of the initial Schwartz set.  Thus, the leader in first place votes (and winner under a plurality system) is actually the first candidate eliminated under Cordorcet.

The problem with finishing the evaluation of this election is that Bullmoose and I tie in our pairwise run-off, and I'm not sure where to go from there.

[minionofmidas]

Here are the results of that controversial election:

Voting is now closed.  

Final results:

Round 1: Keystone Phil 6, Akno21 4, NickG 4, Bullmoose 2.
Round 2: Keystone Phil 6, Akno21 5, NickG 4.
Round 3: Akno21 8, Keystone Phil 6.

Akno21 WINS!

No more posting in this thread, please.

Bullmoose was eliminated first, one of his voters had Akno as second preference.

NickG was eliminated next, three of his voters had Akno as second preference.

I need every single ballot.

[Fritz]

Just click on the part where it says "Quote from: Fritz on JUly 25, 2004, 12:02:20 pm".  That will get you there.