First Past the Post or AV

[Loz]

westdulwich Wrote:

-------------------------------------------------------

> Loz, please tell me - in the example above, do you

> think that B winning is the fairest outcome? (I've

> given my reasons above why I think it isn't). Do

> you think that B winning adequately reflects the

> preference of the voters? Really interested to

> hear from any in the Yes camp on this.

What you are doing here is not actually looking at AV - you are looking at various possibilities that *might* be an alternative to AV. Which is good in a way as you are accepting that electoral reform is A Good Thing. But, the question tomorrow has two options: AV or FPTP, so in your example it is rather irrelevant if you think that Party C has gotten a raw deal. That's not on offer.

What you need to look at is this: which is the more deserving winner in your example: Party A or Party B? It is fairly obvious that between those two, Party B has the legitimacy and the majority support recognition under AV that Party A cannot demonstrate under FPTP.

When you compare AV to FPTP, AV delivers a much more democratic solution. Yes, there might be better systems out there, but we only have two choices tomorrow. If you say no tomorrow you are saying no to any electoral reform.

[westdulwich]

Huguenot Wrote:

-------------------------------------------------------

> There seems to be a group of people who treat

> voting like football teams, where the winner of an

> election is the one that scores the highest number

> of points regardless of whether this is a majority

> or not.

>

> This is not democracy. Democracy is about

> identifying the candidates that have the majority

> support of the electorate.

From your post, it appears that you are assuming that receiving the majority of votes counted in a given round (perhaps after other earlier preferences have been discounted, and other second or subsequent preferences have been ignored) is the same as having the support of the majority of the electorate. This is not the case, and sometimes (probably not often) it is possible to achieve the former without having the latter. The example that I gave above demonstrates this. By using AV, and thereby being selective about which preferences are counted and when, B will win. However, the majority did not want B. Based on first preferences of all voters B would not win (A comes out on top). Taking into account what voters would be willing to accept as a compromise if they cannot have their first choice, B would not win (C comes out on top).

> AV is not about reduced percentage vote value or

> ridiculous 'no' camp fabricated maths, it's about

> weeding out unpopular candidates until only two

> remain and then asking the electorate to vote on

> these final two.

I'm not sure who your post is addressed to. It was posted after mine, so I'm assuming you're responding to me. On the other hand you also appear to be addressing the "no camp", so forgive me if this was meant for someone else and I'm responding out of turn. I don't understand what "fabricated maths" means. Are you suggesting that my example doesn't stack up mathematically? That it doesn't depict a plausible set of circumstances or outcomes? That the conclusions drawn are inappropriate? Of course, if you're referring to some other argument that the "no camp" have advanced and this wasn't meant for me then just ignore this.

> You're not struggling to understand a run off are

> you?

>

> AV allows you to do this without returning to the

> polls.

I'm not struggling to understand a run-off. I understand the concept, and the different systems through which this might be delivered. I'm not sure I understand your line of reasoning though. When I think of a run off, I usually think of a system with multiple rounds of voting, in each of which all voters are allowed to cast a vote (changing their preference from round to round if they so choose). I do appreciate that there are other run-off voting systems in which this is not the case. Can you please tell me:

- is this what you have in mind when you refer to a run-off?

- if so, do you think that AV offers the same "voter experience" (same range of option, same treatment of preferences, etc.) and would always deliver the same outcomes?

- if you think there are differences in the experience offered to voters, or in the outcomes each system delivers, which (if any) do you think is better?

[westdulwich]

Loz Wrote:

-------------------------------------------------------

> westdulwich Wrote:

> --------------------------------------------------

> -----

>

> > Loz, please tell me - in the example above, do

> you

> > think that B winning is the fairest outcome?

> (I've

> > given my reasons above why I think it isn't).

> Do

> > you think that B winning adequately reflects

> the

> > preference of the voters? Really interested to

> > hear from any in the Yes camp on this.

>

> What you are doing here is not actually looking at

> AV - you are looking at various possibilities that

> *might* be an alternative to AV. Which is good in

> a way as you are accepting that electoral reform

> is A Good Thing. But, the question tomorrow has

> two options: AV or FPTP, so in your example it is

> rather irrelevant if you think that Party C has

> gotten a raw deal. That's not on offer.

>

> What you need to look at is this: which is the

> more deserving winner in your example: Party A or

> Party B? It is fairly obvious that between those

> two, Party B has the legitimacy and the majority

> support recognition under AV that Party A cannot

> demonstrate under FPTP.

> When you compare AV to FPTP, AV delivers a much

> more democratic solution. Yes, there might be

> better systems out there, but we only have two

> choices tomorrow. If you say no tomorrow you are

> saying no to any electoral reform.

Thank you for the response. I think you misunderstand me. I am not asking whether (in relation to my example) AV provides a better outcome than FPTP, or what kind of system would deliver the best outcome. Rather, I am asking whether AV (by making B the winner) would have:

- delivered the fairest outcome? (or even a fair outcome)

- reflected the preference of the voters?

Would you mind answering this question?

[Loz]

Thank you for the response. I think you misunderstand me. I am not asking whether (in relation to my example) AV provides a better outcome than FPTP, or what kind of system would deliver the best outcome. Rather, I am asking whether AV (by making B the winner) would have:

- delivered the fairest outcome? (or even a fair outcome)

- reflected the preference of the voters?

Would you mind answering this question?

Did it deliver the fairest outcome? Compared to FPTP (i.e. naming Party A the winner) then, yes, absolutely.

Reflected the preference of the voters? Again, compared to FPTP, then yes, of course, as FPTP ignores any preferences of the voter except their first choice.

AV is often referred as IRV - Instant Runoff Voting. The concept is the same except, as you've noted, those whose preferences are still 'live' as assumed not to change their votes between rounds. It's a tradeoff for the sake of simplicity, speed and cost. That is why examining preferences of live votes is a bit of a red herring - they don't actually 'exist' until needed. It's bit like asking the voters in your 'true' runoff in round 2 which way they are thinking of voting in round 3, 4, etc, and trying to take that into acoount.

As I said, examining other options is no bad thing, but the choice is between AV and FPTP. Which do you think is better and fairer?

[Alan Dale]

http://www.youtube.com/watch?v=tT9rEZqvSu8&feature=player_embedded

[Loz]

Alan Dale Wrote:

-------------------------------------------------------

> http://www.youtube.com/watch?v=tT9rEZqvSu8&feature=player_embedded

Nice try Alan, but the voting system that they used was not AV, so your point is, well, pointless. Had they used AV, then the 'unwanted' yellow would have gone out on the first round.

Score one for AV. :)-D

[Alan Dale]

Is it possible with AV to have the winner that is no one's first choice?

[Marmora Man]

Alan Dale Wrote:

-------------------------------------------------------

> Is it possible with AV to have the winner that is

> no one's first choice?

NO - to be considered the candidate must have polled at least 1 vote more than the candidate with the least votes in order to take part in what is now being termed a "run off" with the 2nd pref votes of the least preferred candidate being apportioned to the remaining candidates.

[Loz]

No, impossible. Even if more than one candidate finished on zero votes, then there would be no votes to redistribute so a candidate with zero first prefs would always be eliminated.

Good question! You had me thinking about various permutations and possibilities until the obvious logical answer came to me.

- edit to say... dammit! MM beat me by two mins!

[Alan Dale]

So could it be a single voter who supports the eventual winner in the first round?

[Loz]

Do you mean can a person with one first preference win the election? Nope. As his/her only vote would be the first vote redistributed. (Unless, of course there is only one voter, in which case that is the winning vote under any system).

Is this the start of the 'theoretical minimum number of votes needed to win under AV'? Because the answer there is completely dependent on the number of voters. But it is worth noting that in Australia, since 1949 there have been 3,354 House of Representatives electorates contests, and in only seven has a candidate placed third after the first round managed to win. No one placed below that has ever managed to win an election (and Australia generally have about 8 to 12 candidates per ballot). Also, that is under Compulsory Preferential - it is expected that under AV this would be an even rarer event.

[motherT]

When we vote tomorrow are we using FPTP or AV to decide . How will the sheet be marked ? Will we have our party signs next to the choices ? Or will it be a simple yes to Nick or Yes to Cameron , hopefully they will be sensible and just use party signs, not sure if using Nicks face would be very clever? Not sure where labour will be are they next to the liberal one on this or separate? Has anyone seen what the form looks like ?

[westdulwich]

Loz Wrote:

> Did it deliver the fairest outcome? Compared to

> FPTP (i.e. naming Party A the winner) then, yes,

> absolutely.

>

> Reflected the preference of the voters? Again,

> compared to FPTP, then yes, of course, as FPTP

> ignores any preferences of the voter except their

> first choice.

Thank you. Could you answer without the caveat? Did the outcome reflect the preferences of the voters?

> AV is often referred as IRV - Instant Runoff

> Voting. The concept is the same except, as you've

> noted, those whose preferences are still 'live' as

> assumed not to change their votes between rounds.

> It's a tradeoff for the sake of simplicity, speed

> and cost. That is why examining preferences of

> live votes is a bit of a red herring - they don't

> actually 'exist' until needed.

But in the example I've given, those preferences would have already been expressed and recorded under AV. They wouldn't all be counted under AV, but they certainly exist.

> As I said, examining other options is no bad

> thing, but the choice is between AV and FPTP.

> Which do you think is better and fairer?

I haven't yet decided which of the two I would prefer.

[Loz]

westdulwich Wrote:

-------------------------------------------------------

> Loz Wrote:

> > Did it deliver the fairest outcome? Compared to

> > FPTP (i.e. naming Party A the winner) then, yes,

> > absolutely.

> >

> > Reflected the preference of the voters? Again,

> > compared to FPTP, then yes, of course, as FPTP

> > ignores any preferences of the voter except

> their

> > first choice.

>

> Thank you. Could you answer without the caveat?

> Did the outcome reflect the preferences of the

> voters?

Yes, I believe it did. As with any set of statistics you can bend and twist them whatever way you want to reach different conclusions, but I think that AV result produced the fairest outcome that represented the voters intentions.

If you start looking at preferences below live preferences you run the risk of getting the winner that no one wanted, rather than trying to reflect a majority wish. Take it to extremes - four candidates, three of them got 33.33% of first preferences each and the fourth got none. Likewise, the second preferences were distributed amongst candidates 1, 2 and 3. But the fourth did get everyone's third preference. There is no way you could argue that person 4 should be the winner.

I think AV is the best way of selecting a single winner of a vote (as opposed to PR, which assigns a number of 'winners' according to the percentage of the vote). As I've said, I've seen AV in use in Australia and I can't think of any example of anyone complaining the wrong 'winner' was selected.

[James Barber]

It's referendum day.

Great vibe from people I've met out in East Dulwich this morning while delivering Good Morning leaflets.

To thE maxxi's of this world voting NO to AV will kill aniy electoral reform for a generation in the way we select our MP's. Voting YES will get us a better voting system now and will keep the possibility of PR alive.

If anyone is at a loose end today plexus do come and help get YES voters out to vote.

[StraferJack]

popped in to the polling station this morning - can't say there was a huge turnout compared to the same time as the general election

I fully expect a NO win today, followed by weeks of gloating. But come the next general election the same old complaints will circulate and the chance will have gone by then. Fools

Given how much people have been complaining in recent elections you would think they would vote yes - you know, for a laugh if for nothing else.

Both campaigns have made me queasy but only one has turned my stomache. Voting Yes won't solve world peace but it would at least increase interest at the next polling day

[Alan Dale]

I am voting YES.

[misscarmelite]

Is it true that an injunction was attempted on the Guardian last night to stop it printing the front page that it eventually ran with? Just picked it up not sure what all the fuss was about?

Headline predicting absolute disaster for Nick clegg, huge no vote for AV along with a rather unflattering picture of Nick Clegg with a rather large Tory like double chin. Still not sure which way to go, James seems rather positive but then so was 'Baghdad Bob'.

[mockney piers]

more no vote dirty tricks missonepostite?

[misscarmelite]

Well I did see james all a whirl in his tweeds , skipping merily down the lane hand delivering leaflets , kissing babies , smelling the flowers blooming on the roundabout. He seemed completely oblivious to the impending shadow of doom that is about to descend from the clear blue skies.

[Alan Dale]

http://i.imgur.com/Zb9tJ.png

[ratty]

I'll be voting yes when I get back form work. But my wife is voting no. Actually, as these votes cancel each other out we would be best not bothering! Damn!

[maxxi]

James Barber Wrote:

-------------------------------------------------------

> It's referendum day.

> Great vibe from people I've met out in East

> Dulwich this morning while delivering Good Morning

> leaflets.

>

> To thE maxxi's of this world voting NO to AV will

> kill aniy electoral reform for a generation in the

> way we select our MP's. Voting YES will get us a

> better voting system now and will keep the

> possibility of PR alive.

>

> If anyone is at a loose end today plexus do come

> and help get YES voters out to vote.

Yikes! Singled out by the behemoth that is the Lib-Dem canvassing machine!

I have already decided on a yes jimmy (see lounge post) but suspect the no vote will triumph if for no other reason than the shameful lack of effort on the part of the yes campaign to debate the issue - they'd rather present loveable actors and comedians for us to agree with and aspire to be like.

I will be more interested in the local govt. elections throughout the country and only wish they were happening here too. Under any voting system i think your party are in for a bumpy night.

[ianr]

westdulwich Wrote:

-------------------------------------------------------

> If, for sake of argument, 10,000 people vote as follows:

> 4,100 vote Party A as their first preference, with 800 voting Party B and 3,300 voting Party C as their second preference

> 3,000 vote Party B as their first preference, with 1,500 voting Party A and 1,500 voting Party C as their second preference

> 2,900 vote Party C as their first preference, with 800 voting Party A and 2,100 voting Party B as their second preference

>

> Have I understood this correctly? Party A has the most first preference votes Party C is the party that most would be prefer, if

> their first choice was eliminated. Under FPTP, Party A is the winner, with 41% of votes cast. Under AV, Party B is the winner, with 51% of votes

> cast (after Party C is eliminated and its votes distributed).

I'm not keen on arguments based on single what-if snapshots. If one's going to compare systems, one really needs to look at all possible comparisons. Some simulations such as this or this I've found helpful.

I did however take this example seriously. Finally, I think my objection to it is that it's very unlikely to be a representative sample of actual voting behaviour.

If you look at, for example , the relative proportions (ex 100) of persons preferring A or C, the values you've chosen for the set of people who voted for C first are inconsistent with the other two.

A first A:C 55:45

B first A:C 50:50

C first A:C 21:79

I suppose it's possible to imagine a scenario in which such a discrepancy could occur, though I wonder how credible it would be. If, on the contrary, one assumes a more likely A:C ratio for the C-first voters of, say, 42:58 (1600 for A as second, 2100 for C as first), one ends up with a round two victory for A.

I don't think anyone with half a mathematical gene will think AV the best of systems; nor, for that matter, FTFP. Nevertheless, once you've adapted to the idea that there are potentially better measures of overall satisfaction than a single choice majority, I think it's worth trying the experiment. Whatever one thinks of it, it's undeniable that it's introducing more information into the decision process than a single choice can. Decisions made on a sole criterion, such as availability of school place or amount of insurance premium depending on one's postcode, can be fairer if more factors are brought into consideration. So too for preferences for candidates.

[Huguenot]

*chuckles*

So in other words ianr, it's not perfect, but it's better than we had before ;-)