Sunday, March 29, 2015

New South Wales Postcount: Lismore

Lismore (Nat, 24.3%): Thomas George (NAT) vs Adam Guise (GRN) and Isaac Smith (ALP)
Assessment:  Nationals favoured to retain.
Result (provisional following button press) Nationals retain

Original Article:

As my paragraph about the Lismore postcount got longer I decided to give it a thread of its own so the potentially many updates do not swamp the main postcount thread.

Lismore is one of the far north-east NSW seats where there have been massive swings to Labor and the Greens as a result of a combination of demographic change and concern about coal-seam-gas and other mining projects. 

This seat caused a lot of confusion on the night as it swayed between being projected to the Greens and Labor on the ABC website.  Today it has swayed between being projected to the Greens and the Nationals.  

What is known is that current primaries as of Sunday are 39.9% Nationals, 29.4% Green, 25.4% Labor, 2.9% Christian Democrat, 1.5% Animal Justice and 1% No Land Tax.  

While the gap between the Greens and Labor may narrow in post-counting, four points is too large to close down and the preferences of the tiddlers (especially the AJP) will most likely be slightly worse than useless to Labor in that regard.  So the seat is between the Nationals and Greens.  

We know that Labor were leading the official 2PP count against the Nationals roughly 52:48 before it was taken offline.  Unfortunately I don't have a record of the primary votes at that point.  I infer that the votes added today would have decreased that lead to about 50:50.

What we don't know is the Nationals vs Greens 2CP.  Currently, the figures showing on the ABC site are ABC estimates only.  Hopefully we will get an official Nat-Green 2CP sooner rather than later, but I don't yet know if or when that will occur.

The key stat here is the gain rate: the rate at which Guise must gain votes over George for each preference cast.  At the moment it's .342 and 82% of the preferences to be thrown are Labor's.  Under compulsory preferences this would be a piece of cake as only a 67% flow would be needed.  Under optional preferencing, if half the Labor and minor party preferences were to exhaust, Guise would need 84% of the rest, which is still quite possible, as we saw in Prahran.

The only comparable situation involving competitive candidates as opposed to lopsided seats last election was in Balmain, where the gain rate to the Greens was only .191, with 58.8% of Labor preferences exhausting and the rest splitting 77:23.  But there was a lot of ALP-Green scrapping in that contest, and it's entirely possible for many reasons Guise will do very much better than that.  Labor's how to vote card in Lismore recommended numbering all boxes with Guise second, and stressed "TO STOP CSG NUMBER EVERY SQUARE".  That and the mood of the electorate seem capable to me of generating a flow from Labor to the Greens similar to the Queensland flow from the Greens to Labor, which is about the scale of flow required to win the seat.

I should also mention Ballina.  Currently the Greens' required gain rate there is only .113.  At the previous election in Ballina the Greens recorded a gain rate of .150 from all third-party preferences, a higher proportion of which were non-ALP than this time, without the assistance of CSG and with a Nationals incumbent.  The Greens' task in Ballina (on current primaries) is a piece of cake by comparison to Lismore.  

At the moment, pending any scrutineering insights or informative official re-throws, I see the asking rate in Lismore as gettable.  With the large size of the postcount it may blow out further - for instance if it goes over .4 that should be too hard.

Poll Bludger relates that postals counted so far have been exceptionally helpful to the Nationals incumbent, and if this continues to be the case the Nationals lead may blow out to uncatchable proportions.  However, very few booth prepolls have yet been counted in Lismore.


30 March 12:15: Nothing to see by way of new figures yet, but note the comment by reader Ifonly re the strong 2011 differences between the on-the-day vote and the prepolls/postals etc in this seat.  It is especially a poor sign for the Greens that in 2011 prepolls were 10 points more favourable to the Nationals, and 8 points less favourable to the Greens, than booth voting.  Prepolls will make up a greatly increased proportion of the remaining votes.  However I would expect at least that the Nationals advantage on prepolls will be proportionally lower than in 2011, simply because of the greater and more representative range of prepoll voters.  The question is how much the picture will have changed and how hard Labor and the Greens campaigned for prepolls in this seat.

A way of looking at it is this: if things stay roughly as they are the Greens are slight favourites, but if it blows out significantly they will have no chance.  Without knowing which of these will be true yet, the Nationals seem more likely to retain the seat.

2:30 I am currently reviewing Ballina where a massive postcount blowout in the Nationals' favour has made the Greens' task much more difficult, with the required gain rate climbing to .290 and the Green candidate in mild danger of being relegated to third.  A separate thread will be added on Ballina later.

5:00 The scathing pre-election comment by poster Nick C on the Tallyroom thread for this electorate suggests we shouldn't expect too much from the Greens' prepoll effort here.  That plus the prepoll blowout in Ballina (which is at a more advanced count stage) suggests this one is about to go pearshape for the challengers.

5:05 Yes and checking for new figures I see now that it has.  With most prepolls in the Nationals have jumped to 41.82% with the Greens on 27.28 and Labor on 25.61.  The required gain rate for the Greens is now 0.471 votes per vote.  That will almost certainly come back when absents are added, which tend to favour the Greens and hurt the Nationals, but I greatly doubt that it will be enough.  The Nationals are highly likely to retain this seat.

5:40 With the gap between Labor and the Greens closing, the ABC has now gone back to modelling the contest as Nationals vs ALP.   With the gap between the two down to 1.1 points now, I'm not so sure anymore that the Greens will be second, but if absent votes help them as usual they still should.  Anyway it makes little difference.  The required gain rate is now .486 if the Greens are second, or .503 if Labor is, both of which are too steep under OPV.

8:55 Or maybe not!  Antony Green has the example of Noosa in Queensland at which a gain rate of about .5 was in fact recorded by the Greens on ALP preferences.  He also notes that on the night combined preferences from the Greens and minor parties in Lismore were flowing 62% to Labor, 8% to Coalition, and 30% exhausting, a gain rate of .54.   On that basis, if they flow about the same from Labor to Green (as was the case in Prahran) it can't be ruled out that the Greens might win even from about as far back as they are now, and will have serious chances if the gap closes on absents.

We're actually not going to know much of use for a while because, as Antony relates, Elections NSW are conducting computerised entry of ballot papers.

11:40 Tally Room projects that there will be little change in the primaries from here, and projects the Greens still winning narrowly on the assumption that the ALP-Green preference flow is as strong as the other way around.

Tuesday 11:30 Some more primaries have seen the Nationals' lead blow out to 16.2%.  The required gain rate is out to .522 for the Greens or .532 for Labor.  However, the crucial absent votes, likely to favour the Greens and reduce the asking rate substantially, are still not included in the count.

Wednesday 12:20 Antony Green reports "National Party scrutineers are claiming that Labor preferences are flowing 55.3% to the Greens, 6.5% to National [and] 38.2% exhausting".  If that is correct it would be a gain rate of .488 for the ALP preferences only.  That's slightly lower than the current required rate (hence the ABC projecting the Nationals as ahead) but when the absents go into the sample I expect the required rate to drop to well below that figure.

4:30 Still waiting for those crucial absents; meanwhile some more postals (presumably) have increased the required gain rate on current votes to .532.

9:30 Still no absents ... meanwhile the required rate is up to .535.

Thursday 1:20 More trickles being added and the required rate is up again to .537 votes per preference.  Once again, this is almost certainly going to come down, as is the Nationals' current lead on the ABC projection (which, for lack of a disclaimer, many people are still taking as representing real vote outcomes.)

Thursday 9:45 The Tally Room has some scrutineering figures that are very similar to Antony's: Labor preferences flowing 57%-6.3%-36.7% (Green-National-Exhaust) and other preferences 15-26-59.  If those are right then that suggests the Greens' gain rate on preferences will be about .393 votes per preference, perhaps slightly lower as the Labor vote drops off in the postcount.

On that basis if the count stopped now we would expect George to survive by about 1980 votes (ABC projects 1700).  As strong as the Greens' performances tend to be on absents in this region it is not likely such a gap will be bridged.  The remaining hope for the Greens is that the preference samples, being from small booths, are not representative of preference flow.  Still I don't think that would make enough difference and I favour the Nationals based on this information and the current and projected final primary votes.

Friday 3:00 Over 1300 absents are now included in the count and they have broken 39% National, 27% Green, 25% Labor, not vastly different to the main count so far (though that may well change).  The required gain rate has come down a little to .523 as a result, but it doesn't look like it will finish much below .5.  The button will be pressed sometime in the coming week.

Friday 5:10 Two days ago a slightly different set of scrutineering figures was posted by Graham Askey on Antony Green's site (65-6 to Greens on Labor preferences, 20-30 on others).  If those figures were correct the Greens would gain at .469 votes per preference.  I'm doubtful the required rate will even come down that far by the end.

Saturday 6:10 Some more figures went up very late last night and the required gain rate came down to .516.  Over 2000 absents are included and the count is at 89.7% (it never gets to 100 because several percent of voters do not vote.)  There may not be much more to add.

Sunday 10:30 It is hoped the count will be done by Wednesday COB.

Wednesday 10:30 am: The Button has been pressed and Thomas George has provisionally retained Lismore.

Wednesday 1:30: The preference distribution for Lismore has been published.  As expected Labor did not bridge the gap to the Greens for second.  On preferences Guise gained 6834 votes, George gained 1679 and 6084 exhausted.  The total gain rate was thus .353 votes per vote, somewhat lower than the projections off scrutineering samples, showing that the Greens would likely have won if only the votes counted on the day were used, but fell behind during the postcount and were actually less competitive in this seat than the ABC projections.  In all George got nearly 53% two-candidate-preferred.  There are no very close exclusions so no potential for a recount.


  1. Pre-poll, postal and absentee votes tend to be more conservative than the vote on the night.

    Last time the polling night figures for Green were 20% and pre-poll/postal etc were 18%, Nats 59%/62% and Labor 13%/12%
    Last time total votes were 45666, to get to this there needs to be 17,903 postal etc votes
    If these go 3% better for Nats (as they did last time) and 2% worse for the Greens, the final numbers will be 12,937/ 18,053/ 11,254

    Last election there were two seats where Labor preferences were counted. They went 27% Green and 8% Liberal with the rest exhausted.

    If the above figures are close it would take the Labor vote to go 60% Green, 15%Nat and 25% exhausted. I think it would be hard to get such a low level of exhausted votes.

  2. Yes if those projected figures end up being right then the gain rate required would be .455 votes per vote which is very likely to be too high, much higher than the Green to Labor flow in Queensland. As it is it is doable but no certainty.

    I was going to run the usual line that there is an important difference between this time and last time which is that a greater proportion of the uncounted votes this time are prepolls, which are generally less favourable to conservatives than postals. However in this electorate in 2015 they were actually more favourable.

    It is very difficult projecting from a seat that was an obvious Nationals retain in 2011 to a seat that was a live contest for some time leading up to this election. We will have to see what the strength of the Greens prepoll campaign was like.

  3. ABC are calling Ballina a safe Green gain Like you I'm not so sure. The Nats made some ground in the postals and even the iVote the Nats are now 4729 votes ahead of the Greens on primaries There are a lot of ALP and Green leaning independent votes, but ballot papers cast with HTVs as Absentee or prepoll are very fickle

  4. I think its unlikely that the flow of presences from Labor to Green will be as strong as Green to Labor. Labor voters are almost "trained" to just vote 1. Nats had some big boosts from the postals. Labor dos ok but most postal voters don't have a HTV so they either exhaust or do their own thing. I noticed a miner of ALP voters deliberately preference coalition after Labor because they don't like Hreens or minor parties

    1. I would have thought this too, but I also thought it at first in Prahran, and in that case it was wrong: the Labor-Green preference flow was stronger than the Green-Labor one. Of course no electorate is an exact model for any other.

      Antony's example of Noosa (Qld, hence also OPV) is interesting: in 2015 the preferences of Labor (including some votes that went PUP-Labor) went 60.5% Green, 9.3% LNP, 30.1% exhaust. (In 2012 it was just 29.4% to the Greens in the same seat. 7.7% Coalition, rest exhaust in the same seat.) Labor supporters might be less "left-wing" than Greens voters but they are also more likely to just do what their party tells them, whether that is just voting 1 or, as in this case, preferencing the Greens. Clearly based on the Noosa case they are not stuck on just voting 1.

  5. Whilst the Thursday news is bad for the Greens in Lismore it may be a good sign for Ballina. If the preference flows are similar in Ballina (and applying the Labor flows to the ex-Green), the Green candidate would be very happy.

  6. Yes a similar sort of flow in Ballina would be fine for the Greens even on present figures and assuming a reasonable flow from the ex-Green.

    Notably in Ballina the Green to Labor flow is weaker than in Lismore so the Greens would be hoping that is not the same case in reverse as well.