'Is AV really so complex? Or is it just confusion marketing?' blog discussion

Ellabelle

bristol_pilot wrote: »

My own opposition to AV is based on potential non-monotonicity a.

I have just found out (I think, at least if I understand the definition correctly:o) what this means.

Does it mean that someone's 1st choice is worth a whole 'point' and someone 4th choice could also end up being worth a whole 'point' when really if a person agrees with yourself a 4th choice should only really be worth a quater of a point?

InMyDreams

Ellabelle wrote: »

Thats the other problem with this - its so complicated that when someone like me (I would say I have an average intelligence) struggles to understand this then what hope is there for others who perhaps are less intelligent or have perhaps a learning disability - Who are as much entitled to vote as anyone else. (No offence to anyone meant, especially those with learning disability - some of whom of course will be able to understand AV, don't want to generalise)

What is complicated (and I'm seeing now more and more why) is trying to work out which system would deliver the fairest result. And trying to agree on what 'fair' even means!

What is relatively easy (if AV was to be adopted) is the mechanics of how to vote. Just rank your candidates in order of preference. 1 next to your favourite, 2 next to your second choice if you can't have 1, and so on. This is what needs to be easy for everyone to do, and for the NO campaign to tell you it's complicated is insulting.

What is also complicated is trying to work out how to vote tactically (and how to get others to vote tactically) to best achieve the result you want. This is not a bad thing at all, I think.

tomwakefield

Ellabelle wrote: »

I have just found out (I think, at least if I understand the definition correctly:o) what this means.

Does it mean that someone's 1st choice is worth a whole 'point' and someone 4th choice could also end up being worth a whole 'point' when really if a person agrees with yourself a 4th choice should only really be worth a quater of a point?

Wiki link
The example given is essentially as follows.

Candidate A wins against Candidate B and C in an election.

In the subsequent election Candidate A wins more votes in the first round, which affects the order of elimination of the other candidates and therefore the distribution of secondary votes, meaning Candidate A fails to win the second round.

In short, because they ended up with more votes, they ultimately ended up worse off.

Ellabelle

InMyDreams wrote: »

What is complicated (and I'm seeing now more and more why) is trying to work out which system would deliver the fairest result. And trying to agree on what 'fair' even means!

What is relatively easy (if AV was to be adopted) is the mechanics of how to vote. Just rank your candidates in order of preference. 1 next to your favourite, 2 next to your second choice if you can't have 1, and so on. This is what needs to be easy for everyone to do, and for the NO campaign to tell you it's complicated is insulting.

What is also complicated is trying to work out how to vote tactically (and how to get others to vote tactically) to best achieve the result you want. This is not a bad thing at all, I think.

I would argue that it is just as important for a voter to understand how their vote is to be used and counted as it is for them to understand the mechanics of filling in the ballot paper.

Ellabelle

tomwakefield wrote: »

Wiki link
The example given is essentially as follows.

Candidate A wins against Candidate B and C in an election.

In the subsequent election Candidate A wins more votes in the first round, which affects the order of elimination of the other candidates and therefore the distribution of secondary votes, meaning Candidate A fails to win the second round.

In short, because they ended up with more votes, they ultimately ended up worse off.

Ta very much for the explanation, so is this example as AV is or not?

Oh dear i'm making myself seem as bit thick now!

tomwakefield

Ellabelle wrote: »

Ta very much for the explanation, so is this example as AV is or not?

Oh dear i'm making myself seem as bit thick now!

The example is AV voting, demonstrating that it fails the monotonicity criterion.

Don't worry about appearing thick - better to ask questions and fully understand it so you can make an informed vote at the referendum.

wozearly

bristol_pilot wrote: »

Furthermore, voters might have ranked their preferences differently had they known in advance which candidates would be eliminated and in which order.

A moot point as guessing the polling order of the smaller candidates is likely to be very, very hard.

And its no different to looking at FPTP results with hindsight and saying "...but if all those Lib Dem supporters had voted Labour, we wouldn't have a Tory MP...if only they'd known their candidate wouldn't win".

bristol_pilot wrote: »

My own opposition to AV is based on potential non-monotonicity and the issue of 2nd/3rd preferences of voters whose 1st preference was for an extremist party potentially deciding the outcome. As we have seen on this thread, these arguments are complex to get across and there are counter- and counter-counter- arguments to each and every point. So if I were running the 'no' campaign I would concentrate on Clegg bashing and alleged expense, which is what they seem to be doing. We live in a dumbed-down world but, hey, that's democracy.

Non-monotonicity is a definite downside of AV, in the same way that the "tyranny of the safe seat" is a definite downside of FPTP. No voting system is perfect, and personally I see non-monotonicity as the lesser of two evils.

...and I never even knew there was a specific term for it until today.


As for extremists deciding the outcome, they can already do that using tactical voting under FPTP. Certainly nothing stopping them.

If anything, its harder under AV as this would only happen if the extremists were large enough as a group or the leading candidate was very close to 50%, and lots of them had similar second preferences.

Plus, whatever happens, there's an incredibly low chance of getting an extremist candidate elected under AV - and even if extremist votes cause their second favourite to win, that candidate would have to be pretty popular across everyone else to cross the 50% mark in the first place. So they're not likely to be Nick Griffins.


As for Clegg bashing, I could never support any campaign which focused more heavily on misinformation, theoretical downsides and attacking the individuals leading the other side because its "too complicated" to explain the merits of their stance...

wozearly

Ellabelle wrote: »

Ta very much for the explanation, so is this example as AV is or not?

Oh dear i'm making myself seem as bit thick now!

The example is using the AV system, and highlights an occasion when it can work in a counterintuitive way - admittedly, the chances of voters having a circular preferencing (1st A, 2nd B : 1st B, 2nd C: 1st C, 2nd A) are somewhat slim. The article mentions only one specific real-life example.

In any event, monotonicity isn't the be-all and end-all of deciding whether a system is fair. Tactical voting can potentially throw up equally 'bad' outcomes - election leaflets often use past data to imply that Candidate X can't win, so you need to vote for Candidate Y to beat Candidate Z (who we know you don't want).

This is essentially trying to second-guess what people's 1st and 2nd preferences are. Of course, it could be completely wrong if people's preferences change.

The 3rd placed candidate / party from the previous election might now have more first preferences amongst voters and would win outright...but if a significant portion of their voting base switches to Candidate Y based on past data and/or electioneering, Candidate Y might win, or the division in support might ensure Candidate Z gets re-elected in any event.

Whatever AV's faults, it does at least remove the need to try to second-guess the level of support your preferred candidate might or might not get in the vast majority of cases.

dan2097

tomwakefield wrote: »

Yes, I think it is simple. More complicated than FTPT, but still simple. And the most simple solution isn't necessarily the best.

As tactical voting is far far less effective in AV for people who really consider their votes AV is simpler, you vote for who you want (and who you would want if you couldn't get them etc.) rather than thinking who you need to vote for to stop someone else winning.

bristol_pilot

http://www.newsgrape.com/a/cest-lav/