Surpluses are treated differently from eliminated candidates in the election

[Brendan Burgess]

Hi Duke

I don't fully follow Scenario 1.

But FF does not get 3 quotas.
John got 3 quotas. Neither FF nor John get to choose what happens his other two quotas. The voters for John do.

Say John gets 1500 votes and the quota is 500
Mary gets no first preferences but 500 second preferences
Peter gets no first preferences but 500 second preferences
500 are non transferable
Joe gets 400 first preferences but no second preferences

The TV in Ireland is Surplus /Transferables which is 1000/1000 So Mary and Peter get 500 each.

The TV in Scotland would be Surplus/Total vote which is 1000/1500 = .66 So Mary and peter get 333 each and the 333 are not transferred.

If those who voted for John gave 750 2nd preferences each to Mary and Peter, then they would get 500 each and that would be fair enough.

 

[Duke of Marmalade]

Hi Duke

I don't fully follow Scenario 1.

But FF does not get 3 quotas.
John got 3 quotas. Neither FF nor John get to choose what happens his other two quotas. The voters for John do.

Say John gets 1500 votes and the quota is 500
Mary gets no first preferences but 500 second preferences
Peter gets no first preferences but 500 second preferences
500 are non transferable
Joe gets 400 first preferences but no second preferences

The TV in Ireland is Surplus /Transferables which is 1000/1000 So Mary and Peter get 500 each.

The TV in Scotland would be Surplus/Total vote which is 1000/1500 = .66 So Mary and peter get 333 each and the 333 are not transferred.

If those who voted for John gave 750 2nd preferences each to Mary and Peter, then they would get 500 each and that would be fair enough.

Yes I got that wrong. The transferable votes are not 1 quota but 2 quotas! I am reworking it.

 

[JohnCondon]

Replying to comments saying that the Irish system is plain wrong, and illogical, I wanted to make this list of advantages for how we do it. You could make another list for the Scottish system too.

1. Voters are encouraged to fill out their ballots, which is good for PR.

2. A non-transferable vote cannot dilute a transferrable vote in surplus when being transferred. (E.g if surplus is 1, which arose from 1000 non-transferable and 1 transferable, then 1 full vote is transferred. It is not wasted or diminished just because others have not specified a no.2. This is not the case when total vote is the denominator in surplus calculation)

3. The surplus is transferred only according to voters who expressed a preference as to where it should go. Seems fair to me! Voters had the option of whether they wanted to fill in 2,3,4 etc. and chose not to. I think on balance it's a good thing that if they do not give lower preferences then they don't control the surplus.

4. Fewer votes go non-transferable and not effective.

5. Voters can, if they wish, express an indifference as to the destination of surplus votes, by only filling in no.1.

To reiterate too, under our system, everyone still has one vote, it is a single transferable vote. Your vote will rest with just one candidate in the end, always with a count value of 1.

 

[Duke of Marmalade]

Replying to comments saying that the Irish system is plain wrong, and illogical, I wanted to make this list of advantages for how we do it. You could make another list for the Scottish system too.

1. Voters are encouraged to fill out their ballots, which is good for PR.

Yes there is a gearing of your vote if you form part of a surplus to be transferred. One might regard that as a moral bonus for voting down the card, but hardly more democratic.

2. A non-transferable vote cannot dilute a transferrable vote in surplus when being transferred. (E.g if surplus is 1, which arose from 1000 non-transferable and 1 transferable, then 1 full vote is transferred. It is not wasted or diminished just because others have not specified a no.2. This is not the case when total vote is the denominator in surplus calculation)

What is at stake is how much of your vote should be transferred or more precisely what are your chances of being transferred? See below.

3. The surplus is transferred only according to voters who expressed a preference as to where it should go. Seems fair to me! Voters had the option of whether they wanted to fill in 2,3,4 etc. and chose not to. I think on balance it's a good thing that if they do not give lower preferences then they don't control the surplus.

That is true for both systems. What is at stake is how much of your vote should be transferred. Under the Irish system you effectively get the value of more than one vote.

4. Fewer votes go non-transferable and not effective.

Well yeah! You have taken the non transferable votes and transferred them.

5. Voters can, if they wish, express an indifference as to the destination of surplus votes, by only filling in no.1.

But following on (1) they are probably not aware that they are missing out on the possible gearing up of their vote,

To reiterate too, under our system, everyone still has one vote, it is a single transferable vote. Your vote will rest with just one candidate in the end, always with a count value of 1.

This is the fundamental error. Those who express a preference and are part of a transferable surplus will actually get more than the value of 1 vote. This is separate to the point that because of the physical process many people will get the value of 2 votes or even more, 1 for a person elected with a surplus and 1 for being lucky to be part of the random process of selecting the transfers. This is a separate point but the Scottish system which is part computerised doesn't even have this distortion.

 

[Duke of Marmalade]

@Itchy Let's consider another system
You can only mark No 1 in the first round.
John gets elected with 10,000 votes, a 2,000 surplus over the quota.
So those 10,000 voters have 20% of a vote left and we now have a second round between them.
Some of the 10,000 won't bother to vote in the second round.
Under the Scottish system these votes are ignored. This is entirely consistent with what happens with eliminations or indeed those who didn't vote at all in the first place.
Under the Irish system these non-votes are in effect spread in proportion to those who did exercise a second vote. Now if they were spread amongst the other candidates according to how they did in the first round that would be fair enough, but a needless complication like spreading those who didn't vote in the first place in the same proportion as those who did vote.
By spreading them only to those who voted for John in the first round one is creating a bias for John's party running mate. If anything the decision not to vote in the second round was probably because the voter liked John but not too keen on his party. This is the point that I think @Brendan Burgess was making.

 

[Brendan Burgess]

To reiterate too, under our system, everyone still has one vote, it is a single transferable vote. Your vote will rest with just one candidate in the end, always with a count value of 1.

Hi John

I think if you got your head around why this is not correct, then you would see why the current system is flawed.

Many people get more than one vote.

I presume you agree that if we can show you that some people have more than one effective vote, that you would then agree that the Irish system is flawed?

Brendan

 

[Duke of Marmalade]

To reiterate too, under our system, everyone still has one vote, it is a single transferable vote. Your vote will rest with just one candidate in the end, always with a count value of 1.

I think this is key. Let's take an example.
A gets elected with 2 quotas i.e. 1 quota surplus.
Only one quota of A's votes are transferable, and so everyone who did express a preference has their whole vote transferred under the Irish system. The way I see it is that someone who voted A with a preference for B has not only had A elected but their full vote survives to elect someone else. So half of their vote elected A and their full vote survives. Looks like they got one and a half votes.
I think you are saying that the full vote of the non-transferable votes were used to elect A and the transferable votes had no part in electing A. That does not make sense to me and it clearly biases the transfers towards A's running mate.

 

[Brendan Burgess]

Hi Duke

I agree with your principle but the figures you used don't show that, unless you are referring back to numbers in a different post.

A gets 2,000 first preferences - twice the quota.

His second preferences split:
B : 0
C: 1000
Non transferrable 1,000

Half of the people who voted for A did not vote for C.

But under our system
The people who voted A 1 and B 2 got the following value

50% of their vote was used to elect A
100% of their vote was used to elect B

So they had an effective vote of 1.5

Or looking at it another way, those who voted A 1 and marked no second preference ...

50% of their vote was used to elect A
50% of their vote was used to elect B although they did not choose to vote for B.

 

[Brendan Burgess]

Using the above figures, the wrong candidate is elected.



B should have received only 500 votes and so their total would be 600, so c would have been elected without reaching the quota.

 

[Brendan Burgess]

Looking at it another way.



Now those who voted A1 and B2 are using 50% of their vote to elect A and 50% towards B

But the non-transferables are using 50% of their vote to elect A and, despite expressing no preference, 50% of their vote is transferred to B.

 

[Duke of Marmalade]

Yes, and the justification (wrong IMHO) would seem to be that the 1,000 non-transferable votes were sufficient to elect A and so the transferable votes were surplus to requirements and therefore transferred in their entirety.
In your examples B was probably A's running mate so the bias is not random, it favours the multi candidate parties.

 

[Itchy]

Using the above figures, the wrong candidate is elected.

View attachment 8952

B should have received only 500 votes and so their total would be 600, so c would have been elected without reaching the quota.

So the "correct" distribution method would have 70% of people who want absolutely nothing to do with C, yet them having 50% of the representation? C could never be elected by open ballot only by secret ballot.

I know its a stylised example, but it illustrates the point I made earlier,

In the Irish system, each of those 200 remaining votes will continue into the church/gym pile as part of the surplus, same as it would if it was an open ballot.
In the Scottish system, even though were already at the quota, because 500 people are happy with either a church or a gym, 100 of the remaining 200 votes must go in the bin (despite those people having a clear preference for one or the other).

You are advocating to discard valid preferential votes in favour of indifferent votes.

 

[Duke of Marmalade]

So the "correct" distribution method would have 70% of people who want absolutely nothing to do with C, yet them having 50% of the representation? C could never be elected by open ballot only by secret ballot.

We all agree that A's votes should be reduced in some proportion on transfer to allow for the fact that A has already been elected.
The Scottish system says the proportion should be the same for all votes i.e. the surplus divided by the total. The Irish system says the surplus came in the first place from non-transferable votes and so the reduction on transferable votes should be enhanced giving them in effect the value of more than 1 vote. Doesn't make sense to me.
As for folk getting elected when 70% of the population can't stand them, it happens.

 

[Freelance]

You are advocating to discard valid preferential votes

This is the nub of the issue.

The mistake being made is looking at it from a candidates perspective and trying to see what is fair to them. The correct view is to look at it from the voter's perspective, and ensuring that every vote is effective

In the example shown in post #89 there are:
1,000 votes which express a first preference for A and no second preference (Lets call them Group 1)
1,000 votes which express a first preference for A and a second preference for B (Lets call them Group 2)

I see nothing wrong with allocating Group 1 to A, and transferring all of Group 2 to B, which is the effect that the Irish system has. In fact this is entirely appropriate. Every one of the 2,000 votes is effective. No votes are discarded.

Running some alternative arrangement which allocates 1,000 votes to A and some (say 500 votes) to B and discards the balance is manifestly unfair and disenfranchises the last 500 voters. The actual effect of this is that A would now represents 1,500 voters, defeating the effect of the quota.

 

[Itchy]

The way I see it is that someone who voted A with a preference for B has not only had A elected but their full vote survives to elect someone else. So half of their vote elected A and their full vote survives. Looks like they got one and a half votes.

Is this not incidental to the process though? We have a minimum threshold, once a person passes that threshold, its not useful for people to keep telling me they like the guy whos already going to get a seat. I need to know who else is liked!

I think you are saying that the full vote of the non-transferable votes were used to elect A and the transferable votes had no part in electing A. That does not make sense to me and it clearly biases the transfers towards A's running mate.

Again, I think this is incidental. We saw from the actual voting patterns in the Euros that transfers were going all over the place, across parties and ideologies. We also know that it is disastrous for parties to run too many candidates as it tends to split the vote and when transfers do not materialise, they fail to get the quota. In fact, it is often cited that a not insignificant amount people vote straight down the ballot

 

[Duke of Marmalade]

In the example shown in post #89 there are:
1,000 votes which express a first preference for A and no second preference (Lets call them Group 1)
1,000 votes which express a first preference for A and a second preference for B (Lets call them Group 2)

I see nothing wrong with allocating Group 1 to A, and transferring allocating all of Group 2 to B, which is the effect that the Irish system has.

Yes that explains the situation perfectly, though my correction makes it clearer that Group 2 is deemed not to have voted for A at all, with which I do see something wrong.

Running some alternative arrangement which allocates 1,000 votes to A and some (say 500 votes) to B and discards the balance is manifestly unfair and disenfranchises the last 500 voters.

The Scots which have come late to the party and can learn from precedent to not agree that there is something manifestly wrong. The "disenfranchisement" is that those who did not indicate a preference only got to use half their vote. Any system is full of this "disenftranchisement" through the elimination process. Group 1 get full value for their vote under the Scottish system - 0.5 in electing A 0.5 transferring to B.

The actual effect of this is that A would now represents 1,500 voters, defeating the effect of the quota.

A "represents" 2,000 voters.

 

[Freelance]

Group 1 get full value for their vote under the Scottish system - 0.5 in electing A 0.5 transferring to B.

You can't transfer half a vote. Under STV, a vote is whole and indivisible no matter how you look at it.

A "represents" 2,000 voters.

That's a FPTP view. The STV as we use it introduced a transferable surplus to avoid this, which would otherwise amount to underrepresentation. A represents 1,000 voters and B represents 1,100 as her surplus cannot be transferred in this case.

 

[Duke of Marmalade]

You can't transfer half a vote. Under STV, a vote is whole and indivisible no matter how you look at it.

The physicals again - that is irrelevant. The Scottish deal in units of 1 in 10,000 but they do use a computer. The NI folk are manual but deal in .01 of a vote. Our Seanad actually deals in 100 units per vote.
The fact is that a surplus is a fraction. In our rather primitive system we express that fraction as a number of whole votes and by a somewhat arbitrary process (albeit set out in law) that number of whole votes are selected for transfer.
What is at stake is the fraction. The fraction is Numerator/Denominator. Both systems agree that the Numerator should be the Surplus, including non-transferables. The Denominator in the Scottish system is all the votes that elected A; the Denominator in the Irish system deducts the non-transferables, totally inconsistent with the approach to the Numerator and effectively gearing up the transferable votes.

 

[Duke of Marmalade]

Is this not incidental to the process though? We have a minimum threshold, once a person passes that threshold, its not useful for people to keep telling me they like the guy whos already going to get a seat. I need to know who else is liked!

So we agree that folk do get 1.5 votes in the example. For many that would be a showstopper.

 

[Duke of Marmalade]

C could never be elected by open ballot only by secret ballot.

I think you mean a rolling ballot rather than a open one. Open suggests somehow more wholesome.
So let's consider your system where the voters are selected at random but in the privacy of their own home they get a Whatsapp message to vote and immediately everybody is told the latest state of play.
I think what you are saying is that C is so detestable and it starts to look like the Scottish system (or indeed any system) will get her elected over B if people continue to express no preference then that will make their minds up for them.
But I could posit the exact opposite situation. B is detestable and under he looks like being elected over C so again their minds get made up.
What you are saying is that there is indeed a preference between B and C which folk might ignore with a blank sheet but when the writing is on the wall they make their minds up.
I think the more reasonable assumption is that they genuinely don't care.