Alternative Voting

[Havet]

The following videos have apparently been created only like a week ago. In short, they explain the flaws in the current voting system, and propose a new voting system, called 'Alternative Voting', where single-candidate voting is replaced with candidate preference.

I know this is not revolutionary, as it tries to soften the status-quo, but it is nonetheless an interesting thought experiment, and could increase the power smaller parties have.

Watch these two videos, and discuss if it has any potential benefits:

s7tWHJfhiyo

3Y3jE3B8HsE

[Kotze]

2x "This video is not available in your country."

I'm pretty sure this is about what is known in the USA as Instant Runoff Voting. In the US it's pushed by FairVote, and it wouldn't surprise me if these videos are influenced by the rhetoric of that group.

Could you paraphrase what they claim about the properties of that system plz? Did they say it solves the spoiler effect or that you are never under pressure to vote strategically? I often encounter this with pieces by FairVote, and both claims aren't quite correct.

[Havet]

2x "This video is not available in your country."

Sorry, but I can't do anything about it. It sucks that youtube keeps enforcing geographical content.


I'm pretty sure this is about what is known in the USA as Instant Runoff Voting. In the US it's pushed by FairVote, and it wouldn't surprise me if these videos are influenced by the rhetoric of that group.

In his blog, the guy says:


In case it’s not obvious enough from the video, I’m a big supporter of the Alternative Vote Referendum (http://en.wikipedia.org/wiki/United_Kingdom_Alternative_Vote_referendum,_2011) and the Yes in May Campaign (http://www.yestofairervotes.org/).


Could you paraphrase what they claim about the properties of that system plz? Did they say it solves the spoiler effect or that you are never under pressure to vote strategically? I often encounter this with pieces by FairVote, and both claims aren't quite correct.

Basically he supports you voting for each candidate in order of preference. In each voting paper, you would rate candidates by preference (ie: 1st, 2nd, 3rd, 4th choice, and so on) rather than just picking one.

He also claims to solve the spoiler effect, though he also recognizes that that system doesn't solve the problem of gerrymandering, proportional representation and condorcet winners

[Kotze]

Here is an example of Alternative Vote (horrible name) aka Instant Runoff having a spoiler problem:

In the first election there are 2 candidates, A and B, and 100 voters.

60 vote for A.
40 vote for B.

A wins.

In the second election there are again candidates A and B, and the same 100 voters with the same preferences regarding A versus B as during the last election, and there is a new candidate, C. People vote with fancy ranked ballots.

45 vote for C first, A second, B last.
15 vote for A first, B second, C last.
40 vote for B first, A second, C last.

As you can see, when we ignore C here we get the same pattern as in the last election. Now who wins? Whether we are talking about Instant Runoff or people literally having an election with a runoff, what happens if they vote according to these preferences is that A gets excluded, then it's B versus C, and the winner of that is B.

This is an example of the spoiler effect. The winner is B and if C hadn't run, the winner would have been A, something preferred by the C-fans to the actual outcome.

[Tim Finnegan]

Decent videos, although the comments on an "inevitable two-party system" are a little off-base; what's more correct to say is that you'll likely get a two-party split in any given geographical area, rather than absolutely. For example, in my constituency, the real choice is between one of the major parties, Labour, and a "third" party, the SNP, with the traditional second major party, Conservative, coming a sorry third. I think this may be a reflection of their country's particular political culture: the American parties are far more diverse and locally autonomous than most parties, so each are able to twist themselves to cater to almost any region, which the narrower, more centralised parties of other countries usually fail to do.

[RGacky3]

In a Parlimentary system, its a little different from the US system, a parlimentary system does not go towards a 2 party system, as for many issues minor parties can make a difference.

But this is a GREAT video, especially to show the team-USA people who spout off all the time "the people could just VOTE communism in if they wanted too," not that people making those arguments would be smart enough to get this video (you have to be pretty dumb to make that argument).

Alternative voting in a presidential system is better, and it makes sense. Its my opinion however, that the NUMBER ONE priority for people wanting to fix the democratic system is campain finance reform and getting rid of corporate personhood, those 2 things alone will do wonders for democracy, then having a better voting system after will help even more.

Thanks for the videos though Havet, very intersting. (out of all of the Free-Market people on revleft, Havet is one of the few honest ones engaging in real discussions, a rarity).

[Arilou Lalee'lay]

Ugh. Calling it the "alternate vote" is a blatant attempt to cover up the fact that it is one among many alternatives. It's also a very illogical one to support because it fails the Condorcet criterion.

And the video seems very fuzzy and unconvincing to me. The way it is usually discussed among theorists in the field is much more clear. I really wish I could post links (just four more posts!), but for now you'll just have to go to the Wikipedia page for the Schulze method and scroll down to the "comparison to other preferential single winner..." chart and it should be obvious why the Schulze or Tideman methods are way better than IRV, which is only slightly better than plurality voting.

Here's the for dummies version:
The Condorcet criterion is really simple. It's that the most preferred candidate should always win. The voting systems of all modern countries fail this criterion. This was made obvious in Florida in 2000.

If you're too lazy to read up on the Schulze method because you don't see where the logical flaw could be in IRV, consider Coombs' method. In IRV, the candidate with the fewest most preferred votes is eliminated. In Coombs' method the candidate with the most least preferred votes is eliminated. Why one and not the other? They're completely arbitrary. What you need is an elegant voting method motivated by the Condorcet criterion, and this is what the Schulze method is. It's a bit tough to explain, but the charts on the Wikipedia page shouldn't be too hard to follow.

This issue should be important to all leftists, not because we want a third party now, but because after the revolution we will still be voting for delegates, representatives, or on the issues directly. The flaws in plurality voting will still effect us then. Also, if we adopt Schulze voting it would draw a big crowd from the open source movement. It would demonstrate to them that we are right about something that requires a lot of thought to figure out, and thus that our other ideas are worth considering.

[Kotze]

The Condorcet criterion is really simple. It's that the most preferred candidate should always win.I agree with you on Schulze over IRV (Instant Runoff Voting), but I don't think most of those who don't already know what Condorcet means will understand it by reading above sentence. It's better to explain it with more words, like this:

Opinions might be very different when it comes to how to deal with situations with more than 2 candidates, but in situations with only 2 candidates I hope we can all agree that the 1 preferred by the bigger group should win.

When the voters give you ballots where they rank candidates, you can infer who probably would have won if it had been an election with only 2 candidates by only looking at how often A got ranked ahead of B and vice versa while ignoring all other candidates. There are many situations where there is a candidate that wins every such pairwise comparison with another one.

Plurality doesn't guarantee to give you that beats-all candidate, and neither does Plurality with Runoff or Instant Runoff. Methods that do are called Condorcet Methods.

Another problem that I have with IRV is the crappy interface. If a ballot has more than 1 candidate in the same rank above bottom (whether it's a voter's honest opinion or the people tabulating the results have problems telling some of your numbers apart), that ballot goes straight into the trash, or if you are lucky they have some additional rules that your ballot only gets into the trash once all the candidates you put above your equal-ranked candidates are disqualified during the counting process.

With Schulze, equal-ranking is interpreted as abstaining from the vote in pairwise comparisons of those sharing the same rank.

A big selling point of IRV is that, unlike with many other alternative approaches like Range Voting or any method from the Condorcet family, stating your lower preferences doesn't affect the winning probabilities of your higher preferences.

Say we use Range Voting, meaning you rate candidates and we take the one with the highest average (arithmetic mean) rating to be the winner. You honestly rate your favourite A 5 stars and a less-liked candidate B 3 stars, the others get 0 stars from you. The winner turns out to be B who received just 2 more stars than A. If you had not expressed your lesser-liked preference, your more-liked candidate would have won. There is some incentive against stating preferences for less-liked candidates with Range Voting. Any method from the Condorcet family suffers from this problem as well (though to a lesser extent than with Range Voting, because with Condorcet, stating a lower preference can also positively affect the winning probability of your more preferred candidates.)

In IRV, the algorithm only looks at one of your preferences at a time. Your second preference is only looked at after your first preference got deleted, your support for your third preference only gets activated after the candidates you ranked first and second got deleted, and so on. In IRV, stating a lower preference doesn't hurt your higher preference, because the algorithm throws away your more-liked candidates before that might happen. Let's reflect on that a bit, people. A kid might bite your hand and hurt you, but the bite won't hurt if before meeting the kid you cut off your hand.

It's more to the point to push for proportional representation than using different single-winner-election procedures and among proposed single-winner alternatives to Plurality IRV is on the lame end (though the worst thing I've seen seriously proposed is forcing people to give a unique rank to each candidate and tabulation giving a fixed number of points based on the rank). If simplicity has a very high priority, a good alternative single-winner method is Approval, otherwise Schulze.

[JTB]

2x "This video is not available in your country."

I'm pretty sure this is about what is known in the USA as Instant Runoff Voting. In the US it's pushed by FairVote, and it wouldn't surprise me if these videos are influenced by the rhetoric of that group.

Could you paraphrase what they claim about the properties of that system plz? Did they say it solves the spoiler effect or that you are never under pressure to vote strategically? I often encounter this with pieces by FairVote, and both claims aren't quite correct.
Care to explain?

[JTB]

Here is an example of Alternative Vote (horrible name) aka Instant Runoff having a spoiler problem:

In the first election there are 2 candidates, A and B, and 100 voters.

60 vote for A.
40 vote for B.

A wins.

In the second election there are again candidates A and B, and the same 100 voters with the same preferences regarding A versus B as during the last election, and there is a new candidate, C. People vote with fancy ranked ballots.

45 vote for C first, A second, B last.
15 vote for A first, B second, C last.
40 vote for B first, A second, C last.

As you can see, when we ignore C here we get the same pattern as in the last election. Now who wins? Whether we are talking about Instant Runoff or people literally having an election with a runoff, what happens if they vote according to these preferences is that A gets excluded, then it's B versus C, and the winner of that is B.

This is an example of the spoiler effect. The winner is B and if C hadn't run, the winner would have been A, something preferred by the C-fans to the actual outcome.
Good point.

What would you propose?

[ComradeMan]

How about one vote for a home vote and two votes for an away vote and in the case of a tie a tele-vote shoot out? :laugh:

[Tim Finnegan]

In a Parlimentary system, its a little different from the US system, a parlimentary system does not go towards a 2 party system, as for many issues minor parties can make a difference.
Would you be able to elaborate? I'll admit to not be overly familiar with the intricacies of the American system.


Ugh. Calling it the "alternate vote" is a blatant attempt to cover up the fact that it is one among many alternatives.
That's not at all what the name is supposed to mean. It's called that in the UK, for example, when the existence of further alternatives is evidenced by their active practice in the devolved assemblies. I think that you're searching for malice where there really isn't any to be found.