chess: (Default)
The AV Yes Campaign are doing a photo shoot on Saturday. I can't get to it because I'm at a player event, so I thought I would post about it. Event details and signup at http://tinyurl.com/purplephoto

I suspect if I actually made any commentary on AV I would mostly be preaching to the converted here, but I'm always happy to host a political argument on this or basically any subject if you like :).

The very short version is: I want the AV Yes Campaign to win because it would mean that I never, ever have to deliver any propaganda with the phrase 'two-horse race' or a stupid bar graph showing how Foo Can't Win Here ever again...

Date: 2011-01-25 03:32 pm (UTC)From: [identity profile] naath.livejournal.com
Ahahahaha. I hate those stupid graphs too.

Date: 2011-01-25 04:54 pm (UTC)From: [identity profile] phlebas.livejournal.com
They are annoying. but surely you'd need some kind of actual PR system rather than just AV to eliminate tactical voting (and thus the stupid graphs)?

Date: 2011-01-25 05:16 pm (UTC)From: [identity profile] phlebas.livejournal.com
Mm. But the tricky case is when X and Y are the parties actually likely to win, Z is third and there are significantly many votes going to minority parties. Then the deferral from Z is less likely ever to happen, and it's more of a priority (for X's publicist) to persuade the Z voter to put X first.

Of course there's also more need for X to avoid offending the dyed-in-the-wool Z voter so much as to lose the second-choice vote, so maybe it works out.

Date: 2011-01-25 05:56 pm (UTC)From: [identity profile] 1st-law.livejournal.com
Unfortunately tactical voting exists in PR. It's the same as in the local election but at the national scale.

E.g. If you want to keep the Tories from forming a government you're still better off voting for Labour than the Green party in PR. Admittedly, however, it is less bad. The vote for Green does dilute the Tory vote a little rather than being wasted.

Arguably the same applies in AV but in AV you effectively always end up casting a vote either for the winner or the first placed loser in your area (if you keep redistributing past the point where the winner has been decided) so it's much better just to honestly list your preferences. Just put the BNP at the bottom of your list and you will always vote for whoever isn't them.

One reason I don't like PR is that you don't have a single person who you elect you instead contribute to a list. It means there isn't an individual who is responsible to a set of voters and it formalises party politics when in theory FPTP and AV both allow independents to be significant.

My personal wish would be to make the House of Lords PR but it doesn't seem like PR in any form is going to be offered as a choice any time soon.

Date: 2011-01-25 08:44 pm (UTC)From: [identity profile] king-of-wrong.livejournal.com
So, coalitions every time and endless people whining that $PARTY has betrayed them by coming third in the election and yet not carrying out every single detail of what they said they'd do if they won? I'll give it a miss, thanks...

Just out of interest, though, how is AV "fairer"? In FPTP, every vote cast has an equal weight and is pretty much fair under any definition you can come up with. As opposed to AV, where disproportionate weighting is given to people whose first preference loses badly.

(Not that it matters, because with the Labour lords filibustering the AV discussion, it's going to miss the deadlines for a May 5th vote...)

Date: 2011-01-25 10:20 pm (UTC)From: [identity profile] 1st-law.livejournal.com
I'm not sure it would always lead to coalitions. Clearly if one party wins by a few votes everywhere they will get a majority of MPs just like under FPTP. Also aside from the gnashing of teeth I don't think coalitions are too bad anyway. We aren't very used to them but they have them on the continent and they work most of the time.

Fair is always hard to define. I really don't agree that anyone gets disproportionate weighting in AV.

I'd phrase it in terms of incentives. In FPTP you have no incentive to vote for the candidate who's only poling at 10% so you might as well vote for someone else. If you chose to vote for the people you want to win your vote would be "wasted". This is undesirable because it means people aren't voting for who they want. They are second guessing who they believe will win and then choosing.

Surely it is better that people are encouraged to be honest in their preferences.

Date: 2011-01-25 10:59 pm (UTC)From: [identity profile] king-of-wrong.livejournal.com
I really don't agree that anyone gets disproportionate weighting in AV.

Really? How about this scenario: 49% of voters choose Red, 49% choose Blue, and nearly all of the remaining 2% pick the BNP. The minority parties are eliminated, second choices tallied, and the election is called based on the second choices of the people who voted for the extremists. Their votes are considered more times, even if they're only tallied once in the end.

What AV does, at least as I understand it, is give the tie-break decision in a close election to the people whose choices are furthest from the mainstream. We see that in coalition negotiations across the world, where a minority party gets to make demands of the first- and second-choice parties, despite polling far far lower - as they say, when things are nearly balanced, it's the guy with his thumb on the scale who gets to decide.

Also, depending on the order in which the losers are eliminated you can tip the results: AV does not meet the Condorcet criterion.
I'm not a big fan of AV because it doesn't have monotonicity - sometimes the only way to get your party to win is to vote against them. At least in FPTP the tactics were simple enough you didn't need a mathematician to explain them.

e.g.
You and a friend want the Labour party to win
Labour: 100 first preference
Lib Dem: 100 first preference, 2nd pref Labour
Conservative: 99 first preference, 2nd pref Lib Dem

If you both vote for Labour the Conservatives will be eliminated leading to a Lib Dem win. If you vote Conservative Lib Dem will be eliminated leading to your desired outcome.

In fact you cannot avoiding tactical voting in an election with more than 2 candidates (see Arrow's Impossibility Theorem. Systems like PR circumvent this problem by not making a decision. Unfortunately this only delays the problem - at some point your government has to make a decision about something, and the tactical voting merely happens then).

Ironically this proves that the only way to achieve 'fair' voting is to eliminate the Liberal Democrats.
In addition the vote has to be very very close for these distortions to occur.

You and your friend want Labour to win and you do this by voting Lib Dem in this case however this only works because the majority of Conservatives would also like Labour to win. If they are happy to have them as second choice this allows them this chance.

Date: 2011-01-26 11:02 am (UTC)From: [identity profile] naath.livejournal.com
That's how I understand it too.

Also that if Z want X *and* Y voters to 2nd-pref Z then they HAVE TO BE NICE (or at least NICER) about X and Y.

Date: 2011-01-26 11:16 am (UTC)From: [identity profile] naath.livejournal.com
AV minimises the total unhappiness with the result, which some people (such as me) think is a fairer way of deciding what the result should be than the current system, which can give a result a plurality of voters prefer but a majority of voters are unhappy with.

Of course whether one believes this to be "fairer" is a matter of opinion since "fairness" is not completely objective.

A different example illustrating the non-objectiveness of fairness - I have 1000 pounds of money to give away to people, I have 10 people asking for money. I could say "it is fair to give each person 100 pounds", or I could say "it is fair to assess each person's needs and apportion the money according to need"; both options are justifiable as "fair".

Date: 2011-01-26 04:08 pm (UTC)From: [identity profile] alextfish.livejournal.com
"coalitions every time and endless people whining that $PARTY has betrayed them by coming third in the election and yet not carrying out every single detail of what they said they'd do if they won?"

Hopefully, "coalitions every time" should lead (over the course of a few years) to the public realising that give and take is a natural part of how coalitions work. I don't think "the public don't understand coalitions" means "we should avoid coalitions" but "we should help the public understand coalitions"; if they're more natural in the voting system, that'll be a likely consequence.
Ironically this proves that the only way to achieve 'fair' voting is to eliminate the Liberal Democrats.

You have to admit that Nick Clegg is doing his best on that score.
He certainly is.

I think he's getting some concessions behind the scenes (otherwise why is the Tory backbench so grumpy?) but they are too few and not public enough to keep the party's supporters on board.

Date: 2011-01-27 10:32 am (UTC)From: [identity profile] 1st-law.livejournal.com
It is true that the people that would prefer minority parties get to contribute if the election is close but they only have a choice of one of the two majority parties that ended up having broad support.

AV strongly favors candidates who appeal to most people and so you wouldn't expect extremists to get elected.

As is being discussed in the other thread the Concordat Criterion can't be satisfied in any system that has independence of irrelevant alternatives. That is vote splitting or other similar distortions.

I would argue that the concordat criterion is less important than independence of irrelevant alternatives as in most cases parties can be ranked in terms of similarity to an ideal or loosely on a political spectrum. This means that if people rank one party slightly above another then a good number of people will also have ranked those parties the other way round. This avoids a number of issues you can get with truly arbitrary sets of preferences like the example given in the other thread.

Finally I would say even though no voting system can be prefect AV is much closer to electing the candidate with the broadest support than FPTP.
The vote has to be very close for one or two votes to make the difference in any reasonable system. Larger blocks can successfully follow this strategy in less close races - obviously the less close it gets the larger the block required!
Tactical voting always relies on having information about how others are going to vote - hence the bar charts with the dodgy axis.

In this case the information you need is what voters for one party are likely to list as their second choice preference. If AV becomes our voting system I expect a rash of surveys to give us that information. In fact if it does starting a polling company would be a smart move!
I think the size of the block needs to be vote difference between the 2nd and 3rd party plus one. Obviously plus chess's requirement that you have to broadly know the 2nd choices of the people in the 2nd and 3rd parties.

  • Pupil Premium

  • Capital Gains Tax

  • AV referendum

  • School Budget increased

  • Inheritance Tax Threshold frozen

  • Student loan payback threshold high and index-linked

  • Pro-Europe foreign policy



Clegg's list of victories is impressive and public. I suspect the problem is that people didn't vote Lib Dem because they wanted the manifesto implemented, but as a protest vote against the Tories.
I agree. The comparison point in FPTP is a block the size of the difference between the 1st and 2nd parties + 1 and knowledge of which the top 2 parties will be.
Clearly I should read the news more as I was only vaguely aware of half of these. Good job Cleggy boy!

Profile

chess: (Default)
Michelle Taylor

January 2025

S M T W T F S
   1234
567891011
12131415161718
19202122232425
262728293031 

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Feb. 1st, 2026 08:41 pm
Powered by Dreamwidth Studios