AV: Yes or No?

[Chris Corby]

Christ Jono, your political understanding has suddenly expanded, having initially said that the only to reason to vote for AV was that David Cameron was against it!

[Michael Wallace]

Christ Jono, your political understanding has suddenly expanded, having initially said that the only to reason to vote for AV was that David Cameron was against it!

Given that you haven't countered it, I'm assuming you accept that the 'the BNP voters get more votes!' argument is wrong, then?

[Jon O'Neill]

Christ Jono, your political understanding has suddenly expanded, having initially said that the only to reason to vote for AV was that David Cameron was against it!

Well it's pretty obvious that both systems have drawbacks. Listening to the arguments presented on here is quite enlightening though. You can make contrived examples either way, and that's fine, but coming out with the "one person one vote" crap that Cameron and co. use to trick the thick-shit part of the country who might actually get off their arse and vote is just ridiculous. I thought we were above that level here.

[Charlie Reams]

I think it's possible to make inferences about someone's opinion beyond what they state explicitly.

. . . puts you at significant risk of confirmation bias.

And I was completely right. I'm sure you will humbly acknowledge this.

Your example (unlike mine) bears no relation to the result of any election - the winner has 11% of the vote???? Never ever!

But I thought...

The assertion that my results are unlikely to be reflected in a real local or general election miss the point.

So which is it?

The trouble with this campaign nationally is the absolute garbage spouted. I heard John Reid this morning (a NO to AV man) compare the AV system as the equivalent of buying a Mars Bar, taking a bite out of it and returning it to the shop and exchanging it for a Twix (WTF?)

Agreed. Crap has been spouted on both sides. It's perfectly possible to ignore it though.

These could now include some transferred BNP votes so the argument against AV says that those who voted for BNP have now had two votes and some are now heading for a third! If the combined votes for the BNP and UKIP parties have all now voted LibDem as their second preference that candidate has passed the 50% threshold and is elected (unlikely) so I would predict the count goes on. Now the second preference Labour votes are taken into account. Some BNP voters could be heading for a fourth vote!

This is completely wrong though, as has already been explained several times in this topic and elsewhere.

I now think it highly likely that the LibDem candidate will get over 50%. So really no different in this instance to first past the post.... but a better system? This is what you have to decide.

Yes, I understand that we have to decide which is the better system. I don't think anyone was confused about that.

heard Paddy Ashdown on Newsnight say that under AV, no vote is wasted? Really? What about all those who voted Conservative above? Were their second preference votes ever taken into account?

That's a fairly loose definition of "wasted" which I wouldn't really agree with. But the fact that Paddy Ashdown likes AV for a bad reason doesn't mean that AV itself is bad, any more than John Reid talking crap about Mars Bars makes FPTP wrong. It's irrelevant.

[Ian Volante]

Under AV, the BNP votes are looked at again and their second choice votes are added to the remaining four candidates. Even if all these votes go to the same candidate, there is still no 50% target reached so the UKIP second preference votes are counted. These could now include some transferred BNP votes so the argument against AV says that those who voted for BNP have now had two votes and some are now heading for a third! If the combined votes for the BNP and UKIP parties have all now voted LibDem as their second preference that candidate has passed the 50% threshold and is elected (unlikely) so I would predict the count goes on. Now the second preference Labour votes are taken into account. Some BNP voters could be heading for a fourth vote! I now think it highly likely that the LibDem candidate will get over 50%. So really no different in this instance to first past the post.... but a better system? This is what you have to decide. I heard Paddy Ashdown on Newsnight say that under AV, no vote is wasted? Really? What about all those who voted Conservative above? Were their second preference votes ever taken into account?

But they haven't had two votes, have they???????????????????? Becuase their first vote doesn't count any more????????

Honestly, it's not that fucking difficult. That whole several votes argument is one of the most moronic things I've ever heard.

Aye.

[David Williams]

I think it's possible to make inferences about someone's opinion beyond what they state explicitly.

. . . puts you at significant risk of confirmation bias.

And I was completely right. I'm sure you will humbly acknowledge this.

Being right (whether you are or not) doesn't mean you didn't have a significant risk of confirmation bias. Let me know what I think. Will save me a bit of time deciding which way to vote.

Yours humbly, as ever.

[Soph K]

Dick

Forgive me for not actually reading anything but the quoted bit but...er, why does one of the names have to be Dick?!????!!

Dick is also slang for penis.

EXACTLY! That's exactly why I said why does it have to be Dick!!!!!

[Charlie Reams]

I'm sure you will humbly acknowledge this.

Yeah, I thought not.

[Chris Corby]

This is a great ding-dong, bringing out the best and worst in all of us. I am genuinely surprised by some of the vitriol spouting forth from our collective mouths. Politics or religion - can't beat it! Back to my real general election result and the BNP "multi-vote" argument. Remember we had this:

LD 16,806 43.7%
Con 14,116 36.7%
Lab 5,061 13.2%
UKIP 1,243 3.2%
BNP 1,213 3.2%

Let's say a voter, we'll call him Jono, goes into vote under the present system and votes BNP. He adds 1 vote to the BNP party and that's it, his election is over. 1213 votes is not enough to take the seat. At the declaration, his one vote is included in the BNP's grand total. Under AV, he votes like this: 1 BNP, 2 UKIP, 3 LAB, 4 LIB DEM and stops there because he can't bring himself to vote Conservative.

Counting starts. It is still 1213 votes for BNP (being last in this poll). His BNP vote does not go in the bin but is passed to the UKIP pile. The UKIP vote has now gone up 1. Jono has added 1 vote to BNP and 1 vote to UKIP. Under FPTP the only way this could happen is for one person to vote BNP and another person to vote UKIP. But we haven't stopped there. On the third count, his ballot paper in the UKIP pile is now added to Labour. So far, he has now incremented 3 parties by one vote each. Finally his ballot paper in the Labour pile gets passed to the Lib Dems adding 1 to their total. So his ballot paper has added one vote to four parties, which wil be reflected in the final tallies.

Now to me he has had the power of four voters. But I am keen to listen to arguments why this is not so.

[Jon O'Neill]

As far as the second round goes, BNP has no votes. It's not like there are 6 MPs in the end, all in order of how much power they have, is it.

[Jon O'Neill]

I am genuinely shocked that you have been hoodwinked.

[Charlie Reams]

This is a great ding-dong, bringing out the best and worst in all of us. I am genuinely surprised by some of the vitriol spouting forth from our collective mouths. Politics or religion - can't beat it! Back to my real general election result and the BNP "multi-vote" argument. Remember we had this:

LD 16,806 43.7%
Con 14,116 36.7%
Lab 5,061 13.2%
UKIP 1,243 3.2%
BNP 1,213 3.2%

Let's say a voter, we'll call him Jono, goes into vote under the present system and votes BNP. He adds 1 vote to the BNP party and that's it, his election is over. 1213 votes is not enough to take the seat. At the declaration, his one vote is included in the BNP's grand total. Under AV, he votes like this: 1 BNP, 2 UKIP, 3 LAB, 4 LIB DEM and stops there because he can't bring himself to vote Conservative.

Counting starts. It is still 1213 votes for BNP (being last in this poll). His BNP vote does not go in the bin but is passed to the UKIP pile. The UKIP vote has now gone up 1. Jono has added 1 vote to BNP and 1 vote to UKIP. Under FPTP the only way this could happen is for one person to vote BNP and another person to vote UKIP. But we haven't stopped there. On the third count, his ballot paper in the UKIP pile is now added to Labour. So far, he has now incremented 3 parties by one vote each. Finally his Labour vote gets added to the Lib Dems adding 1 to their total. So his ballot paper has added one vote to four parties, which wil be reflected in the final tallies.

Now to me he has had the power of four voters. But I am keen to listen to arguments why this is not so.

There are four separate rounds of voting. Everyone gets to vote four times, some changing their vote between rounds and some not. I don't see how you can argue that Jono has gained a great advantage in this system when now he's faced with having either his 4th or 5th preference elected.

[Chris Corby]

As far as the second round goes, BNP has no votes. It's not like there are 6 MPs in the end, all in order of how much power they have, is it.

You really don't understand AV, do you?

[Jon O'Neill]

As far as the second round goes, BNP has no votes. It's not like there are 6 MPs in the end, all in order of how much power they have, is it.

You really don't understand AV, do you?

I dunno, don't I? I feel like I do and you don't.

[Michael Wallace]

As far as the second round goes, BNP has no votes. It's not like there are 6 MPs in the end, all in order of how much power they have, is it.

You really don't understand AV, do you?

I dunno, don't I? I feel like I do and you don't.

Yeah, it seems that way to me. I'm genuinely interested to know what it is we're apparently missing, though.

[Liam Tiernan]

This is a great ding-dong, bringing out the best and worst in all of us. I am genuinely surprised by some of the vitriol spouting forth from our collective mouths. Politics or religion - can't beat it! Back to my real general election result and the BNP "multi-vote" argument. Remember we had this:

LD 16,806 43.7%
Con 14,116 36.7%
Lab 5,061 13.2%
UKIP 1,243 3.2%
BNP 1,213 3.2%

Let's say a voter, we'll call him Jono, goes into vote under the present system and votes BNP. He adds 1 vote to the BNP party and that's it, his election is over. 1213 votes is not enough to take the seat. At the declaration, his one vote is included in the BNP's grand total. Under AV, he votes like this: 1 BNP, 2 UKIP, 3 LAB, 4 LIB DEM and stops there because he can't bring himself to vote Conservative.Counting starts. It is still 1213 votes for BNP (being last in this poll). His BNP vote does not go in the bin but is passed to the UKIP pile. The UKIP vote has now gone up 1. Jono has added 1 vote to BNP and 1 vote to UKIP. Under FPTP the only way this could happen is for one person to vote BNP and another person to vote UKIP. But we haven't stopped there. On the third count, his ballot paper in the UKIP pile is now added to Labour. So far, he has now incremented 3 parties by one vote each. Finally his ballot paper in the Labour pile gets passed to the Lib Dems adding 1 to their total. So his ballot paper has added one vote to four parties, which wil be reflected in the final tallies.

Now to me he has had the power of four voters. But I am keen to listen to arguments why this is not so.

Wrong. Since the BNP candidate has now been eliminated, he still has one vote, which has transferred to the UKIP. Its's a simple system. We've had it here for 80 years. It's really not that difficult to understand.What I find difficult to understand is how a party with a 25% share of the vote nationally can end up with less than 4% of the seats.Nationally that adds up to a discrepancy of over 100 seats. And that is a real election result I'm quoting. Can anybody reasonably justify that?

[Peter Mabey]

Unfortunately it looks as if the NO people are likely to win as the result of their campaign of fiction (the imagined expensive voting machines) misrepresentation (pretending that an unrealistic distribution of preferences preferences would let in the BNP where they'd come nowhere under FPTP is probable) and a personal attack on Nick Clegg.

[David Williams]

I do agree that everyone gets the same number of votes, it's just that some people get to vote for the same person in several rounds, and some people have to vote for different people. But what about this.

Tom 48%, Dick 30%, Harriet 22%, and all Harriet's second preferences are for Dick. What about Tom and Dick's second preferences? It looks as if Dick and Harriet have split the anti-Tom faction, with disastrous and undemocratic consequences under FPTP. If so, we can assume that all Dick's supporters would put Harriet as second choice. Now suppose that Tom's supporters favour Harriet over Dick by three to one. In a straight fight against Tom either Dick or Harriet would win by 52% to 48%. But in a straight fight between Dick and Harriet, Harriet wins by 58% to 42%. So shouldn't Harriet win under a properly devised AV system?

The most fun of all is if Dick's 30% splits 26 to 4 in favour of Harriet over Tom, and Tom's 48% splits 30/18 in favour of Harriet over Dick.

[Charlie Reams]

I do agree that everyone gets the same number of votes, it's just that some people get to vote for the same person in several rounds, and some people have to vote for different people. But what about this.

Tom 48%, Dick 30%, Harriet 22%, and all Harriet's second preferences are for Dick. What about Tom and Dick's second preferences? It looks as if Dick and Harriet have split the anti-Tom faction, with disastrous and undemocratic consequences under FPTP. If so, we can assume that all Dick's supporters would put Harriet as second choice. Now suppose that Tom's supporters favour Harriet over Dick by three to one. In a straight fight against Tom either Dick or Harriet would win by 52% to 48%. But in a straight fight between Dick and Harriet, Harriet wins by 58% to 42%. So shouldn't Harriet win under a properly devised AV system?

The most fun of all is if Dick's 30% splits 26 to 4 in favour of Harriet over Tom, and Tom's 48% splits 30/18 in favour of Harriet over Dick.

This is a reasonable criticism of AV although FPTP doesn't do any better in this case. Nevertheless those are the two options we're faced with here, and I think it's fair to assume that if AV gets rejected then we'll never get a referendum on further alternatives. And although it would be nice to solve all of these problems, that's also logically impossible, so there will always be some scenarios in which a given system behaves undesirably.

[Martin Bishop]

This is a great ding-dong, bringing out the best and worst in all of us. I am genuinely surprised by some of the vitriol spouting forth from our collective mouths. Politics or religion - can't beat it! Back to my real general election result and the BNP "multi-vote" argument. Remember we had this:

LD 16,806 43.7%
Con 14,116 36.7%
Lab 5,061 13.2%
UKIP 1,243 3.2%
BNP 1,213 3.2%

Let's say a voter, we'll call him Jono, goes into vote under the present system and votes BNP. He adds 1 vote to the BNP party and that's it, his election is over. 1213 votes is not enough to take the seat. At the declaration, his one vote is included in the BNP's grand total. Under AV, he votes like this: 1 BNP, 2 UKIP, 3 LAB, 4 LIB DEM and stops there because he can't bring himself to vote Conservative.

Counting starts. It is still 1213 votes for BNP (being last in this poll). His BNP vote does not go in the bin but is passed to the UKIP pile. The UKIP vote has now gone up 1. Jono has added 1 vote to BNP and 1 vote to UKIP. Under FPTP the only way this could happen is for one person to vote BNP and another person to vote UKIP. But we haven't stopped there. On the third count, his ballot paper in the UKIP pile is now added to Labour. So far, he has now incremented 3 parties by one vote each. Finally his ballot paper in the Labour pile gets passed to the Lib Dems adding 1 to their total. So his ballot paper has added one vote to four parties, which wil be reflected in the final tallies.

Now to me he has had the power of four voters. But I am keen to listen to arguments why this is not so.

Let's say a voter, we'll call him Chris, goes into vote under the present system and votes Lib Dem. He adds 1 vote to the Lib Dem party and that's it, his election is over. 16,806 votes is enough to take the seat. At the declaration, his one vote is included in the Lib Dem's grand total. Under AV, he votes like this: 1 Lib Dem, 2 Conservative, 3 LAB, 4 UKIP and stops there because he can't bring himself to vote BNP.

Counting starts. 1213 votes go to the BNP (being last in this poll) and they are eliminated. The BNP votes do not go in the bin, their second preferences are now counted. The votes are counted again. Chris has now voted twice for the Lib Dems. Under FPTP the only way this could happen is for one person to vote Lib Dem and another person to vote Lib Dem as well. But we haven't stopped there. On the third count, the ballot papers in the UKIP pile have their next preferences counted. So far, Chris has voted for the Lib Dems three times. Finally, the Labour pile gets distributed to Lib Dems and Conservtives. So his ballot paper has been counted as a Lib dem vote four times, which wil be reflected in the final tallies.

[Mike Brown]

And although it would be nice to solve all of these problems, that's also logically impossible, so there will always be some scenarios in which a given system behaves undesirably.

This is exactly the sort of thing I love about your posts, Charlie. Either you know everything, or you're extremely good at googling.

[Chris Corby]

Wrong. Since the BNP candidate has now been eliminated, he still has one vote, which has transferred to the UKIP. Its's a simple system. We've had it here for 80 years. It's really not that difficult to understand.

Really? I thought Ireland had the STV (Single Transferrable Vote) system which is simliar to AV but not the same.

[Charlie Reams]

But I am keen to listen to arguments why this is not so.

So keen that you've ignored at least three posts explaining exactly why this is not so!

And although it would be nice to solve all of these problems, that's also logically impossible, so there will always be some scenarios in which a given system behaves undesirably.

This is exactly the sort of thing I love about your posts, Charlie. Either you know everything, or you're extremely good at googling.

Haha, well thanks very much. I did some reading about voting theory at some point in the distant past and the impressive generality of that result stuck with me. I had a feeling I'd linked it here before and... yep, guilty as charged.

[Chris Corby]

There are four separate rounds of voting. Everyone gets to vote four times, some changing their vote between rounds and some not. I don't see how you can argue that Jono has gained a great advantage in this system when now he's faced with having either his 4th or 5th preference elected.

Everyone? The Conservatives in this example don't get even a second chance!

[Charlie Reams]

There are four separate rounds of voting. Everyone gets to vote four times, some changing their vote between rounds and some not. I don't see how you can argue that Jono has gained a great advantage in this system when now he's faced with having either his 4th or 5th preference elected.

Everyone? The Conservatives in this example don't get even a second chance!

Yes, everyone. If you voted for Conservatives in the first round, you can vote for them again in the second round.

[Jon O'Neill]

And although it would be nice to solve all of these problems, that's also logically impossible, so there will always be some scenarios in which a given system behaves undesirably.

This is exactly the sort of thing I love about your posts, Charlie. Either you know everything, or you're extremely good at googling.

He's actually a Wikipaedophile.

[Liam Tiernan]

Wrong. Since the BNP candidate has now been eliminated, he still has one vote, which has transferred to the UKIP. Its's a simple system. We've had it here for 80 years. It's really not that difficult to understand.

Really? I thought Ireland had the STV (Single Transferrable Vote) system which is simliar to AV but not the same.

When STV is used for single-winner elections, it is equivalent to the non-proportional instant-runoff voting (alternative vote) method

Wrong again. It's how our Presidential elections are decided.



Edit:
Maybe another way to solve the argument about votes being counted multiple times is to say that there is no point in the counting process where your vote has gone to two people who are still in the running.

[Chris Corby]

This is a great ding-dong, bringing out the best and worst in all of us. I am genuinely surprised by some of the vitriol spouting forth from our collective mouths. Politics or religion - can't beat it! Back to my real general election result and the BNP "multi-vote" argument. Remember we had this:

LD 16,806 43.7%
Con 14,116 36.7%
Lab 5,061 13.2%
UKIP 1,243 3.2%
BNP 1,213 3.2%

Let's say a voter, we'll call him Jono, goes into vote under the present system and votes BNP. He adds 1 vote to the BNP party and that's it, his election is over. 1213 votes is not enough to take the seat. At the declaration, his one vote is included in the BNP's grand total. Under AV, he votes like this: 1 BNP, 2 UKIP, 3 LAB, 4 LIB DEM and stops there because he can't bring himself to vote Conservative.

Counting starts. It is still 1213 votes for BNP (being last in this poll). His BNP vote does not go in the bin but is passed to the UKIP pile. The UKIP vote has now gone up 1. Jono has added 1 vote to BNP and 1 vote to UKIP. Under FPTP the only way this could happen is for one person to vote BNP and another person to vote UKIP. But we haven't stopped there. On the third count, his ballot paper in the UKIP pile is now added to Labour. So far, he has now incremented 3 parties by one vote each. Finally his ballot paper in the Labour pile gets passed to the Lib Dems adding 1 to their total. So his ballot paper has added one vote to four parties, which wil be reflected in the final tallies.

Now to me he has had the power of four voters. But I am keen to listen to arguments why this is not so.

Let's say a voter, we'll call him Chris, goes into vote under the present system and votes Lib Dem. He adds 1 vote to the Lib Dem party and that's it, his election is over. 16,806 votes is enough to take the seat. At the declaration, his one vote is included in the Lib Dem's grand total. Under AV, he votes like this: 1 Lib Dem, 2 Conservative, 3 LAB, 4 UKIP and stops there because he can't bring himself to vote BNP.

Counting starts. 1213 votes go to the BNP (being last in this poll) and they are eliminated. The BNP votes do not go in the bin, their second preferences are now counted. The votes are counted again. Chris has now voted twice for the Lib Dems. Under FPTP the only way this could happen is for one person to vote Lib Dem and another person to vote Lib Dem as well. But we haven't stopped there. On the third count, the ballot papers in the UKIP pile have their next preferences counted. So far, Chris has voted for the Lib Dems three times. Finally, the Labour pile gets distributed to Lib Dems and Conservtives. So his ballot paper has been counted as a Lib dem vote four times, which wil be reflected in the final tallies.

What? Are you winding me up Martin? Chris the voter under AV has cast his vote for the Lib Dems as his first preference. They have topped the poll. His ballot paper is not counted again. The only ones added to this second stage are the ones where the first preference was BNP. Their second preference is added to the remaining candidates. After that. still not 50% for any one candidate so the second preference of the UKIP party is added to the remaing candidates. If some of them voted BNP as their second, then their third choice is taken instead. Still no 50%, so now the Labour vote is taken to add votes to the two remaining candidates, Lib Dem and Conservative. If anyone one voted BNP or UKIP on their papers, their fourth choice is taken (if indicated, otherwise the ballot paper is ignored). You are now left with two candidates, the Lib Dems and the Conseravtives - one of whom now has over 50% and is declared the winner.

[Chris Corby]

There are four separate rounds of voting. Everyone gets to vote four times, some changing their vote between rounds and some not. I don't see how you can argue that Jono has gained a great advantage in this system when now he's faced with having either his 4th or 5th preference elected.

Everyone? The Conservatives in this example don't get even a second chance!

Yes, everyone. If you voted for Conservatives in the first round, you can vote for them again in the second round.

I refer the honourable gentleman to my previous answer.

Anyway, I am off to the Lake District for a few days now so I will be back after the election.

At least it's all a lot clearer to you now

[Charlie Reams]

His ballot paper is not counted again.

Yes it is.

[Chris Corby]

His ballot paper is not counted again.

Yes it is.

No it's not. The original Lib Dem vote was 16,806. This includes Chris' original vote. This total of 16,806 is now added to and increased by votes from BNP, UKIP and Labour voters. Chris is not part of this. He therefore only voted once. I need to go......

[Charlie Reams]

His ballot paper is not counted again.

Yes it is.

No it's not

I'll be charitable and assume you're genuinely finding this hard to understand, rather than being obstinate.

So, if you prefer, you can think of the election being re-run after each lowest party is eliminated. People who voted for the eliminated party would change to another one; everyone else, of course, still prefers the same party they did in the previous round. At the end, everyone has cast the same number of votes -- one in each election. This produces exactly the same result as AV, and AV doesn't require the hassle of actual re-runs (and also limits the tactical voting that can occur in real multi-stage elections, qv the French system).

[Martin Bishop]

This is a great ding-dong, bringing out the best and worst in all of us. I am genuinely surprised by some of the vitriol spouting forth from our collective mouths. Politics or religion - can't beat it! Back to my real general election result and the BNP "multi-vote" argument. Remember we had this:

LD 16,806 43.7%
Con 14,116 36.7%
Lab 5,061 13.2%
UKIP 1,243 3.2%
BNP 1,213 3.2%

Let's say a voter, we'll call him Jono, goes into vote under the present system and votes BNP. He adds 1 vote to the BNP party and that's it, his election is over. 1213 votes is not enough to take the seat. At the declaration, his one vote is included in the BNP's grand total. Under AV, he votes like this: 1 BNP, 2 UKIP, 3 LAB, 4 LIB DEM and stops there because he can't bring himself to vote Conservative.

Counting starts. It is still 1213 votes for BNP (being last in this poll). His BNP vote does not go in the bin but is passed to the UKIP pile. The UKIP vote has now gone up 1. Jono has added 1 vote to BNP and 1 vote to UKIP. Under FPTP the only way this could happen is for one person to vote BNP and another person to vote UKIP. But we haven't stopped there. On the third count, his ballot paper in the UKIP pile is now added to Labour. So far, he has now incremented 3 parties by one vote each. Finally his ballot paper in the Labour pile gets passed to the Lib Dems adding 1 to their total. So his ballot paper has added one vote to four parties, which wil be reflected in the final tallies.

Now to me he has had the power of four voters. But I am keen to listen to arguments why this is not so.

Let's say a voter, we'll call him Chris, goes into vote under the present system and votes Lib Dem. He adds 1 vote to the Lib Dem party and that's it, his election is over. 16,806 votes is enough to take the seat. At the declaration, his one vote is included in the Lib Dem's grand total. Under AV, he votes like this: 1 Lib Dem, 2 Conservative, 3 LAB, 4 UKIP and stops there because he can't bring himself to vote BNP.

Counting starts. 1213 votes go to the BNP (being last in this poll) and they are eliminated. The BNP votes do not go in the bin, their second preferences are now counted. The votes are counted again. Chris has now voted twice for the Lib Dems. Under FPTP the only way this could happen is for one person to vote Lib Dem and another person to vote Lib Dem as well. But we haven't stopped there. On the third count, the ballot papers in the UKIP pile have their next preferences counted. So far, Chris has voted for the Lib Dems three times. Finally, the Labour pile gets distributed to Lib Dems and Conservtives. So his ballot paper has been counted as a Lib dem vote four times, which wil be reflected in the final tallies.

What? Are you winding me up Martin? Chris the voter under AV has cast his vote for the Lib Dems as his first preference. They have topped the poll. His ballot paper is not counted again. The only ones added to this second stage are the ones where the first preference was BNP. Their second preference is added to the remaining candidates. After that. still not 50% for any one candidate so the second preference of the UKIP party is added to the remaing candidates. If some of them voted BNP as their second, then their third choice is taken instead. Still no 50%, so now the Labour vote is taken to add votes to the two remaining candidates, Lib Dem and Conservative. If anyone one voted BNP or UKIP on their papers, their fourth choice is taken (if indicated, otherwise the ballot paper is ignored). You are now left with two candidates, the Lib Dems and the Conseravtives - one of whom now has over 50% and is declared the winner.

Here goes.

Let's say 100 of the BNP voters put the Lib Dems as second preference, 613 UKIP, 300 Conservative and 200 Labour.

In your argument, we only count the original BNP votes in round 2, so UKIP win with a majority of all the votes counted.

In reality, we count all the original ballot papers again, including those who didn't vote BNP. So in my example round 2 ends up with the following.

LD 16,806 + 100 = 17,906
Con 14,116 + 300 = 14,416
Lab 5,061 + 200 = 5,261
UKIP 1,243 + 613 = 1,856

Chris's vote is included in those listed in bold.

[Peter Mabey]

BTW. when MPs elect their own leader they don't use FPTP, so even they don't think it's the best system

[Liam Tiernan]

Wrong. Since the BNP candidate has now been eliminated, he still has one vote, which has transferred to the UKIP. Its's a simple system. We've had it here for 80 years. It's really not that difficult to understand.

Really? I thought Ireland had the STV (Single Transferrable Vote) system which is simliar to AV but not the same.

When STV is used for single-winner elections, it is equivalent to the non-proportional instant-runoff voting (alternative vote) method

Wrong again. It's how our Presidential elections are decided.



Edit:
Maybe another way to solve the argument about votes being counted multiple times is to say that there is no point in the counting process where your vote has gone to two people who are still in the running.

This is a breakdown of the share of seats vs. the share of first preference votes, for the Irish general election of 1992:
Party % of seats % FPv
Fianna Fáil 40.9 39.1
Fine Gael 27.1 24.5
Labour Party 19.8 19.5
Progressive Democrats 6.0 4.7
Democratic Left 2.4 2.8
Green Party 0.6 1.4

This is a breakdown of the share of seats vs. share of votes in the British general election in the same year:
Party % of seats % vote
Conservative 51.6 41.9
Labour 41.6 34.4
LibDem 3.07 17.8


Again, the year was chosen at random, but gives a pretty good example of the intrinsic unfairness of FPTP. I guess when David Cameron talks about traditional British values, fair play isn't one of them.

[Gavin Chipper]

I do agree that everyone gets the same number of votes, it's just that some people get to vote for the same person in several rounds, and some people have to vote for different people. But what about this.

Tom 48%, Dick 30%, Harriet 22%, and all Harriet's second preferences are for Dick. What about Tom and Dick's second preferences? It looks as if Dick and Harriet have split the anti-Tom faction, with disastrous and undemocratic consequences under FPTP. If so, we can assume that all Dick's supporters would put Harriet as second choice. Now suppose that Tom's supporters favour Harriet over Dick by three to one. In a straight fight against Tom either Dick or Harriet would win by 52% to 48%. But in a straight fight between Dick and Harriet, Harriet wins by 58% to 42%. So shouldn't Harriet win under a properly devised AV system?

The most fun of all is if Dick's 30% splits 26 to 4 in favour of Harriet over Tom, and Tom's 48% splits 30/18 in favour of Harriet over Dick.

For ranked preferences with a single winner, I would regard Kemeny-Young as probably the best method. It is a Condorcet method which means that if there is a candidate that would beat all others in a head to head, then that candidate is guaranteed to win. With your example, it could be Tom 34%, Dick 34% and Harriet 32% - not much in it at all really - but all of Tom and Dick's supporters put Harriet second. Harriet would win each head to head by nearly 2:1 but lose under AV (and FPTP obviously).

This is a reasonable criticism of AV although FPTP doesn't do any better in this case. Nevertheless those are the two options we're faced with here, and I think it's fair to assume that if AV gets rejected then we'll never get a referendum on further alternatives. And although it would be nice to solve all of these problems, that's also logically impossible, so there will always be some scenarios in which a given system behaves undesirably.

Actually Arrow's Theorem only works for ranked preferences so Range Voting (where scores are given rather than ranks) bypasses it (although it is not without other problems).

[Ryan Taylor]

And although it would be nice to solve all of these problems, that's also logically impossible, so there will always be some scenarios in which a given system behaves undesirably.

This is exactly the sort of thing I love about your posts, Charlie. Either you know everything, or you're extremely good at googling.

Agreed. I learn so much stuff from posts like this although probably can't remember much of it. Pretty cool though. So why haven't you been on UC?

[Charlie Reams]

For ranked preferences with a single winner, I would regard Kemeny-Young as probably the best method.

Haven't seen this before. Pretty nice. Has it ever been used for a real election?

Actually Arrow's Theorem only works for ranked preferences so Range Voting (where scores are given rather than ranks) bypasses it (although it is not without other problems).

I'm quite a fan of range voting as GotW might suggest, but if we're already telling people they're too stupid to count 1-5 then imagine having to pick the numbers without prompting. Clearly far too difficult.

So why haven't you been on UC?

Assuming that's at me and not Mike, it's basically because there are too many things I don't know about, and also it sounds like hard work.

[Gavin Chipper]

For ranked preferences with a single winner, I would regard Kemeny-Young as probably the best method.

Haven't seen this before. Pretty nice. Has it ever been used for a real election?

Don't think so actually.

I'm quite a fan of range voting as GotW might suggest, but if we're already telling people they're too stupid to count 1-5 then imagine having to pick the numbers without prompting. Clearly far too difficult.

You may be right. But there's also Approval Voting which is the simplest form of Range Voting, where you just vote for as many candidates as you want. I think it has some merit, and can be modified into a more proportional system.

[David Williams]

I find it hard to see why there is such a difference of opinion about something so seemingly straightforward as to whether some people get more votes than others under AV. If you view the process as a series of elections with one person eliminated each time everyone gets the same number of votes, one each time. Seems obvious to me. But is the other camp arguing that at the end of the process the second (and maybe third and fourth) choices of the supporters of candidates eliminated in the earlier rounds are counted, but the second choices of the supporters of the candidate who comes second are totally ignored? That's not actually how I read what's being said, but it's a fact. If your guy isn't going to win, better for you if he comes third rather than second.

[Charlie Reams]

I find it hard to see why there is such a difference of opinion about something so seemingly straightforward as to whether some people get more votes than others under AV. If you view the process as a series of elections with one person eliminated each time everyone gets the same number of votes, one each time. Seems obvious to me.

I think calling it a "difference of opinion" is generous. No one gets more votes than anyone else under AV. There are some ways of presenting it in which it appears that some votes count more than others, but these are really analogous to the missing pound riddle or something like that. It's a trick. The worst thing is that it's deliberately misleading propaganda, although I must say I think it's pretty cool how few of the voters in the C4C poll have fallen for it. Shame the same can't be said at the national level.

But is the other camp arguing that at the end of the process the second (and maybe third and fourth) choices of the supporters of candidates eliminated in the earlier rounds are counted, but the second choices of the supporters of the candidate who comes second are totally ignored? That's not actually how I read what's being said, but it's a fact. If your guy isn't going to win, better for you if he comes third rather than second.

I don't think that's the point that Chris is making, and I know that responding to it is likely to give him more ammunition to keep changing the subject. If you like you can imagine a final round of voting in which the everyone's vote gets transferred to the eventual winner, and thereby everyone's second choice counts (except people who are presumably satisfied anyway since the winner was their first choice).

If that seems like pedantry (which it probably does), a better way to think about it is that with AV it's not as simple as whether your guy wins or loses. If "your guy" (presumably meaning your first choice) comes second (in the penultimate round) and gets into the final two, he's then backed by everyone who prefers him to his opponent -- and if he can't win that straight head-to-head then presumably you can't complain at that point. On the other hand if he comes third then he's eliminated, which of course is bad news for you -- but now your vote gets transferred to your preference among the remaining two, which is better for you than having essentially no say in the final head-to-head as would happen under FPTP.

[Clive Brooker]

If your guy isn't going to win, better for you if he comes third rather than second.

I think this is one of the arguments used by the "No" camp (in their more sophisticated moments) to show that tactical voting lives on under AV, but is more complicated. All the examples I've seen suggest to me that you can't vote tactically under AV without 20-20 hindsight or perfect polling, so you're better off just going with what you believe. Please contradict me if I'm wrong!

Just for a moment I was wavering: if I find I would have made a difference by voting tactically under AV, but could never have known it, won't I be incredibly frustrated? But then I realised that under the current system it's far worse seeing people failing to vote tactically when the possibility is staring them in the face.

[Gavin Chipper]

If your guy isn't going to win, better for you if he comes third rather than second.

It's not even as simple as that. If your guy isn't going to win and you also prefer the candidate who he might come third to to the one who would win against your guy in the final round, and also if that other guy would actually win in a head to head against the candidate who would win if your guy made the final round, then it is better for your guy to come third rather than second. Simple.

[Rhys Benjamin]

Leaflet from the No camp through our door:

AV would lead to more hung parliaments, backroom deals and broken promises.

On Thursday May 5th,
Vote No to AV

Vote NO to the Alternative Vote on Thursday because......

--> It will produce more coalitions. Under the Alternative Vote system, we would have coalitions most of the time, with Nick Clegg deciding who would be Prime Minister by cutting a deal behind closed doors after the election.

--> It is used by only 3 other countries in the world - Fiji, Australia and Papua New Guinea - and Australia want to get rid of it.

--> It allows the second or third placed candidate to win. We would end up with third-best candidates becoming MPs.

--> It will cost the country £250 million, at a time when money is tight.

--> It means that someone else's 5th preference is worth the same as your 1st preference.

--> It will mean that supporters of the BNP and other fringe parties would decide who wins, because they will be eliminated first and then their votes could be counted again and again for other parties. That will encourage other candidates to pander to the likes of the BNP.

Remember the core principle in our democracy: every person gets an equal vote and the candidate with most votes wins.
Defend equal votes by voting No on Thursday.

http://www.no2av.org

[Charlie Reams]

Well, you certainly win on formatting. However,

--> It will produce more coalitions. Under the Alternative Vote system, we would have coalitions most of the time, with Nick Clegg deciding who would be Prime Minister by cutting a deal behind closed doors after the election.

Disingenuous.

--> It is used by only 3 other countries in the world - Fiji, Australia and Papua New Guinea - and Australia want to get rid of it.

Irrelevant.

--> It allows the second or third placed candidate to win. We would end up with third-best candidates becoming MPs.

Wrong.

--> It will cost the country £250 million, at a time when money is tight.

Wrong.

--> It means that someone else's 5th preference is worth the same as your 1st preference.

Wrong.

--> It will mean that supporters of the BNP and other fringe parties would decide who wins, because they will be eliminated first and then their votes could be counted again and again for other parties. That will encourage other candidates to pander to the likes of the BNP.

Wrong.

Remember the core principle in our democracy: every person gets an equal vote and the candidate with most votes wins.

Circular argument.


Thanks for playing!

[Ryan Taylor]

Has anyone on this forum ever thought of going into politics? There are definitely a few who I think would be good politicians.

[Rhys Benjamin]

Me

[Martin Bishop]

--> It is used by only 3 other countries in the world - Fiji, Australia and Papua New Guinea - and Australia want to get rid of it.

On the flip side of that, we are the only country in the EU that still uses FPTP.

[Ryan Taylor]

Me

You weren't amongst the people who I thought could be politicians.

[Mark James]

Has anyone on this forum ever thought of going into politics? There are definitely a few who I think would be good politicians.

Me

I think Billy Connolly said it best when he said the desire to become a politician should automatically exclude you from ever becoming one.

[Liam Tiernan]

Me

You weren't amongst the people who I thought could be politicians.

Actually, I think Rhys is good politician material. A successful politician needs to be impervious to criticism ( or, in Rhys' case, oblivious).He would also need to have supreme confidence in his own abilities, the capacity to ignore inconvenient facts , and a strong right arm, capable of flogging a dead horse until it moves, qualities that Rhys has in abundance.

[Graeme Cole]

For those who still think AV gives some people more votes than others, here's a little cartoon which hopefully clears things up (click for bigger)...

[Ian Volante]

Has anyone on this forum ever thought of going into politics? There are definitely a few who I think would be good politicians.

Me

I think Billy Connolly said it best when he said the desire to become a politician should automatically exclude you from ever becoming one.

That was my recent argument for retaining the monarchy.

[David Williams]

Interesting.

The stated objective is to find a leader that the majority are happy with, but all that has actually been proved is that the majority are less unhappy with Alice than Bob. Charlie's supporters have not said they are happy with Alice. It is quite possible that the majority are unhappy with Alice as leader, but would be happy with Bob or Charlie as leader.

But everyone gets the same number of votes.

[Charlie Reams]

The stated objective is to find a leader that the majority are happy with

Yeah, that's not really how I'd state it. I mean, it might be that everyone is unhappy with all of them. No voting system is going to resolve that.

[Clive Brooker]

Sorry, Graeme, but I'm marginally less happy after seeing that cartoon. It looks like an Exhaustive Vote (right term?) scenario rather that AV, where everyone gets a chance to vote afresh in each round in the light of what's happened before. I think it's a lot easier to see that everyone gets the same number of votes under those rules.

The way I prefer to look at AV is to say that all votes from earlier rounds are completely discounted. It's as though previous rounds never happened, and the election is starting again with one fewer candidate.

But I'm becoming increasingly worried about the votes cast for the guy who eventually loses the final two-horse run off (if it comes to that), whose second preferences are never considered. If the Exhaustive Vote is what AV would like to be but can't be for practical reasons, then the cartoon (to me anyway) does rather highlight the fact that AV can't quite do the same job.

I'm sure all this has been covered many times further up the thread, so sorry again.

[Charlie Reams]

Sorry, Graeme, but I'm marginally less happy after seeing that cartoon. It looks like an Exhaustive Vote (right term?) scenario rather that AV, where everyone gets a chance to vote afresh in each round in the light of what's happened before. I think it's a lot easier to see that everyone gets the same number of votes under those rules.

If you were happy with Candidate X in Round N then why would you suddenly stop being happy with them in Round N+1? There's no legitimate reason to change from someone who wasn't eliminated. AV is exactly like an exhaustive vote except that you can't vote tactically using information from the previous rounds, so it's strictly better.

But I'm becoming increasingly worried about the votes cast for the guy who eventually loses the final two-horse run off (if it comes to that), whose second preferences are never considered. If the Exhaustive Vote is what AV would like to be but can't be for practical reasons, then the cartoon (to me anyway) does rather highlight the fact that AV can't quite do the same job.

It's getting quite hard to follow this thread now, but see the second half of this post. In short: their second preferences aren't considered because their first pick was given every possible chance to win, and lost in a straight head-to-head with the overall winner.

[Clive Brooker]

If you were happy with Candidate X in Round N then why would you suddenly stop being happy with them in Round N+1? There's no legitimate reason to change from someone who wasn't eliminated. AV is exactly like an exhaustive vote except that you can't vote tactically using information from the previous rounds, so it's strictly better.

I presume that by "legitimate" you mean non-tactical? Clearly there's potentially a benefit in switching away even from the first-placed candidate if the pattern of voting suggests he/she is ultimately doomed.

I maintain that it's much more intuitively obvious that everyone has an equal voice in each round under an exhaustive vote (which is what the cartoon depicts) than under AV. Maybe the confusion with AV arises because people think a second preference is a fresh choice, thereby mixing up elements of AV and exhaustive voting in a way that certainly would be unfair.

[Phil Reynolds]

The stated objective is to find a leader that the majority are happy with

Yeah, that's not really how I'd state it. I mean, it might be that everyone is unhappy with all of them. No voting system is going to resolve that.

I'm reminded yet again (as I have been several times while following this thread) of the words of W.C. Fields: "I never vote for anybody. I always vote against."

[Charlie Reams]

If you were happy with Candidate X in Round N then why would you suddenly stop being happy with them in Round N+1? There's no legitimate reason to change from someone who wasn't eliminated. AV is exactly like an exhaustive vote except that you can't vote tactically using information from the previous rounds, so it's strictly better.

I presume that by "legitimate" you mean non-tactical?

Yep, exactly.

I maintain that it's much more intuitively obvious that everyone has an equal voice in each round under an exhaustive vote (which is what the cartoon depicts) than under AV. Maybe the confusion with AV arises because people think a second preference is a fresh choice, thereby mixing up elements of AV and exhaustive voting in a way that certainly would be unfair.

Probably, yep. The AV crowd haven't done a great job of explaining it either, and they're faced with opposition who are apparently unembarrassed to say "don't worry guys, you're too dumb to understand this!"