Monday, February 15, 2016

It's true, folks : If the SNP take all of the list votes, they'll take all of the list seats

As long-term readers know, I firmly believe that "No to AV" was one of the most disreputable campaigns in the whole of British political history.  Their landslide victory in the 2011 referendum was entirely based on a small number of extremely basic lies and distortions which were scribbled down on the back of an envelope, and which they got away with because hardly anyone cared enough about the issue to research it for themselves.  Perhaps the lowest point of all was a broadcast which depicted a teacher trying to "explain" AV to a class of bright sixth-form students.  To begin with, she made genuine points about what is a very simple system in a ridiculously garbled way, but then she quickly degenerated into reading total gibberish from a supposed "AV manual"...

"The candidate with the most votes may or may not win depending on whether they gain 50% of the vote.  Or not.  After that, it's down to your second preference vote, providing you vote for an unpopular candidate.  Then your second and third choices will be added to the most popular ones, which means that now..."

"If the winner gets 38%, and the second gets 32%, then the contest is still not finished, because what they do now is if the runner-up gains more second votes than the winner, then number 2 or 3 is now the winner."

"After that, what happens now?"

"You count everyone's third choice?"

"No you can only use the third preferences for those people who have already been eliminated twice."

The latter claim has no basis whatever in fact - it's impossible under AV for anyone to be "eliminated twice".  This really begs the question of whether parties or political campaigns should be allowed to get away with flat-out lies (as opposed to mere distortions) in election broadcasts and literature.

Something similar appears to be happening with the increasingly exotic claims that are currently being made about the Additional Member System used for Scottish Parliament elections, and more specifically about the potential impact of any attempts at "tactical voting on the list".  I'm not clear whether some people are following the example of No to AV and making deliberate attempts to mislead, or whether we're just seeing a very extreme "Chinese whispers" effect.  Either way, the electoral system seems to be mutating rapidly before our eyes.

I've just been speaking to a Solidarity supporter called Angela Kerrigan on Twitter, and I was very struck that she made almost exactly the same claim as Green supporter Morag Hannah did the other day - that the SNP can't win all of the list seats even if they win all of the list votes.  I pointed out to her that the opposite is true, and that if the SNP win 100% of the list votes they'll take 100% of the list seats.  But she stuck to her guns and insisted I was wrong.  Her explanation was that the d'Hondt method "deducts" list seats from parties that have done well in the constituencies.

That, of course, is gibberish on a par with the stuff about candidates being "eliminated twice" under AV.  But given the way that the tactical voting lobby have misrepresented the d'Hondt calculation, you can kind of see how people might wrongly think there is some sort of rule about "deducting list seats".

Let's clear this up once and for all by going through the arithmetic in full.  The reality is that even on 95% of the list vote, the SNP would be absolutely guaranteed to win 100% of the list seats in any region.  To demonstrate the point, let's take a region with ten constituency seats (that's the maximum) and assume that the SNP win them all.  The SNP take 95% of the list vote, but to make it as hard as possible for them to claim the final list seat, let's assume that every non-SNP list vote is won by a single opposition party - and in Angela's honour, we'll also assume that party is Solidarity (although it doesn't make a jot of difference whether it's Labour, Tory, Lib Dem or whatever).

*  *  *

List result -

SNP : 240779 votes (95%)
Solidarity : 12673 votes (5%)
All Others : 0 votes (0%)

*  *  *

First count -

Because the SNP already have 10 constituency seats, their original vote is divided by 11 (10 + 1).

240779 divided by 11 is 21889.

As Solidarity have no seats yet, their vote is unchanged.

SNP : 21889 votes
Solidarity : 12673 votes

The first list seat is won by the SNP.

*  *  *

Second count -

Because the SNP already have 10 constituency seats and 1 list seat, their original vote is divided by 12 (11 + 1).

240779 divided by 12 is 20065 (after rounding).

As Solidarity have no seats yet, their vote is unchanged.

SNP : 20065 votes 
Solidarity : 12673 votes

The second list seat is won by the SNP.

*  *  *

Third count -

Because the SNP already have 10 constituency seats and 2 list seats, their original vote is divided by 13 (12 + 1).

240779 divided by 13 is 18521 (after rounding).

As Solidarity have no seats yet, their vote is unchanged.

SNP : 18521 votes 
Solidarity : 12673 votes

The third list seat is won by the SNP.

*  *  *

Fourth count -

Because the SNP already have 10 constituency seats and 3 list seats, their original vote is divided by 14 (13 + 1).

240779 divided by 14 is 17199 (after rounding).

As Solidarity have no seats yet, their vote is unchanged.

SNP : 17199 votes 
Solidarity : 12673 votes

The fourth list seat is won by the SNP.

*  *  *

Fifth count -

Because the SNP already have 10 constituency seats and 4 list seats, their original vote is divided by 15 (14 + 1).

240779 divided by 15 is 16052 (after rounding).

As Solidarity have no seats yet, their vote is unchanged.

SNP : 16052 votes 
Solidarity : 12673 votes

The fifth list seat is won by the SNP.

*  *  *

Sixth count -

Because the SNP already have 10 constituency seats and 5 list seats, their original vote is divided by 16 (15 + 1).

240779 divided by 16 is 15049 (after rounding).

As Solidarity have no seats yet, their vote is unchanged.

SNP : 15049 votes 
Solidarity : 12673 votes

The sixth list seat is won by the SNP.

*  *  *

Seventh count -

Because the SNP already have 10 constituency seats and 6 list seats, their original vote is divided by 17 (16 + 1).

240779 divided by 17 is 14163 (after rounding).

As Solidarity have no seats yet, their vote is unchanged.

SNP : 14163 votes 
Solidarity : 12673 votes

The seventh list seat is won by the SNP.

*  *  *

At that point we stop counting, because there are only seven list seats to be distributed.  So how is it possible that the SNP have won every single seat?  Basically the aim of the Additional Member System is to make the overall number of seats won by any party in a region (constituency and list seats combined) roughly proportional to the percentage of votes won by that party on the list ballot.  In the region we've been discussing, there are seventeen seats in total, which means that any party that wins more than sixteen-seventeenths of the list vote (approximately 94.1%) can't be beaten to the final list seat, even if the opposition vote isn't split at all.

And what if, more realistically, the opposition vote is split between multiple parties?  In that event, the seventh count in the above example might look something like this...

SNP : 14163 votes
Labour : 8527 votes 
Conservatives : 3289 votes
Liberal Democrats : 1002 votes 
Greens : 925 votes
Solidarity : 231 votes
RISE : 189 votes

As you can see, the SNP have won the final seat even more easily.

100 comments:

  1. You are 100% correct. You are my hero for typing this out.

    ReplyDelete
    Replies
    1. You know it peeps, RISE are idiots and the greens are idiots not to be trusted, we don't need them any more anyway! Stick to the plan, both votes SNP for an independent Scotland is the only way forward.

      Delete
  2. Your patience is truely to be commended. When i read the gibnerish that Supporters of the minority parties are spraphing I am reminded of the phrase, "never attribute to malice that which may more easily be explained by incompetence"

    ReplyDelete
  3. Your patience is truely to be commended. When i read the gibnerish that Supporters of the minority parties are spraphing I am reminded of the phrase, "never attribute to malice that which may more easily be explained by incompetence"

    ReplyDelete
  4. .....However, what is more likely is that SNP will get around 50% of list vote and votes going to smaller pro-independence parties,(Greens, RISE, SNP, Solidarity) are in fact likely to cost SNP list seats and instead allow unionists scrape another few critical seats across Scotland. This could allow the pro independence parties to have no majority and thus absolutely no chance of a second referendum for at least a generation. Given this, I am suprised that the unionist parties and their banking cronies do not bankroll the smaller independence parties!

    ReplyDelete
    Replies
    1. What makes you think they don't?

      Promoting the small parties all demanding to have their cake and eat it rather than campaigning to win votes from the unionists could be far more effective than voting wheels had any hope of being.

      Fortunately it seems that only a relatively small number of folk are listening to this nonsense so far

      Both votes SNP

      Delete
    2. Well I have been suspicious of RISE since their inception.

      Beware pop up parties.

      Delete
    3. "This could allow the pro independence parties to have no majority and thus absolutely no chance of a second referendum for at least a generation"

      A ridiculous assertion bearing no relation to reality. The chances of the SNP gaining less seats than last time must be vanishingly small. The Greens will also win a handful of seats regardless (and I know some question their commitment to indy but there is no credible suggestion that they would block another referendum) so there WILL be a majority for an indyref in the next parliament. Suggesting otherwise is nonsensical and misleading at best.

      Not to mention that it possibly doesn't even matter, as barring Brexit, the SNP won't be calling for one anyway!

      Delete
  5. True enough, though the probability that you will get all that SNP support for the list is as poor as being selective with the list vote in favour of the best placed other Indy party.... I my opinion worse .... Fundamentally the system is designed to stop One party taking all the seats. The repercussions of not hitting achieving a clean sweep of the list votes ... is not linear! Its negatively exponential ... fail to achieve it by just a small margin and you will very rapidly make room for a substantial unionist representation in Holyrood. I would rather a larger Multi Party Indy block, with a smaller SNP majority ... than a one party, one target, SNP super majority. I vote SNP , have always done so .... for independence ... not policy!

    ReplyDelete
  6. What about UKIP list voters? Do these have the potential to reduce the number of list 'wins' by the main unionist parties, or are they less of a threat because they are not expected to win many constituency seats anyway (thereby reducing the effect)?

    ReplyDelete
    Replies
    1. My hope is that UKIP cost the Tories a list seat or 2. If the EU referendum is only a month after, UKIP will be keen to use the Scottish, Welsh, and London elections as a pre-campaign launch pad.

      Delete
    2. My hope is that UKIP cost the Tories a list seat or 2. If the EU referendum is only a month after, UKIP will be keen to use the Scottish, Welsh, and London elections as a pre-campaign launch pad.

      Delete
  7. With Brexit very much in the UK media's eyes, UKIP will take enough Tory and Labour votes to win a seat in one or two regions just as they shocked everyone in 2014. With Europe as the backdrop, even in a Scottish Parliament election, they will rise...

    ReplyDelete
    Replies
    1. I doubt UKIP will win a list seat. They need to score at least 6% to have a chance in any region.

      Delete
    2. I doubt UKIP will win a list seat. They need to score at least 6% to have a chance in any region.

      Delete
    3. Based on the most recent poll from MORI, they just need to up their share about 7x to be in with a chance.

      They used to get a few % but Farage's anti-Scottish attitude in the GE - notably live on TV - screwed them.

      And what's the point now anyway? Tories are holding an EUref. that renders UKIP meaningless given they're a one trick pony.

      Delete
  8. How the hell does this "vote anyone but SNP on the list" infomercial count for news at the BBC?

    http://www.bbc.co.uk/news/uk-scotland-35573771

    ReplyDelete
  9. James all your arguments are sound. But if we really want to get the SNP x 2 message across what we really need is a catchy slogan that captures the reality. I've been racking my brains. I hope someone can come up with better.

    SNP one and two, for independence we must do

    SNP in both your votes, a savvy citizen this denotes

    SNP in seat and list, for powers and freedom, you get the gist

    ReplyDelete
    Replies
    1. 'Nicola Sturgeon for fist minister' on the ballot should do the trick. After all, her sat ratings are normally mid 60's.

      Delete
    2. Yeah but the reason that worked so well for Alex Salmond was because the ballot paper was alphabetical!

      Delete
    3. This trick would still put the SNP higher up the ballot than the billions of parties beginning with "S", though. And above RISE. In fact, are there any parties in Scotland whose name is higher up the alphabet than "Nicola"?

      Delete
    4. "This trick would still put the SNP higher up the ballot than the billions of parties beginning with "S", though."

      It wouldn't anymore. The rules were changed after what happened in 2007.

      Delete
    5. For me its SNP one and two,that'll do.

      Delete
  10. "Basically the aim of the Additional Member System is to make the overall number of seats won by any party in a region (constituency and list seats combined) roughly proportional to the percentage of votes won by that party on the list ballot."


    Interesting James. By the way by your own logic if SNp in a region win all the constituency seats or all but one and then say in a region get less than 50% on the list does that also mean they are likely to win phuqall on the list? It would seem so or does this logic only apply when disproving ludicrous RISE type claims.

    ReplyDelete
    Replies
    1. If the SNP narrowly miss a constituency seat and the list vote holds up then they will get a list seat instead of the constituency seat

      If the SNP narrowly miss a constituency seat and the list vote is eroded by SGP, RISE, SSP, Solidarity, Morningside Liberation Front et al then Unionists will almost certainly be the beneficiaries.

      If in doubt think of the List Vote as being like an insurance policy and vote for the party you want to form the Government.

      Tactical voting in the Holyrood electoral system, unlike straight FPTP, just doesn't work. And those parties trying to convince me that it does are proving why I should not trust them with my vote.

      Both votes SNP

      Delete
    2. As we've discussed many times before, Jam, if you're going to put total faith in the opinion polls when they tell you that the SNP are going to win every constituency seat in a region, it's logically consistent to also believe the polls on what they say about the list. And at the moment, they clearly suggest that the SNP will not win "phuqall" on the list. The TNS poll that RISE prayed in aid in their notorious press release actually had the SNP on six list seats (and RISE on zero, naturally).

      Delete
  11. I think we should make more of the last EU election when trying to explain the dangers of "tactical" voting on the lists. A campaign by the Scottish Greens to gain an MEP and keep UKIP out backfired and gave us David Coburn MEP instead of a third SNP MEP.

    ReplyDelete
    Replies
    1. You can't be sure of that. It could equally be postulated that people favoured SNP as the 'big party' because they generally think the Greens have little chance of being elected and so the Greens missed out because of SNP1&2 campaigning.

      The reality is both parties fell 30000 votes short of the seat...

      Delete
    2. CMac : The problem with your latter point is that it's a hell of a lot more plausible that a large party like the SNP could locate an extra 30000 votes from somewhere.

      Delete
    3. in this straight proportional vote yes I think that point is valid. In the two vote system it's perfectly conceivable that a well intentioned and well received (This bit is unlikely I concede) campaign to vote e.g. SNP-1 Green-2 could work with relatively low risk and relatively high rewards.

      I don't actually advocate a tactical vote by any means - I personally think people should vote for policy on the list - I just take exception to the fanatical opposition the suggestion receives when it is in fact a perfectly reasonable proposition for consideration.

      I understand your general point about misleading campaigning and I think it's valid for you to call it out - unfortunately most of your output on this seems to be accompanied by a whole lot of equally misleading stuff like "has to be SNP1&2 for independence". It has a bit of a greedy playground bully feel to it (given the SNPs total dominance of the polls) that I think will probably be off putting to a lot of people. I don't level that at you by the way but those who use your analysis in that way.

      Delete
    4. "I don't level that at you by the way"

      I'm grateful for that ultra-late disclaimer, because up to that point it certainly sounded like you were levelling it at me. I have never, ever used the phrase "SNP 1 & 2" (except to say that I don't think it should be used), even though it's been wrongly attributed to me about a million times - mostly, it has to be said, by mischief-making Green and RISE supporters on various Facebook groups.

      As you know, I fundamentally disagree with you that tactical voting on the list is feasible or a reasonable proposition for consideration, but I won't rehearse the reasons yet again.

      Delete
    5. Apologies - I see now it does read to be quite accusatory to that point. Genuinely not my intention!

      Delete
    6. On the fundamental disagreement I would tend to agree with you in general but would suggest that in this particular election we will have good foreknowledge (to a high degree of certainty - but obviously not perfect) of the constituency result.

      Basically with SNP sitting at ~50% in polls it would (probably) work, at ~40% it (probably) wouldn't.

      Of course it is a gamble - but it would be an educated gamble not a random punt...

      Delete
    7. I'm noticing more passive/aggressive comments online, really masquerading trolls, feigning 'genuine concern' and pretending to be 'the voice of reason'. Wee Ginger Dug has one such, (Sister Anonymous). Sounds like Cmac11 is one of your trolls in disguise. It's almost endearing that such disguised trolls think they have succeeded in hiding their true (troll) natures.

      I wish there was a term for the pretendy trolls. AS trolls were outwitted by goats maybe we should call pretendy trolls, 'Nannies'? In honour of those who can see right through their disguise.

      Delete
    8. Aye mate - bang on the money. There was me thinking I was sparking some useful debate by putting across reasons that I disagree with some of James' output and much of the social media spin it subsequently generates. turns out I'm just a passive aggressive troll and wee ginger dugs sister... or something!

      Delete
    9. I think many people have become so irritated by the wilful blindness and misrepresentation going on that they have become a bit strident. Suffering fools for any length of time tends to take people that way.

      Delete
    10. Bibbit, there is a term for people like that. It's been around for quite a long time. It's "concern troll".

      Delete
  12. Thanks for that, James. Very well explained.

    To be honest, I really don't know why this is so difficult for the likes of Ms Kerrigan to comprehend.

    Put simply, if the SNP receieved 100% of votes (ergo all other parties received zero votes), WHY should any other party expect to be rewarded with any seats on zero votes?

    It's not rocket science, Ms Kerrigan.

    SNP x 2 = IndyRef 2

    ReplyDelete
  13. First things first: independence.

    SNP x 2

    ReplyDelete
  14. No Drew

    Don't think of it as an insurance policy as that is not what it is. It's a vote that takes into account the seats already allocated in both the constituency seats in the region and then in addition iteratively the lists seats in the region. If you believe Green are going to get over 10% in a region on the list then SNP will likely have to get mid-50s on the list to get a list seat before Green get 2 list seats. In Lothian that is very likely and very possible in Mid and Fife and Highlands and Glasgow. In most regions green will be in line for a list seat before any SNP one will be and in some it might be 2 seats before SNP. if there is evidence of the latter we should switch and drivel about pointless scenarios about SNP getting all the list votes 9which they won't) isn't going to disguise it.

    How about we use the Scotgoespop monthly rolling figures of the below and see how many SNP get on the list:

    Regional list ballot :

    SNP 46.6% (+0.6)
    Labour 19.4% (n/c)
    Conservatives 17.0% (-0.2)
    Liberal Democrats 6.8% (n/c)
    Greens 6.4% (-0.2)

    You can assume do two scenarios on the constituency (SNP win all and SNP win all but one). Then work out how much more Green need to take an additional seat and how much SNP need to get a seat..you can do it here:

    http://www.dhondt.eu/php/

    The answer is revealing and much more interesting than stupid scenarios albeit peddled by some solidarity lunatic..

    the answer is if SNP win every seat then green get a list seat and SNP get nothing. Indeed if every green list vote went to SNP they'd still get nothing. If green get 13% they get 2 list seats. In young, areas with high numbers of social class As and Bs greens will be approaching 10% and higher. That is surely the interesting point. But keep jibbering about SNP * 2 while contributing to electing a nawbag....lalalalalala

    ReplyDelete
    Replies
    1. "Don't think of it as an insurance policy as that is not what it is. It's a vote that takes into account the seats already allocated in both the constituency seats in the region and then in addition iteratively the lists seats in the region."

      That's an utterly contradictory statement. There are no constituency seats "already allocated" at the moment you cast your vote. You don't know how many constituency seats the SNP will win, and therefore by extension you have no idea what scope the SNP have to win one, two, three or even more seats on the list. Therefore, it's perfectly reasonable for Drew to describe an SNP list vote as an insurance policy.

      How about this example : In 1999, Dennis Canavan stood as an independent candidate in both the Falkirk West constituency and on the Central Scotland regional list. By definition it was impossible for him to be elected twice. But many thousands of people voted for him twice, because they wanted to maximise his chances of getting in. In other words, an insurance policy.

      Delete
    2. Jam

      You said "... But keep jibbering about SNP * 2 while contributing to electing a nawbag....lalalalalala"

      I believe I had already answered that point when I said "... and vote for the party you want to form the Government.

      Tactical voting in the Holyrood electoral system, unlike straight FPTP, just doesn't work. And those parties trying to convince me that it does are proving why I should not trust them with my vote.

      Both votes SNP"

      The smartest thing the Unionists ever did in the early days of the Scottish Parliament was convince a large number of people that the list vote is of lesser significance than the constituency vote (when if anything the reverse is true) and that it can somehow be used for a second preference party without harming the chances of the first preference party forming a government.

      For the avoidance of doubt based on experiences before during and after Indyref I do not trust the SGP / SSP / RISE / Solidarity etc commitment to independence. I live in a marginal seat which could easily go back to a Unionist if they Rainbow of Tories can get their collective act together. Consequently I'll be very happy to have my insurance policy thank you very much.

      Both votes SNP

      Delete
  15. why do you publish polls james? You have a rolling average your Scotgoespop monthly on the list and the constituency. you could make these useful with illustrative scenarios and who would get what on the list. That would be informative. It would obvously hav to have health warnings around the distribution of the list vote across regions i.e. low greens in central, west scotland, north east and south. Higher greens elsewhere and very high in Lothian. But it would be informative. It would also make use of the regional list figures which are currently not very useful at all.

    ReplyDelete
    Replies
    1. "why do you publish polls james?"

      What a strange question. Because I find them interesting, I suppose. What you're driving at is you think they can somehow be used as a tool for "tactical voting on the list" - but they can't. Tactical voting on the list isn't feasible, because the level of foreknowledge required is too precise. There are no polls that provide the extreme level of accuracy that you seem to ascribe to them.

      Delete
  16. you don't know but you can guess. I'm going to guess that SNP take every constituency seat in mid and Fife (my constituency) and none on the list. if it was all random the chances of that would be astronomical. Anyone fancy an even money wager on it?

    ReplyDelete
    Replies
    1. You're proving my point. An even money wager implies a 50% chance - that's gambling voting, not tactical voting.

      But if you want to offer 1/20, you're on.

      Delete
  17. Thank you so much for doing all the work. I have tried and tried to explain this but now all I have to do is to link to you!

    ReplyDelete
  18. Is it because using every single monthly Scotgoespop monthly rolling average list poll since you started making them that SNP wouldn't win a single seat on the list? Hmm call me a cynic..

    ReplyDelete
    Replies
    1. That's simply not true.

      Delete
    2. Jam, when are you going to realise that no matter what you say, people are going to consider their list vote as insurance (a la Dennis Canavan scenario). I don't CARE if the polls show the SNP at 80% or even 90% in the constituency vote - I'm STILL going to vote SNP/SNP - simply because a SNP majority is (IMO) VITAL to the yes campaign. The odds against a small child getting injured in a car accident are fairly small, but any parent not bothering to place a child in a car seat is considered criminally negligent. It depends on how important the result is to each individual. If the SNP majority is considered absolutely vital, then you strap your second vote in the car seat.....

      Delete
    3. @Jam, that's nonsense. If you input any f the polls over the last week or so the SNP wins between 4-6 Regional List MSPs. If you put in the numbers suggested by the sub-sample and assign 15% of regional votes to the Greens, SNP come close to losing their majority.

      Given that this spread will NOT be consistent across Scotland, it becomes impossible to try and forecast what the actual voting will be in YOUR constituency/region, and therefore impossible to predict with accuracy whether voting SNP/Green (for example) is better than SNP/SNP if you are wanting more pro-Indy MSPs (this assumes you consider the Greens to be pro-Indy in principle. I don't)

      Add to this uncertainty the SSP/RISE/Solidarity factor - who might get the transferred SNP votes, and the most likely outcome is additional Unionist/Pro-Westminster Managed Dependency (i.e pro-WMD) MSPs.

      You might consider this a risk worth taking... I don't

      Delete
  19. I'd like to see the figures for the rolling average polls that give SNP a list seat..

    ReplyDelete
    Replies
    1. You're the one that made the original claim. How are you calculating it? Why are the SNP never winning a list seat in the north-east, for example?

      Delete
    2. I've just pumped the current Poll of Polls figures from this blog into the Scotland Votes calculator, and it produces 5 SNP list seats and 5 Green list seats. So I really don't know what you're talking about.

      Delete
    3. I used the figures in your monthly rolling list average and assumed SNP win all or all but 1 of the constituency seats. In both scenarios SNP don't get a list seat. In the former if SNP got all of the green list vote they'd still not get a list seat. I know there is a possibility of SNP in the north east and I'm not advocating green on list. i'm advocating green on list in Lothian and a watching brief in Mid and Fife, Weej and Highlands. Elsewhere SNP * 2. You could use historic rolling average trends on the list to give likely party specific lows and highs. In south and highlands you could give SNP losing 2 constituency seats as an illustrative example (and in these examples you probably would find SNP picking up list seats).

      Delete
    4. Sorry, but you're really going to have to explain why pumping the current Poll of Polls figures into the Scotland Votes calculator produces no fewer than FIVE list seats for the SNP.

      OK, we know it's not a perfect predictor, but it seems highly unlikely it's as inaccurate as you seem to be saying. Your claim was that not a single update of the Poll of Polls has ever shown the SNP winning even one list seat - that's clearly, unambiguously untrue.

      Delete
    5. I assumed SNP win every seat and all but one constituency seat in each region and used your figures as an illustrative region. The Scotland votes calculator is making distributional assumptions around the constituency vote which they are not revealing. It's perfectly obvious. You can trust theirs if you want. But I can tell you if you use your figures but assign EVERY Labour constituency vote to SNP such that SNP have 72.8% of the constituency vote and Labour has none then SNP somehow still don't pick up 2 seats (the northern isles presumably). Ergo, their model is shite.

      Delete
    6. Well yes, because Labour's support in the Northern Isles is negligible, so giving their vote to the SNP doesn't change the strength of the Lib Dem vote.

      So it's not their model that's shite - it's your understanding of modelling.

      Delete
    7. A model that assigns SNP 72.8% of a vote nationally should ensure that they win all the constituency seats. it doesn't so it's sh ite. I'm perfectly aware of how too many of those nawbags vote. But it's changing. And SNP even on current polls (low 50s nationally) are anecdotally likely to take Shetland and Orkney is close. They've got a ridiculously high figure of Lib Dem concentration in the Northern isles probably based on a historic number (2011) and then adjusted uniformly by the national figure. That's my guess. And it's totally unrealistic.

      Delete
    8. Okay, the SNP got 50% nationally in 2015. That means they got 50% in every constituency and therefore won all 59 MPs, aye?

      Delete
    9. No but on 49.98% of the vote they were a bawhair awa fae winnin all of them including in Orkney and Shetland so on 72.8% any reasonable model would predict they win them all (as would anything over high 50s actually). But keep it up you're really pittin forrit the enlightened case for SNP twice. A shit model like Scotland votes doesn't though. Ergo, it's shite.

      Delete
    10. No, because your 72.8% is the combination of SNP and Labour votes, but doesn't touch the Lib Dem vote. The Labour vote is negligible in the Northern Isles, so although your weird methodology may give the SNP an extra 20% or so nationally, in those seats it's only adding an extra few points - not enough to take the SNP past the Lib Dems.

      These models take local deviations into account - which is the correct thing to do when you consider the Lib Dems still won those two seats despite only getting 5.2% nationally in the constituency vote in 2011.

      So no, the model is not shite, or if it is, it's not because of the reasons you state.

      Delete
  20. I'm thinking of making a come back. I have lots on Spanners I want to tell everyone.

    ReplyDelete
  21. Angela Kerrigan is a nutter! Best not to engage!

    ReplyDelete
  22. The Scotland votes calculator has some non-transparent assumptions about how the constituency vote is allocated. Basically, it probably gives liberals the northern isles, Edinburgh South and lots of seats in South region. SNP assumed to get a list seat in NE, 2 in highlands and 2 in South or something like that. Totally non-transparent methodology. And I wasn't recommending a Green list vote in South or NE anyway as SNP isn't likely to win all the constituency seats there. Funny, how you trust their guesses as to the distribution but disparage my distribution assumptions that come from ACTUAL survey data rather than made up and hidden behind a front end. Is it because it gives you the result you want?

    ReplyDelete
    Replies
    1. Oh, for heaven's sake. How are your assumptions any more transparent? How on earth do you justify your apparent assumption that the current Poll of Polls figures would translate into the SNP winning at least 72 out of 73 constituencies? It's like you're clicking your fingers, and saying in Jean-Luc Picard fashion : "Make it so."

      I presume you'll concede that if you're wrong about the constituencies, the claims you made about the list are by definition also wrong?

      Delete
    2. I'm campaigning in the South of Scotland. It will take a miracle of biblical proportions to win all the constituencies here. The constituency I happen to live in is relatively safe though, so much so that our candidate isn't even on the list.

      I'm going hell for leather for the list votes, because that's what we need to counterbalance the constituencies we'll probably lose. If pig-headed contrarians like Jam are listened to to the extent that we don't pick up the list seats we should get according to our general level of support, I swear I'll murder someone.

      Delete
    3. I was advcating an SNP list vote in South Scotland for that very reason Rolfe. Lothian where I'm advocating a Green vote is not South Scotland. Mid and Fife, highlands and Glasgow where I'm advocating wait and see are not South Scotland. NE I'm advocating SNP as I am in Central and West. They're not South Scotland either. I'm going to put my trainers on and go for a run. My trainers are not South Scotland either. After that I'll have my dinner. I'm not having South Scotland for dinner tonight.

      Delete
  23. I don't claim that i make a claim of an illustrative region that SNP win all the constituency seats or all but 1. there is no region in Scotland that has 73 constituency seats James. they have 8, 9 or 10. We don't know who gets them James so we have to assign them. You might like to use Scotland votes which has a threshold so high in some seats that using your own figures but assigning all the labour vote on the constituency to SNP Liberals still hold the northern isles despite SNP having a whopping 72.8% nationally on the list. Their model isn't credible. It's obviously rubbish. try it for yourself. You might like it as it assigns snp list votes but it's clearly not based on reality.

    so I developed a simple illsutrative and transparent way of doing so using scenarios. You could use past monthly polls for party specific highs and lows.

    ReplyDelete
    Replies
    1. Well, if you think the Scotland Votes model is rubbish and non-transparent, I'd suggest it's incumbent upon you to explain your own model in quite a bit of a detail, because so far you've been extremely vague about it.

      For example, the Scotland Votes calculator says that the current Poll of Polls would give the Dumfriesshire constituency to the Tories - which intuitively seems pretty reasonable to me, although it could be a very tight contest. On what statistical basis are you concluding that the SNP would win it instead? Are you using a crude extrapolation of the national swing, or what?

      Delete
  24. And my contentions is in a region where SNP win all or all but one of the constituency seats there isn't a single monthly poll of polls on the list that gives SNP a list seat. But they all give Green a list vote. Don't they?

    ReplyDelete
    Replies
    1. My answer is no. If you think you're right, explain your working.

      Delete
  25. I'm using your most recent list poll of poll figures and giving all the seats to SNP (first scenario) and all but one (2nd scenario). In the 2nd scenario you can assign the other seat to any of the 3 nawbags and the result os the same. No list SNP seat. It's obviously just illustrative but if anywhere is like that then SNP get nothing on the list. There are problems with that assumption in Highlands and South on the constituency - both have a couple of dodgy seats. In South green isn't high on the list so vote SNP list but in highlands it could be so watching brief here. Everywhere else I think my constituency seat assumptions are reasonable.

    ReplyDelete
  26. Is anyone likely to do region by region polling at any point? The Greens' national share is likely to be substantially lower than what they get in Lothian, for example. The Lothian region this time around is particularly interesting because Margo's votes - presumably mostly people who voted SNP in the constituency vote - are up for grabs. I suspect Sillars will say people should switch to RISE, but the Greens look like they could be in with a good shout of getting a second seat with only a relatively small swing. In that case, might "tactical" voting for the Greens be a fair bet?

    For me, the other regions get a lot sketchier. The Highland vote is distorted because current independents were elected as SNP list members. (I personally think this is really unfair. If you're elected as a list member of a party and you leave the party, you really shouldn't keep your seat. It's not like you got there on any kind of personal vote.) Glasgow is an odd one too. Tommy Sheridan might have some personal sway, but the SSP vote, such as it is, is now with RISE. All of that points to SNPx2 for anyone whose main goal is a pro-independence parliament in most places. But Lothian presents a different option. So taking off your SNP hat for a minute, which do you think offers Margo voters the biggest bang for their buck - SNP or Greens on the list?

    ReplyDelete
    Replies
    1. Honestly think it depends on why they voted for Margo in the first place. I.e. Which of the policies she espoused that were most important to the individual voter. Then look at the parties available and select the closest match. E.g. If independence is number 1 issue then SNP; if land reform maybe SGP but will they be in a position to do anything about it?

      Delete
    2. There's no answer to that question. You're implying that Margo voters can switch as a bloc, but of course in reality they're all individuals, and part of the challenge (if anyone is reckless enough to think about this "tactically") would be to work out what your fellow Margo voters are actually likely to do in the real world. My guess is that more of them will go to the SNP than to the Greens. It's also possible that Margo's support included traditional Labour voters who would never normally have voted SNP - although what they'll do now is anyone's guess.

      Delete
  27. I went to a small community hall meeting during the IndyRef where the only speaker was Patrick Harvie. To be honest he spoke in riddles and his policies seemed so far removed from reality but what really put me off the Green party was when he was asked how committed he was to independence for Scotland and he replied, Well, I'm not really that bothered about that, to be honest'.

    Since then, I'm not really that bothered about the Scottish Green Party. I'll vote for green policies once we have independence.

    SNP/SNP

    ReplyDelete
    Replies
    1. Very good point. For me the Greens are a Unionist Trojan Horse. Fluffy Mundell's win is a case in point. SNP X 2.

      Delete
  28. TNS do this:

    http://www2.tnsglobal.com/l/36112/2015-12-18/f92hf8/36112/140613/TNS___Holyrood_Voting_Intention_Poll___21st_December_2015.pdf

    It has sample size issues for some regions but there are timelines which offer some comfort. See p5 and p14 (list). I did a 95% confdience interval on the green 24% list figure in lOthian and it had a bottom end of 15%. There could be other sampling frame error issues too who knows..

    Ipsos Mori has interesting demographics if you wanted to correlate with actual data by seat and region for the really interested ha!

    https://www.ipsos-mori.com/Assets/Docs/Scotland/scotland-opinion-monitor-august-2015-tables.pdf

    ReplyDelete
    Replies
    1. To state the bleedin' obvious, regional subsamples of 100 or so people are not statistically reliable. As I used to say in the run-up to the general election, subsamples are better than nothing in the absence of anything else, but to have any real confidence about the likely distribution of list seats you'd need something a hell of a lot better than that.

      Delete
  29. Point of order: You are only allowed 12 names on the list, if 6 or more of the constituency candidates were on the list as well. then there won't be enough people to fill the 7 list spots. In a region with 17 seats where all the constituency MSPs were on the list as well, then even with 100% of the vote there would have to be 5 MSPs chosen from the other parties. Even if that meant by lots.

    ReplyDelete
    Replies
    1. I'll give you a badge for a legitimate point of pedantry, although of course in principle the SNP could win all seventeen seats if they put up enough non-constituency candidates on the list. The only reason they don't do that is they know they're not going to get 70%, 80% or 90% of the vote (ie. the same reason they typically only put up two candidates in local government wards that elect three or four councillors).

      Delete
  30. Snap, Bibbit.
    I heard him say exactly the same thing in Knightswood, Glasgow. My question was how did the speakers, 2 for, 2 against, one of the fors, being Mr. Harvie, view Scotland.
    Can't say I was surprised.
    I thought, nae votes for them, ever. Nice to see it confirmed.

    ReplyDelete
  31. Glasgow Working ClassFebruary 15, 2016 at 7:42 PM

    Anything less than Both votes SNP will put the SNP majority at risk. Simples.

    ReplyDelete
  32. There is a massive problem here, based on one very simple mistake:
    Yes, the SNP are forecast to win 95% of constituency seats BUT that does not mean they will get 95% of votes.
    It's possible to win first past the post constituency seats with less than 50% of the votes. Look at the Tories in Westminster, who ended up with a majority of seats even though a majority of people didn't vote for them, UKIP and the Greens complained that they got 5 million votes between them but only 1 seat each.
    See https://www.youtube.com/watch?v=-CKfS_Wb9Pw for a simple explanation

    ReplyDelete
    Replies
    1. Bill, you've completely misunderstood the article, and the discussion that's taking place here. It did briefly cross my mind to add the disclaimer "please note that this is a purely hypothetical example to demonstrate how the voting system works, and that there is absolutely no chance whatsoever of the SNP getting more than 60% of the vote", but I thought that point was so obvious it didn't need to be said. Except in dictatorships, no party ever gets 95% of the vote. Not even the ANC has got close.

      Delete
    2. James
      Your numbers only work if SNP get 95% of the vote, which we both agree is impossible.
      You've made a mistake.
      Recalculate with SNP getting 60% (even that is unlikely) and you will see where the other parties get their votes and seats from.

      Delete
    3. Seriously, Bill, you are MISSING THE WHOLE POINT of the blogpost. I've just explained it to you very clearly - if it's still going over your head, then I fear there's not much more I can do.

      There's always one.

      Delete
  33. How much do you trust the polls after the GE. They said it would be a hung parliament. They were wrong.

    SNP X 2

    ReplyDelete
  34. A bit o class fae the working man fae Glasgow.

    ReplyDelete
    Replies
    1. Glasgow Working ClassFebruary 15, 2016 at 9:19 PM

      I am the Glesga Wurkin Class not that idiot impersonator coward in hiding. Vote Labour and keep the two Tory parties out.


      Delete
    2. Glasgow Working ClassFebruary 15, 2016 at 9:52 PM

      No you are not I had this name long before you abused it.

      Delete
  35. There is something completely wrong about 'Glasgow Working Class's' identity being taken over by someone else. His post at 9:19 is clearly the person we have come to recognise, the post at 7:42 is obviously not.

    Quite why people feel the need to assume someone else's identity is beyond me. It is maybe one of those post modern things where there are no identities and there are no opinions, because some idiot can subsume them.

    Whoever is subverting his identity, gonna give it a miss? It makes you, the anonymous one, look like an idiot.

    ReplyDelete
  36. MY CAT'S BREATH SMELLS LIKE CAT FOOD

    ReplyDelete
  37. Which makes a mockery of communication. The above comment at 12:20 AM was not made by me. It is the simplest thing out to play this identity theft game. For the absence of doubt, no comment left here using my name will have been from me for the forseeable future, they will all have been from the identity thief. It is the oddest hobby, but as I have no copyright on my name there is nothing to stop someone doing this.

    Best wishes James.

    ReplyDelete
  38. You get just about what you deserve with d'hont.

    Get 50% of the list votes and in a 'normal' place (9 constituency, 7 list = 16 in total) you would deserve 8 seats. If SNP win 9 constituencies they would keep those 9, slightly more then their list vote justified

    If Labout then got 25% of list they would get 4 list seats. 4 out of 16 is fair for 25%.

    If Tories got 12% they'd get 2 list seats. Again deserved as 12% of 16 is 2.

    If the next party gets ~6%, they will get the last list seat, and 6% of 16 is one.

    It all works out perfectly.

    Doing a sample calculation with 95% going to one party may prove a technical point but it isn't really anything more than being technically correct.

    Generally if the SNP get all 9 constituency seats they will need to be over 55% to get a single list seat. This can be seen by using your examples.

    I would use the % figures rather the actual expected figures. It is easier to understand/follow what has 'happened' to the numbers.

    So suing latest poll guesses, lets say SNP win all constituencies, and divisor will then be 10.
    SNP list 55%, becomes 5.5%
    Labour at 20%, had 20% first pass, if they win, it become 10%, then third pass would be 6.7%, 4th pass would be 5%
    Tories at 12%, so 12% first pass, 6% second, 4% third
    LD at 5.6%
    RISE 6.2%
    Others 1.2%

    First pass Labour 20% wins
    Second pass Tory 12% wins
    Third pass Labour 10% wins
    Fourth pass Labour 6.7% wins
    Fifth pass RISE 6.2% wins
    Sixth pass LD 5.6% wins
    Seventh pass SNP 5.5% wins

    SNP get 10/16 (62%) for their 55% list vote
    Labour get 3/16 (18%) for their 20% list vote
    Tories get 2/16 (12%) for their 12% list vote
    RISE get 1/16 (6%) for their 6.2%
    LD get 1/16 (6%) for their 5.6%

    ReplyDelete
  39. RISE are a good laugh. As long as they're around, we know where all the lefty loonies are.

    I don't care what James Kelly wants, I say SNP 1 & 2.

    ReplyDelete
    Replies
    1. The reason I don't like it is that it conveys the inaccurate message that the list vote is a second preference, which is counter-productive from the SNP's point of view. It's better to get across the fact that the list vote is in many ways the more important vote, and in most circumstances will decide who forms the government.

      Delete
  40. Hmmm... one party winning all the seats. That sounds like a healthy kind of democracy to aspire to!

    ReplyDelete