Sunday, March 31, 2019

Legislative Council 2019: Pembroke

PEMBROKE (ALP vs Lib 7.45% - pre-redistribution by-election margin)
Incumbent: Jo Siejka (ALP)

Welcome to the second of my three Legislative Council guide pages this year.  The one for the most interesting-looking contest, the Nelson vacancy, has been doing business for some time, and the guide for Montgomery is now up too.  I will be updating my voting patterns assessment as well but am waiting for the upcoming session to complete in view of the lack of data in the last 12 months so far, so I expect to do that update in the third week of April. [EDIT: I ran out of time to do this before the election!]

And there will be live coverage here of all three seats on the night of the election, expected to be Saturday May 4th.

This piece will be edited through the campaign from time to time for updates, campaign information, added candidates and changed assessments.

(Note: candidates may contact me once only to request a change to the link their name goes to, or additional links which will be added, or not, at my discretion.  No other changes will be made on request except to correct clear factual errors.  Candidates are welcome to comment in the comment section. Differences in the length of different candidate sections reflect differences in amount of available/(in my view) interesting material.)

Seat Profile

Pembroke (see map) is a small suburban seat that falls entirely within the City of Clarence on Hobart's eastern shore.  The electorate extends from Tranmere in the south to Geilston Bay in the north.  A redistribution has removed the most pro-Labor booth of Risdon Vale, making the seat about 0.7% closer "on paper" than the 7.45% margin from last time.  However, that's an unreliable margin anyway (see below). The suburb of Tranmere is wealthy and Liberal-leaning, Warrane is blue-collar and good for Labor, and the rest is all middle suburbia, with a fairly high Green vote in Bellerive.  At the 2018 state election Labor picked up swings from both Liberals and Greens in Pembroke, with the Liberals polling 47.4%, Labor 39.2% and the Greens 11.2%.  (These votes do not include postal votes, which tend to favour the Liberals.)  That places the seat somewhat on the left side of the Tasmanian average, but not hugely so.  The Liberal vote would have been boosted by Premier Hodgman's huge personal vote, so in 2PP terms the seat might be considered a little bit ALP-leaning.

Pembroke is a swing seat that has been held by both major parties back and forth in the last 30 years, with a brief independent interruption.  The resignation of the late Vanessa Goodwin for health reasons in 2017 resulted in the Pembroke by-election (see guide, see coverage thread).  This was a three-cornered contest between the Liberals' James Walker (a local councillor), Labor's Jo Siejka and the local mayor Doug Chipman, an independent and former Liberal Party state director.  In scenes that had seasoned Tasmanian politicos reeling, the Liberals attacked Chipman over his age and lifestyle, and seemed more concerned about beating him than beating Labor.  The attack succeeded in getting Walker into second by a very small margin, but destroyed any hope of a friendly preference flow from the mayor.  The Liberals were thrashed on the 2PP count (Chipman vs Labor would have been 52.4% to Labor) and later apologised.


Jo Siejka (ALP) (Facebook, Twitter, candidacy announcement)  is the most recently elected of the four Labor MLCs.  Prior to winning Pembroke, Siejka had no previous electoral form that I could find.  She had earlier been the CEO of the Youth Network of Tasmania, Chair of the National Youth Coalition for Housing, and a board member of TasTAFE.  In a recent Shadow Cabinet reshuffle, Siejka was promoted to be Shadow Minister for Disability and Shadow Minister for Ageing.  The Ageing portfolio is especially relevant to the Pembroke electorate, which does not have a high average age overall but which has some suburbs (eg Lindisfarne) with high retired populations. In 2017, Siejka polled 32.4% of primaries in a field of 7 candidates, and received 54% of preferences to just 36.5% for Walker, with the rest exhausting.

Siejka's brief parliamentary career so far has been uncontroversial, but she has often been seen as surprisingly low-profile, especially in contrast to neighbouring MLC Sarah Lovell.  Siejka has had a lot more media over the last six months but she would still not be on the list of Labor household names across Tasmania generally.  She has flagged health services, TAFE, penalty rate cuts and cost of living pressures as issues of concern. Siejka lives within the electorate.


Kristy Johnson (Liberal) (Facebook, linkedinACL questionnaire, candidacy announcement) is the owner-manager of the all-female 24-hour Fernwood Gym in Bellerive (within the electorate), and also of Glenorchy Health and Fitness, and has 20 years' involvement in the fitness industry.  Johnson, not to be confused with Glenorchy Mayor Kristie Johnston, contested Denison (now Clark) for the Liberals at the 2018 state election, polling a very respectable 3234 votes at the first attempt.  Her vote was far higher in the Glenorchy booths than the Hobart booths, but this was probably caused by Sue Hickey being the former Lord Mayor for the latter rather than the name-confusion issue.

As of the state election Johnson lived in West Moonah, on the other side of the river. Johnson has flagged support for parents, education and literacy and a 20-year infrastructure plan as issues.

Tony Mulder (Independent) (Twitter) was MLC for the adjacent seat of Rumney from 2011 until his rather narrow defeat by Labor's Sarah Lovell in 2017.  Mulder is a former police commander and was a Liberal state election candidate in Franklin in 2010; for more on the complex history of his association with the Liberal Party see my Rumney preview.  After contesting Prosser (where he finished a disappointing fourth with 9.7%, well behind competing independent Steve Mav), Mulder romped back onto Clarence Council, polling second on primaries for Councillor and also finishing second to incumbent Chipman in the mayoral contest (44:56).

I've rehashed my description of Mulder as a forthright and somewhat grumpy contrarian with small-l liberal/libertarian tendencies often enough that anyone interested can click on the Rumney link above for more details.  At this election Mulder is describing his position as of the "sensible centre" and a moderating influence over party politics, and is running especially on the independence of the Upper House.  Other issues flagged by Mulder include a proposal for a new hospital at Cambridge, cost of living issues, public sector wage setting and poker machines.  On the latter Mulder's position is unusual in a very polarised debate - he supports the freedom to gamble but also supports stopping addictive programming, reducing the house edge and stopping "the government rake-off". Mulder lives in Howrah, within the electorate.

Ron Cornish (Independent) (candidacy announcement) was a member of the Tasmanian House of Assembly for 22 years.  First elected to Braddon for the Liberal Party in 1976, Cornish was a solid electoral performer for the party who was re-elected in Braddon six times, generally comfortably, though he never polled a quota in his own right.  Cornish held 14 different portfolios at various times (see list), most prominently Attorney-General in the Ray Groom government.  He was one of six Liberals to vote against gay law reform in 1997. He has remained vocal about gun control, introduced by the Rundle government following the Port Arthur massacre and recently topical.  Cornish lives at Rose Bay, within the electorate.

Cornish quit the party in 2014 after Tony Abbott decided not to proceed with changes to racial discrimination law (the "18C" controversy). He appears to be running to the right of the party on "political correctness" issues, and is also opposed to the Lower House's recent amendments concerning genders on birth certificates.  His campaign has created some bemusement because its centrepiece is opposition to federal Labor's franking credits changes. This is a federal issue of concern to quite a few Pembrokers (disclosure: especially my dad!) but which a Tasmanian MLC could have no impact on even if they were declared elected before the federal election.  Cornish states he's running to give voters a way to protest, and possibly influence Labor policy.  The Cornish campaign continues a trend of nostalgia candidates, with a few names from the past also cropping up in Prosser last year.

Carlo di Falco (Shooters, Fishers and Farmers) (announcement).  is from Forcett in the adjacent electorate of Rumney and also ran for Pembroke in 2017, polling 3.1% (which was actually more than I expected in an area where the SF+F normally poll poorly).  He is a target shooter, hunter and gun collector.  A previous bio said he has "been involved in the State National Service Rifle discipline for 6 years hosting a National event in the position of discipline chair in 2013." He has written op eds in the Mercury arguing against gun control and to raise concerns about restrictions being placed on gun owners because of thefts.  He has also appeared on Tasmania Talks.

At this election, di Falco has flagged support for "proscribed" (I think he means prescribed) burnoffs in the wilderness World Heritage Area, strengthening whistleblower legislation and increasing funding for the Ombudsman and Integrity Commission. He has also flagged an intention to work for better mental and preventative health services, and to better represent "rural voters".  (There are no rural voters in Pembroke, though I think I've seen the odd sheep round the edges.)

Not Running

Perennial candidate Hans Willink has taken the Sherman pledge on this one with the memorable line "the more people get to know me, the less they appear to like me."

Also out: Chipman, Richard James, Bill Harvey.

The Greens have shown no signs of interest in what is not a very strong area for them.


Some of the issues canvassed in the Nelson campaign may also crop up here.  Some others to be raised in this one are as follows:

Traffic Congestion: Traffic is a long-running issue on the Eastern Shore and was a factor in the fall of neighbouring Rumney to the ALP in 2017.  Labor has accused the government of inaction on the issue, with Siejka bringing up the Mornington roundabout as an example.  They also have organised a petition leading to an independent review.  The same roundabout was also a political football when Goodwin won the seat with Labor in office back in 2009.  Tony Mulder supports a Flagstaff Gully link road and suggests the eastern shore has been "forgotten", and also alludes to a mysterious funding commitment somewhere in or near the electorate that the government has not detailed.  The Liberals have promoted an upcoming ferry service.

Balance of the Legislative Council: The Liberals are likely to continue the tactic seen at many recent Council elections of trying to attack perceived obstruction of the Government's mandate in the Upper House by Labor and the four left-wing independents.  Johnson refers to these eight on Facebook as the " 🛑/ bloc" (highlight the red x and look what happens!)  So far the Liberals really haven't had a lot of success with this message, though perhaps that's just because they haven't found the right Facebook icons previously.  Mulder is running against both parties and promoting independence.

Firearms: With Cornish in the campaign, firearms were bound to appear, with Cornish slamming the Liberal Party's handling of the review of gun laws as "inept".  Unsurprisingly di Falco has a different view, claiming that the proposed changes contentious at the end of the state election campaign are a non-issue.

Dying with dignity: In the ABC candidates' debate, all the candidates showed various only subtly different shades of in-principle support subject to safeguards.

Mandatory sentencing: The Liberal Party supports mandatory imprisonment for child sex offenders, a measure blocked by the Legislative Council. The rug was pulled out from under the Liberal Party's feet on this issue in the final week when their own rebel Speaker Sue Hickey voted the government's reintroduced bill down.  None of the non-Liberal candidates support mandatory imprisonment.  Mulder has argued against the measure based on his police experience that those prosecuted are sent to jail anyway.

Other issues will be added as I notice them.


Notes on anything interesting in the campaign, if there is anything, may be added.  I expect this to be a sleepier affair than 2017.  Overshadowing by the federal election will not help with getting media publicity for campaign matters, and in comparison to Nelson this seat has already been pretty much ignored.

The Liberal Party's messaging is interesting - the bio of Johnson leads not with her career achievements but with her status as a single mother and experience of single-parent family struggles.  This is a tactic much more often seen from the left.  It may be a pitch to the lower-income parts of the electorate but may also be a strategy to put distance between this campaign and the last one.

Siejka appears to have a serious advantage over the Liberals in ground game.  She has been doorknocking for some time (not just in the immediate leadup), and has had small doorknocking teams out with thousands of residents already contacted as of mid-February.  At the candidate's forum, Siejka reported doorknocking a staggering 10,000 houses as well as visiting schools and other centres.  Johnson in contrast was not preselected until early March.

I have heard that Mulder signs are very common in the electorate, Siejka signs are fairly common and Johnson signs are rare.

Mulder is running partly on his work rate.  A letterbox flier compares his presence on the floor of parliament in terms of speeches, amendments and Hansard mentions to Siejka and Prosser Liberal MLC Jane Howlett (neither of whom he is allowed to name).

The Liberals have attracted criticism for including two policies together under the heading "Protecting your children".  One of the policies is described as "Mandatory sentencing for pedophiles" and the other as "Against removal of gender on birth certificates".  Leaving aside that both these criticisms are misleading, the placing of the two together implies that making gender an opt-in field on birth certificates is somehow a threat to children that deserves bracketing with sexual abuse.

(Cornish supports the Liberal position on the birth certificate issue, as obviously does Johnson.  In the ABC debate Johnson said voters were confused about the issue, but Siejka accused Johnson of "choosing not to understand" the issue and misinforming voters. Mulder sees it as pretty much a non-issue.)


Labor's biggish margin from 2017 is an inflated baseline.  Aside from the redistribution that has removed Risdon Vale, the 2017 margin owed much to the high-risk nature of the Liberal campaign that resulted in them having no friends in the preference flow.  It's possible that the by-election being called a by-election also had something to do with it (though in some ways all LegCo elections have a by-election flavour), but a straight Liberal-Labor contest without Chipman on the scene would have been a lot closer.  Nonetheless I think Labor should have the Liberals covered.

Whether they have Mulder covered is less clear.  Given the relatively close Chipman vs Labor 2CP last time it could be that Mulder vs Labor would be a problem for Labor if Mulder can make the final two and then benefit from Liberal preferences.  Mulder could poll a good vote here on his home Council turf but the question is whether he can beat Chipman's 20% and then sneak into the final two.  Chipman was higher-profile as mayor but was also dragged down by the Liberal attacks.  Mulder in contrast has had a clean run with no adverse publicity during this campaign.  I think he has some chance but it would be brave to bet against a Labor retain, probably easily.

Sunday, March 24, 2019

2019 New South Wales Postcount: Legislative Council

Legislative Council NSW Summary:

Seats assumed won based on incomplete count Coalition 8 Labor 6 Green 2 One Nation 1 Shooters Fishers and Farmers 1

(2015 results and continuing MLCs were Coalition 9 Labor 7 Green 2 Shooters 1 Christian Dems 1 Animal Justice 1).

Summary of contest:

Close multi-party contest for remaining three seats.  Count is extremely complex.  The ABC website has been projecting seats as won directly off the check count and this is very unreliable.

Labor (#7) and One Nation (#2) currently appear well placed for two of these seats after preferences, though One Nation is vulnerable to leakage.

A preference race will occur between four parties for the final seat: Christian Democrats (now leading), Liberal Democrats, Animal Justice, Keep Sydney Open.   Keep Sydney Open appears to be struggling to catch Animal Justice.  Animal Justice appear to have at least realistic prospects of beating the other two after preferences.

Sustainable Australia was competitive earlier in the count but its primary vote does not appear to be sufficient.

Note: Button press pushed back to Monday 15 April


If you want a go at picking the result yourself, feel free to have a go on the Not-A-Poll in the sidebar (but if you want bragging rights for getting the trifecta, you have to post your pick in comments!)

Donations welcome!

If you find my coverage useful please consider donating to support the large amount of time I spend working on this site.  Donations can be made by the Paypal button in the sidebar or email me via the address in my profile for my account details.  Please only donate if you are sure you can afford to do so.

The comments below are classified Wonk Factor 5/5.


Thursday 7:00

The check count is finished!  Final check count standings for the parties I've been tracking:

Coalition 34.82 (7.660)
Labor 29.69 (6.531)
One Nation 6.90 (1.517)
CDP 2.28 (.501)
LDP 2.18 (0.479)
AJP 1.95 (0.429)
KSO 1.83 (0.403)
SUS-A 1.46 (0.322)
SFF 5.44 (1.218)
Greens 9.73 (2.140)

To highlight how weak the flows between LDP and CDP might be, in 2016 Senate only 1.8% of LDP above the line preferences were #2 for the Christian Democrats, while 4.5% of CDP preferences were #2 for the LDP.  If one of these gets eliminated it's going to be very hard for the other to gain much with the very high rate of exhaust.

I'll be looking in more detail at how realistic it is for Animal Justice to gain .072 of a quota on preferences and will probably post up a new thread.  It seems to me that all of CDP, LDP and AJP have realistic chances here and the outcome won't be callable in advance of the button press.  But based on the large volume of preferences, the increase in preferencing, and the two largest preference sources being left-wing, it seems to me that Animal Justice can get this.

Thursday 12:00

Current check count standings (very nearly finished):

Coalition 35.02 (7.705)
One Nation 6.92 (1.523)
Labor 29.62 (6.517)
CDP 2.28 (0.501)
LDP 2.16 (0.475)
AJP 1.93 (0.424)
KSO 1.79 (0.393)
SA 1.45 (.320)
SFF 5.56 (1.222)
Greens 9.64 (2.120)

The Liberal Democrats need to both beat CDP and stay ahead of AJP.  My modelling below suggests the first is not impossible but might be difficult; the second is not trivial either.

Wednesday 7:00

I've removed some comments from this entry because of a calculation error.

In the 2016 Senate race the Liberal Democrats gained preferences at 1.45 times the rate of the Christian Democrats.  In 2015 NSW the CDP gained .015 quotas on preferences before the Liberal elimination, but that was out of a preference pool worth 6.7% that time compared to about 11% this time, and with only half (15/28 according to Antony Green) the rate of ATL preferencing.  Applying the obvious crude multipliers the CDP might be expected to gain .046 quotas, so if the LDP maintains the ratio from 2016 that would give them only a recovery of about .02 quotas - which is not quite enough.  At present the gap is .026 quotas.

As I started to suspect on Monday, the initial count appears to have been inaccurate, but it seems it was so for some parties more than others.  This especially applies to AJP and CDP who seem to have been undercounted by about 5%, probably because counters doing manual sorting encountered their votes less often, tended to forget they were a counted party and put them in the Others pile.  The Liberal-National and One Nation votes also seem to have been slightly undercounted (1.4% and 2% currently, though this may change with the remaining 300K primaries to go into the check count) while the Labor, Greens and Shooters votes seem about right.  The other possibility is that there is some sort of bias in the remaining votes that will drag these apparently undercounted parties back to scale.

The effect of all this is that it looks like at all points the combined vote for Others parties was about 0.85% lower than my estimates, meaning that the multiplier for LDP and KSO in my live estimates has been about 9% too high all along, so while I had LDP tracking for .57 quotas that should really have been .52.  They were still tracking to be ahead for a long time, but not by as much as it looked,

Wednesday 4:00

Orange has arrived!  Every division is now over 60% counted, with Bega and Goulburn among the least complete.  The current check-count totals are:

Coalition 35.20 % (7.7438 Q)
Labor 29.61% (6.5149)
One Nation 6.88 % (1.5146)
Christian Democrat 2.29 % (0.5039)
Liberal Democrat 2.16 % (0.4747)
Animal Justice 1.90 % (0.4191)
Keep Sydney Open 1.78% (0.3924)
Shooters 5.48 % (1.2066)
Greens 9.60 % (2.1109)

I'll just watch the check count for now as projecting off the live count seems unreliable because of the possible error issue mentioned below, and we should get the final check count tomorrow.  Animal Justice are in this if they're good enough on preferences.

Wednesday 2:45

Ross Leedham has suggested there may be an undercount of the main parties caused by some votes for them being in the Others pile.  If so (and I've felt for a number of days that there was something not quite right with the continual finding that Others were being undercounted in the check count) then this would be good news for Labor, One Nation, CDP and AJP and bad news for LDP and KSO.  It would mean not only that the latter were stuck where they are (apart from any further decline in the rest of the count) but also that CDP and AJP are going to finish slightly better than the main count suggests.

Wednesday 1:45

The initial count is now showing estimated quota figures for the seven listed parties, but the figures are grossly misleading because they add to 22 quotas for the seven, inflating all of them and allowing no room for Others.  Hopefully this nonsense will be removed soon.

The check count quota is up to 174K.  The district of Orange is still virtually uncounted and there are a few more that have not had much counting.  The LDP are on 2.19% (almost exactly where I expect CDP to finish) and KSO are on 1.83% (almost exactly where I expect AJP to finish) - barring large gaps opening up, we may have two preference races there.  The Others vote is still supposedly under-represented in the check count but I wonder if this is really the case to such an extent (we'll find out if the Others parties start rising soon!)  It might be worth noting some of the other Others parties at this point: SUS-A 1.47  VEP 1.06 AC 0.57 SBP 0.68 (percentages).

Tuesday 10 pm

Counting is expected to complete on Thursday, allowing a few days to model things before the button press on Monday! 

Keep Sydney Open have now dropped behind AJP in the check count but it is not clear they will finish there.  KSO are currently on 1.88% in the check count and LDP on 2.22 (precariously ahead of CDP).

The informal vote shows no signs of going down, with non-blank informals now at about 2.5%.  As the blank informals dropped off slightly near the end of the initial count, they may do so in the check count too, so I am still using 2.4%.  On this basis these are the current live estimates:

L-NP 34.71 (7.367), LDP 2.38 (0.523), ALP 29.54 (6.499), PHON 6.8 (1.496), CDP 2.17 (0.477) KSO 2.01 (0.443), AJP 1.82 (0.400), SUS-A 1.56 (0.343), SFF 5.51 (1.212), Greens 9.6 (2.112),

The most undercounted seats remain rural and regional (Bathurst and Orange are below 10% and none of the eleven below 50% are Sydney seats).  The apparent undercount of Others parties may mean that the uncounted votes in Sydney seats include a lot of absents. That's how KSO might still finish above AJP although projected not to.  If the Liberal Democrats can finish at around .510 quotas as per Ross Leedham's projection, then they should be safe from CDP.  They should also be safe from AJP but I would not be completely sure about that, if preferencing has increased markedly from 2015.

Antony Green has tweeted that the rate of preferencing is roughly twice what it was last time, according to his sources.  In comments I noted that the following were the quota gains on preferences in 2015 to the point where the Liberal candidate was eliminated (Liberal preferences will play no role this time):

Labor .085 quotas, Animal Justice .065, Shooters .044, Coalition .042, No Land Tax .025, Christian Dems .015.

In 2019 there will be more preferences thrown - 10.7% on current numbers, perhaps even pushing up to 11.0%, compared to 6.7% in 2015.  They will be going between these parties:

Labor, Animal Justice (or KSO), Shooters, Coalition, Liberal Democrats, Christian Democrats, One Nation

It is basically the same line-up except with One Nation added.

AJP outdid CDP in the preference distribution by by .05 quotas (about .23%) last time.  On a simple ratio to the above they could go right over the top of both CDP and LDP from where they are, gaining as much as .16 quotas (they'll probably need about .11).  As against that, however, their biggest gain on CDP was on one specific preference source (Greens) on which they gained at a rate of about .1 votes/vote; the rest of the time they gained at only about .02.  Once this is plugged in, the AJP probably won't gain much more than they did last time unless there is a very strong flow from KSO or some combination of other micros. But about a 15% flow from KSO could still do it, if they can get that.


Tuesday 13:00 

Further changes to the main count did roughly nothing.  My live standings continue to estimate that the Others parties vote is undercounted and will improve in what remains, but for both LDP and KSO this assumption won't necessarily hold as they are dropping rapidly.  I am also wondering whether the initial count is reliable enough for these purposes or whether it might contain small errors - it will be interesting to square it with the final result.

Here are my current adjusted live estimates:

L-NP 34.71 (7.636 Q), LDP 2.37 (0.522), ALP 29.54 (6.499), PHON 6.81 (1.498), CDP 2.17 (0.477), KSO 2.02 (0.445), AJP 1.82 (0.400), SUS-A 1.56 (.344), SFF 5.51 (1.212), Green 9.59 (2.110)

If we look at the actual live standings in the check count, it's currently CDP .5062 quotas, LDP .4911, ALP .4791 and PHON .4643.  The initial count says that the CDP will tank a bit from here (as they should with more rural seats still undercounted) - but how precise is it?

Tuesday 10:10

The Liberal Democrat vote continues to fall in the check count, now down to just 2.25, only a sliver ahead of C Dems in the main count and behind them in the check count.  The Others parties remain apparently slightly undercounted in the check count but the LDP position is very vulnerable to both CDP and the left micros if they cannot maintain their share of the Others vote.

Monday 10:10

The initial count reached 4,717,570 tonight and is probably just about finished.  Changes were too small to update my live estimates, but Labor has got ahead of One Nation in them, and both LDP (now just 2.28% in the check count) and KSO (1.98%) suffered.  With the Christian Dems and Animal Justice now pretty much "in the clubhouse" on what looks like about 2.17% and 1.83% the question is where those two can finish relative to them.

Monday 3:30:

A lot more has happened while I was writing my federal poll roundup.  The main count has passed 4,663,000 and might have only 50,000 or so (if even that) to go.  The Coalition continues to drop back and the Greens' surplus grows more significant (but is still not worth that much).

The check count has a quota of 133,050. Keep Sydney Open have fallen to 2.00% and the Liberal Democrats have dropped to 2.29% - the latter have been crawling in recent hours.  The Christian Democrats are still ahead of the LDP on the check count on 2.32%.  Based on the main count, the CDP can be expected to drop back a little from here (perhaps in line with the Coalition also doing so).  That is, provided the initial count is accurate as concerns the CDP vote (it is not subject to the checks of data entry in the check count).  My projections are that the Others party vote is slightly under-weighted in the main count at the moment, on which basis the Liberal Democrats should come up.  But maybe that under-weighting (if it's even real) is because the left micros are going to do well on absents to come, in which case the LDP might keep falling.  I'm still not yet sure that the LDP are home vs the CDP, or that one of KSO or AJP won't be able to get over both LDP and CDP.  When whoever loses out of KSO/AJP is cut out, preferences will be being thrown to the other one, ALP and four right-wing parties.  There just might be enough there to get one of them up if there is, say, a 15% flow and similar from the Greens.

The other issue is that the gap between non-blank and blank informals remains stubbornly around 2.4%, so I've set it to that level (for now).

Current live-standing estimates (adjusted for check count skew):

Coalition 34.77% (7.649), LDP 2.39 (0.525), PHON 6.82 (1.500), ALP 29.52 (6.494), CDP 2.17 (0.477), KSO 2.08 (0.458) - still dropping back, AJP 1.83 (0.403), SUS-A 1.54 (0.339), SFF 5.51 (1.212), Greens 9.59 (2.110)

Sunday 5:10:

Counting has kept going with the check count now on a quota of 111,547.  Keep Sydney Open have scored 1.21% in the last c. 500K added.  If that continues for the rest of the check count they will fall to 1.73%, which would probably put them behind Animal Justice, in which case they are likely to be out of business.  (Even if they then jump AJP on preferences, the gap to the leaders would be too large.) Ross Leedham's projection yesterday had them very slightly worse than this.

My current live estimates are:

Coalition 34.86% (7.669 Q), LDP 2.46 (0.541), PHON 6.82 (1.500), ALP 29.48 (6.486), KSO 2.20 (0.484) - expected to fall further, CDP 2.15 (0.473), AJP 1.80 (0.396), SUS-A 1.54 (0.339), SFF 5.46 (1.201), Green 9.53 (2.097)

Sunday 11:30:

Counting continuing today.  KSO have continued to drop and are now on 2.19% in the check count (the Liberal Democrats have also dropped slightly).  On the initial count the Greens continue rising and the Coalition dropping back slightly.  In the check count some parties have seen significant declines in their BTL vote, which if it continues might help the Others parties.  These are my current live estimates (not projections of the final totals), using 2.2% for non-blank informals:

Coalition 34.98% (7.696 Q), LDP 2.46% (0.541), PHON 6.80 (1.496), KSO 2.22 (0.489), ALP 29.47 (6.483), CDP 2.17 (0.477), AJP 1.79 (0.394), SUS-A 1.54 (0.339), SFF 5.43 (1.195), Greens 9.46 (2.081)

I expect KSO to fall behind CDP, probably substantially so, over coming days.  We'll need to see where KSO lands to see what hopes it might have of overtaking more than one other party on preferences.

Saturday midnightish:

Quite a bit went into the initial count tonight, with the Coalition dropping back a bit (somewhat irrelevant) and the Greens continuing to build providing a few more preferences for the left parties (though there won't be that much to see there with their excess unlikely to be more than say .1Q, of which no one will get more than maybe a quarter.)

The Others vote in the main count has made gains during today's counting, apparently off absents which are likely to be good for KSO.  That may help them to at least stay above Animal Justice, which in my view is extremely important for what chances of winning after preferences from behind they may have.

Saturday 3:15 pm:

KSO had better start swimming or you know what will happen; they're down to 2.24% already.

Saturday 2:15 pm

Another large increase in the check count which is now up to 99748 votes (about half complete).  KSO dropped to 2.30% in the check count.  The Greens improved in the main count - last time there was a trend towards the Greens and left-micros doing well at the end of the count, but this time it won't help that much as Greens preferences are worth so little.

The size of the non-blank informal rate is quite a significant issue for the position of the parties fighting for the final seat.  It decreased slightly in the check count overnight.  These are my current live estimates if I assume it finishes at 1.7%:

Coalition 34.98 (7.696 Q), LDP 2.543 (0.559), KSO 2.387 (0.525), PHON 6.82 (1.500), CDP 2.17 (0.477), ALP 29.35 (6.457), AJP 1.76 (0.387), SUS-A 1.56 (0.342), SFF 5.43 (1.195), Green 9.33 (2.053).

But those estimates assume that Others parties are currently under-represented in the check count, which seems unlikely.  These are my current live estimates if I assume non-blank informals finish at 2.2%:

Coalition 35.15 (7.733), LDP 2.42 (0.533), PHON 6.85 (1.507), KSO 2.28 (0.501), ALP 29.51 (6.492), CDP 2.18 (0.480), AJP 1.77 (0.389), SUS-A 1.48 (0.327), SFF 5.47 (1.203), Green 9.38 (2.064)

So it is possible that PHON would really be in front of KSO and Labor would really be in front of CDP.

Friday 9:30 pm

Quite a lot happened today.  There was another big update to the initial count, which now has over 4,357,000 votes in it (based on the Lower House count there may be about 360,000 votes to go, not all formal.)  Changes were minor, but the Greens gained .06%.  Secondly the check count advanced to a quota of 88,678 (about 40%).  It has become much more representative as concerns the Shooters vote, though it still has a way to go there.  Keep Sydney Open fell further to 2.38% in the check count.  The below-the-line rates for both Shooters and One Nation dropped.  The other issue coming up is the strong evidence of a higher non-blank informal rate than 2015.  In 2016 non-blank informals were 1.5%.  They are currently running (based on the difference between the two counts) at 2.4%, in a count that's only very slightly skewed to areas of naturally high informal voting.  This skew might be because votes with low informal rates are undercounted in the check count, but I'm getting wary of that assumption, and I've bumped the assumed non-blank informal rate up to 1.7%.

These are my current adjusted standings:

Coalition 35.05 (7.711), LDP 2.55 (.562), KSO 2.41 (.530), PHON 6.84 (1.505), CDP 2.17 (.477), ALP 29.36 (6.459), AJP 1.75 (.385), SUS-A 1.53 (.336), SFF 5.44 (1.197), Greens 9.26 (2.037).

So a big loss for KSO today on that projection, though some of that's caused by changed handling of informal votes.  It might be worth keeping an eye on the rate they gain votes at from here (on a large enough sample).  They're currently on 46,360/1,950,913.

Friday 11:30 am

Again I am working through the day today so updates may be slow but it is notable that the Shooters have climbed to 3.45% in the check count, which means the anti-rural/regional bias in the thing is starting to dilute.  The Liberal Democrats are now on 2.56% in the check count while Keep Sydney Open have fallen to 2.44%.

In other news I am now treating Coalition-8 as an assumed win following further gains in the main count today.  Also, the informal vote in the check count is stubbornly high still at 6.5% - which may suggest Others parties are now slightly over-represented in the check count.

Thursday 11:30 pm

OMG! We have an update in the main count.  About another 70,000 votes have gone into it and Labor has gone backwards a little bit.  This is where I have things currently:

L-NP 35.00 (7.700 Q), LDP 2.60 (.569), KSO 2.52 (.554), PHON 6.86 (1.509), CDP 2.15 (0.473), ALP 29.33 (6.453), AJP 1.74 (0.383), SUS-A 1.51 (0.333), Shooters 5.49 (1.208), Green 9.19 (2.022)

I'm still tracking the Greens in case they bounce up and look like having .1 of a quota surplus, but it doesn't seem like it will happen.  The big mover here is Labor - down.  Even after adjusting for leakage I now have them nearly .04 Q behind One Nation, which they would probably overturn but I couldn't be absolutely sure.  If they drop back to the low 6.4 range, things start looking dicey.

Thursday 2:00 pm

Worth keeping an eye on Ross Leedham's projections based on matching votes gained in the Lower House to counting in the same districts in the upper house, I understand with an assumed 0.9% for KSO in districts not contested. Currently these project KSO to crash out, dropping to below .4 Q, and Sustainable Australia to come up to a point where if the flow between the left (or confused-for-left) micros was strong enough then SA might be competitive (but still I doubt it; I think the micro preferences will flow more to Labor whatever is on their cards).

KSO have now fallen below 2.5% in the check count; a few more tenths of a point will see them drop into the mix with PHON, ALP and CDP.

Thursday 12:00 pm

Very busy with work and meetings today so not expecting to update projections until tonight.  However, in the check count today the Liberal Democrats have moved out to lead KSO by 2.61% to 2.51% (these figures should be only very slightly inflated by pro-Others bias now).  This would take KSO down to .544 Q on my crude projection, bearing in mind that they seem likely to drop further.  The check count now overrepresents the Coalition and Greens but no longer much overrepresents Labor.  Shooters are still very under-represented.

Wednesday 7:00 pm

Quota in the check count is now 65,535.  We're now at a point where the total Others vote in the check count has just about got down to where I project it will finish.  However, my current projection of where it will finish might prove optimistic if, for instance, (i) the non-blank informal rate stays higher than last time or (ii) Others parties generally performed badly in the votes presumably not yet added to the initial count (for instance if these included a lot of postals).  If time permits I may look for more clues on this on the weekend.

One Nation's BTL share keeps going up and the Greens' keeps going down, so Greens preferences look like hardly being a factor at all.  Revised crude projections:

Coalition 34.96 (7.691), LDP 2.56 (0.562), KSO 2.50 (0.549) (may decline), PHON 6.86 (1.509), CDP 2.15 (0.473), ALP 29.39 (6.466), AJP 1.75 (0.385), SFF 5.51 (1.212), Green 9.22 (2.028)

Both majors and the Greens are now overrepresented in the check count, and PHON and especially SF+F underrepresented.  The question for Keep Sydney Open is how well they can hold their projected 2.50% (which has so far been more or less stable) before we get into the really bad seats for them.

There are 31 seats still with less than 10% counted.  KSO contested seven of these in the Lower House, with an average vote of 1.6%, less than half their average in all seats contested.

Wednesday 1:00 pm

Quota is over 57,000 now but the anti-rural and pro-ALP/Green skew is still just as apparent.  The Liberal Democrats and KSO continue swapping positions in the check count for now. I'll hold off til tonight for a projection update because of work commitments during the day.

Tuesday 8:45 pm

Something a bit interesting did happen in today's check-count: One Nation's below-the-line vote as a share of its overall vote rose.  Because BTL votes are not in the currently stalled "initial count", a high BTL to ATL ratio suggests a party is stronger than the initial count suggested.  Early on, One Nation's multiplier to account for ATLs was as high as 1.094; this went down to the high 1.07 range but has now come up again to 1.083.  I'm expecting PHON to drop .02 quotas to leakage, but they're getting into the range where Labor's task to catch them on preferences, while still pretty easy, is starting to be non-trivial.

The other useful thing that's happening is that the total vote for the Others parties is becoming less and less out of whack with where it seems it should be based on the initial count.  It's not far off parity in this regard.  Once it reaches parity, any loss for one of the Others parties (eg KSO) will be a gain, on average, for the remainder (eg LDP).

Here's my current crude projection:

Coalition 34.96 (7.691 Q), KSO and LDP both 2.50 (.550) - KSO expected to decline, PHON 6.84 (1.505), CDP 2.15 (.473), Labor 29.39 (6.466), AJP 1.75 (.385), SUS 1.56 (.344), SFF 5.52 (1.214), Green 9.23 (2.031)

Also, see Ben Raue for an excellent look at how over-represented certain regions are in the count so far.

Tuesday 6:30 pm

Still not a lot to see today.  The check count is at 53,802 on quota but is no more representative than it was at the start of the day as concerns rural seats.  It is becoming much more accurate as concerns the Coalition vote, but still has Labor too high and the Shooters too low.

Tuesday 1 pm

The check count has now passed 50,000 on quota but the massive underrepresentation of rural and regional districts has if anything got worse. KSO and LDP continue to swap the lead among those competing for the last three seats but we really need to start seeing a better representation of rural seats to make the figures meaningful. Also there is still no progress on the main count.  I will hold off updating projections until later because of these issues.

Monday 10 pm

It was a big day in terms of the size of the check count which more than doubled, but not such a big day in terms of learning about the vote strength of the Others parties, because the check count is so skewed.  Also there wasn't any update to the main count, so I wonder if it will be completed.  At the moment all that affects my estimates for parties in the main count is evidence regarding their above the line rates.

Current live projections LNP 34.96 (7.690 Q), KSO 2.52 (.554 - expected to drop), LDP 2.49 (.548), PHON 6.82 (1.500), CDP 2.15 (0.473), ALP 29.42 (6.472), AJP 1.75 (.385), SUS 1.54 (.339), SFF 5.52 (1.214), Green 9.24 (2.033).

I'm keeping an eye on the informal rate in the check count.  This is currently at 6.24%, compared to 5.6% in 2015.  However I think this could result from the check-count being weighted towards divisions with high informal votes as well as from it being against forms of voting with low rates, so I haven't done anything about that.

Monday 4:30

The Liberal Democrats have overtaken KSO on the check count again.  [update 5:30: KSO back in front]

Monday 3:30

Another big update in the check count with the progressive quota now at nearly 40,000 and more votes going in as I write.  These have pushed Keep Sydney Open back in front of the Liberal Democrats but this is again unreliable because the updates included a lot of counting from strong KSO seats including Sydney, Vaucluse, Summer Hill, Rockdale and (albeit less strong) Wakehurst.  Revised projections on the false assumption that the check count is representative:

Coalition 34.93 (7.685), KSO 2.51 (.553), LDP 2.50 (.551), ON 6.81 (1.498), CDP 2.15 (0.473), ALP 29.39 (6.466), AJP 1.75 (.385), SUS-A 1.67 (.337), SFF 5.52 (1.214), GRN 9.25 (2.035)

The graph below shows that KSO are over-represented in the check counts in seats they contested.  None of the nine seats where they ran but polled below 2% in the Lower House have hit 20% yet, while four of their best six (assuming an artificial 10% for Sydney where they actually didn't run) are overcounted.

However Others parties as a whole are still overrepresented too, though only by 9% now, and my projection does take account of that.

Monday 12:30:

The check count has had a large update with the progressive quota now at 31,109.  The main count still hasn't moved.  Revised projections:

Coalition 35.00 (7.700), LDP 2.50 (.549), KSO 2.39 (.526) (may still be over-projected), PHON 6.80 (1.496), CDP 2.15 (0.473), ALP 29.39 (6.466), AJP 1.75 (0.385), SUS-A 1.57 (0.346), SFF 5.51 (1.212), GRN 9.25 (2.035)

Keep Sydney Open's raw vote in the check count has dropped to 2.69% but their projected vote has actually gone up a little bit.  This is partly because the overall vote for Others parties has declined, but also because a very favourable seat for them, Heffron, has become one of the best-counted seats in the check count.  Seats KSO did not contest are no longer overcounted, and among the seats they did contest there is still a slight weighting towards their better seats, so I expect they will actually do worse than my projected 2.39% above.  This is possibly also an issue in the seats they did not contest, because the Shooters, Fishers and Farmers are undercounted in the check count.  However even after accounting for this, KSO might still finish in the low 2% range, which would give them chances to win with the aid of Animal Justice and other left preferences.  Since I recorded KSO's vote early in the count, they have got 1.96% of votes added.

Ross Leedham, who is modelling this count by matching Lower House votes, notes on Twitter that KSO are only just beating their Lower House count over the last 380,000 votes.  Since KSO get votes for free on this comparison every time a seat they didn't contest is counted, that sort of result suggests they're not matching their Lower House count in the seats they did contest.

Saturday midday:

The check count has had a substantial update with the progressive quota now at 21,117.  There is no change to the main count. Revised projection:

Coalition 35.03 (7.707 Q), LDP 2.42 (.532). KSO 2.35 (.516), PHON 6.79 (1.494), CDP 2.15 (.473), ALP 29.41 (6.470),  AJP 1.75 (.385), SUS-A 1.643 (.362), SFF 5.50 (1.210), Green 9.26 (2.037)

I am no longer tracking VEP who have fallen below 1%; they together with Small Business Party, Aus Conservatives etc account for a projected total of 3.7% (0.807 Q) that will be available in preferences to other parties.

Labor and the Greens have been dropping back in the projection because their below the line vote is falling in the check count as it takes in more outer suburbia and less inner urban voting.  I'm not sure how representative that is.

KSO have now dropped .036 of a quota on this projection since the iVotes came in as their share of the Others vote has fallen.  There is a long way still to go and there are some strong KSO areas (such as Vaucluse and the northern beaches) that are now a bit undercounted.

Friday 7:20 pm:

Not a lot has happened today.  There was a thrilling moment when the Liberal Democrats got within 37 votes of Keep Sydney Open on the check count, and I expected a swap for the lead in the race for the final three seats, but it didn't happen and KSO are now 90 votes ahead! (The check count hasn't changed since 2:30).

These are the current projections based on the methods I've been using, but noting that they are prone to various skews especially for the Others parties (as already discussed):

Coalition 35.03 (7.707 quotas)
Keep Sydney Open* 2.39 (.527 Q)
Liberal Democrats* 2.38 (.524)
One Nation 6.79 (1.494)
Christian Democrats 2.15 (0.473)
Labor 29.42 (6.472)
Animal Justice 1.75 (0.385)
Sustainable Australia* 1.65 (0.363)
Voluntary Euthanasia* 1.00 (0.219)
Shooters Fishers and Farmers 5.49 (1.208)
Greens 9.26 (2.037)

The initial count is now at 4.234 million. If it gets to 4.35 million without substantial change in the Coalition's position I intend to call their 8th seat then.

KSO and LDP are on just over 3% in the check count, and KSO have been dropping rapidly.

The Greens preferences looked significant earlier in the count but are tracking to be more or less irrelevant.

Having continually noted things that might cause KSO to go down in the projection I have now noticed something that might pull the other way a bit.  The check count is currently higher on average in seats that KSO did not contest in the Lower House and excluding the Sydney seat (12.67%) than seats where they did contest plus the Sydney seat (10.62%).  Presumably the former would be worse seats for them.  On the other hand, KSO's best 12 seats that they did contest plus Sydney are as well counted as the seats they didn't contest excluding Sydney.  So are the remaining seats KSO ran in (where they averaged 2.4% and where on average below 10% is counted) better or worse for them than the average of their 13 best seats (including Sydney) combined with the remaining 50 seats that they didn't run in?    I think the answer is actually worse, because for it to be better they would have to have polled not more than 1.2% in the Lower House on average in the remaining seats, but they only polled below that in a few far western seats.  But it's possible that KSO support strength is not greatly overrepresented geographically in what has been counted.  We also don't know if the relationship between KSO's upper and lower house votes (or what it would have got in the upper house) is proportionate.

Friday 11:30:

The Liberal Democrats continue to improve in my projection; they are now up to 0.522 Q vs 0.530 Q for Keep Sydney Open.  It may well be that they will overtake KSO soon in the check count.  The Coalition is also dropping back a bit (now projected to 7.696 Q) although they would need to underperform by several points in the remaining main-count votes to miss out.

Thursday 6:33 pm:

The check count hasn't been updated since about 2:30 but the main count has, with Labor now overtaking Christian Democrats on my projection.  Labor would lose a few votes to below the line leakage but nothing significant.

The current projection (normal disclaimers apply, especially to Keep Sydney Open):

Coalition 35.03 (7.707 Q), KSO* 2.42 (.532), LDP* 2.34 (.514), PHON 6.79 (1.494), ALP 29.44 (6.477), CDP 2.16 (0.475), AJP 1.75 (0.385), SUS* 1.66 (0.365), VEP* 1.00 (0.220), SFF 5.46 (1.201), Green 9.29 (2.044)

The Greens are drifting back to the point that their preferences could have virtually no impact (they may not even end up with two quotas).  This is bad for Labor, but there are other parties that may give Labor useful preferences.

Some seats are hugely oversampled in the check count, which is making it skew to Labor. Those with over 40% counted are Bankstown, Blacktown, Fairfield, Granville and Londonderry - all Labor seats (Fairfield has a high Christian Democrats vote.)  Those with 20-40% counted are Blue Mountains, Cabramatta, Seven Hills, Strathfield and Sydney - all Labor on a 2PP basis except for Seven Hills.

Thursday 3:30 pm:

I haven't run a full update but in the votes counted since the update below Keep Sydney Open have taken a large hit in my projection, dropping back to a projected 0.529 quotas.  This is the lowest value they have been at since the iVotes were added and again it is hard to be sure they are going to stay high in the race for the final three seats.  All the top four Others parties dropped but the one that dropped least was Liberal Democrats, who might plausibly overtake KSO on primaries soon.

I am also thinking that the Christian Democrats' position is rather difficult unless both KSO and LDP crumble, because their only natural preference source is Australian Conservatives, who basically have no votes.

Thursday 1 pm:

The quota in the check count is now 15,810 (6.9% counted).  The check count is heavily skewed against the Coalition at present, such that Labor is actually leading on it, although we know from the main count that Labor should actually be about six points behind.  What is notable is that the over-representation of the Others parties in the check count is declining, so that even though the vote for Keep Sydney Open in the check count has slumped from 3.84% to 3.32%, their standing in my projection isn't changing all that much (indeed it is up on yesterday).  They only received 1.53% of the votes since my capture yesterday, so perhaps whatever votes have been added are unusually weak for Others generally.

We still need to see a lot more data but the fact that Keep Sydney Open have actually made a small net gain on projection since the iVotes were added is not a bad sign for them.  Here is the current primary vote projection, which assumes that (i) the current main count is representative (at this stage not much is left to count there, so there is a high confidence level on this) and (ii) the current check count is representative in terms of the proportions of votes for the Others parties relative to each other (this assumption is a lot more dubious because of the iVotes issue and the possibility that left Others are overrepresented compared to other Others, which would be great news for the Liberal Democrats):

Coalition 35.1 (7.722 Q), KSO* 2.58 (0.568 Q), LDP 2.36 (0.518 Q), PHON 6.77 (1.489 Q), CDP 2.17 (0.477 Q), ALP 29.39 (6.466 Q), AJP 1.78 (0.396 Q), SUS A 1.78 (0.392 Q), VEP 1.07 (0.234 Q), Shooters 5.42 (1.197 Q), Greens 9.34 (2.055 Q).

When it comes to preferences, I am still expecting One Nation to lose about .02 Q off leakage from Latham's surplus, which means that One Nation are effectively behind CDP in the projection.  Also, the Liberal Democrats are better placed in the projection now than they were, having overtaken One Nation.  However, the Liberal Democrats often perform poorly on preferences, and lack obvious preference sources in this contest.  While preferences play a limited role in NSW, they would not necessarily be safe with this sort of lead.


The check count is continuing to unroll slowly so I will probably rerun the "live count" projections a few times today, not every time there are slight changes.

Wednesday midnight

It's been a slow day so maybe we'll just get a few percent rolled out per day through data entry (I haven't checked to see how this proceeded last time as I wasn't following so closely then.) Quota is now 14,000.  My latest "live count" projections (again, not adjusted for things I'd love to adjust them for but NSWEC hasn't published the data):

Coalition 35.13 (7.729 Q). KSO* 2.45 (.540), PHON 6.83 (1.503), LDP* 2.24 (0.493), CDP 2.17 (.477), ALP 29.31 (6.448), AJP 1.79 (.394), SUS* 1.69 (.372), VEP* 1.01 (.222), SFF 5.45 (1.199), Green 9.35 (2.057) (* = Others parties)

Note that the ABC site is now projecting seats directly off the check count and has therefore called seats for KSO and Liberal Democrats (and not called the Coalition's seventh when they will clearly win seven and are just about callable for eight).  Both KSO and LDP could well win but to call them would be totally premature in both cases.

Wednesday 8:10: The check count has advanced only slowly today and quota is now 13,729.  (Final quota should be just over 200,000).  Keep Sydney Open are still sliding gradually on the whole, now down to 2.47% in my model (3.59% in the check count now).  I'm noting here their current vote (10,832 of 302,020) so as to track what they get over the next, say, 200,000, as a guide to where they might end up. I'm still not sure how competitive they are.

Wednesday 6:30: These are my latest "live count" projections (not adjusted for vote location or iVote skew) following small updates to both counts.

Coalition 35.13 (7.729 Q), KSO* 2.50 (.549) (may drop), PHON 6.8 (1.496), LDP* 2.24 (.492), CDP 2.16 (.475), ALP 29.33 (6.453), AJP 1.8 (.396), SUS* 1.71 (.376), VEP* 1.03 (.226), SFF 5.4 (1.188), Green 9.36 (2.059).   (* = Others parties)

Keep Sydney Open had been dropping back, but the last update to the check count has reduced the non-Others party total and put them back almost where they were.

There is a possibility that my model is overstating BTL votes for some parties based on where the check count votes are from.  If this is so it would slightly favour the Others parties compared to the above.

Wednesday 4:00: Hoped this would be a great day with votes piling in to the check count but alas still very little to report.  The ABC website is showing Labor's 6th and the Greens' 2nd seats as in doubt because these parties have fallen below the relevant quota (which doesn't take account of some of the Others votes being informals or being BTLs for these parties).  This is just a silly bit of programming; these seats are in no doubt whatsoever.

Wednesday 3:30: Still no update to the check count from the slight change earlier today.  However with another 400K added to the main count it is notable that the Greens and Others have gone down and Coalition has gone up slightly.  This if it continues has implications for Labor's potential to overhaul LDP, CDP and PHON on preferences, and Labor has dropped 0.01 of a quota in my estimate compared to the last update.  I'll post updated estimates when there is meaningful change in the check count.

Wednesday 11:30:

Some more votes are coming in but the check count still has a severe over-representation of micro-parties, so a re-run of projections is unlikely to be useful until the count advances significantly.  I'll be keeping an eye on it through the day and will update when I think it is useful (possibly in an hour or two).  A suggestion from Nick Casmirri is that iVotes (which would be likely to favour progressive micros) may have been added in a chunk last night, so it's quite likely some micros are being over-represented even relative to their booth vote more than others.

12:50 No NSWEC update since 11:30, except a very minor update to overall count that is not worth processing yet.

2:20 Still not much to see.

Wednesday morning:

I'm going to be working on this count on and off through the day because I think today is the day when things could get a lot clearer.  Regarding KSO, Ross Leedham did a neat calculation finding what KSO would be on if their Legislative Council result matched their Lower House result, on the assumption that they ran in Sydney and scored 10% there.  (Note that in many seats they didn't run and scored zero, but in most of those they would have only got 1-2% anyway).  This calculation shows that the current check count (which has KSO on 3.84%) is indeed skewed to them, but not by as much as might be suspected.  Indeed it projects KSO to about 2.8-3%, higher than my 2.5% - which I thought was probably inflated! 

Without more information on exactly how the check count is being conducted I wonder if there might be something going on that is skewing the output in favour of small parties more than would be suggested just from the range of seats included.   An obvious candidate is that the booths counted are predominantly booth votes (and don't include many prepolls on which small parties tend to do badly.)  However another one might be some kind of distortion in the rate in which small party votes are getting into the check count compared with larger party votes.  If this is the case then as the count fills out it will normalise.

Wednesday 1 am (sigh):

The check count just had a huge update (its quota is now 11877, with some votes included from every division), but it included a lot of inner-city votes and has swung to over-representing the Greens, with a massive lift for Keep Sydney Open at the same time.  So now Keep Sydney Open are likely to be over-represented in the check count!  Unfortunately we don't have any seat by seat data on what each party is getting in each seat.  We could use Lower House votes to estimate how under/over-represented the micros are based on the seats included, but this also has to be adjusted for number of Lower House candidates and parties not running in every Lower House seat - it all gets very messy.

I'm going to start posting the estimated adjusted live totals for parties in order of projected spare quotas:

Coalition 34.93% (7.685), KSO 2.51 (0.552) (likely to drop), PHON 6.78 (1.492), CDP and LDP both 2.20 (0.484), ALP 29.43 (6.475), AJP 1.82 (0.400), SUS 1.65 (0.363), VEP 1.05 (0.230) [Note later: unintentionally omitted SFF and Green from this list.]

VEP have gone down further in this projection, and might come up but are not looking competitive now.  KSO are probably over strength given the apparent link between their vote and the Greens', and given that Sydney (seat) is the most overcounted now, but where they will settle remains to be seen.  LDP might be under-represented in the projection now if anything.

The other point is that Labor are improving in their quest to get a seventh seat, on Greens preferences if not on primaries, and might well end up jumping all the micros.

Tuesday midnight:

The check count has been updated, and while the numbers are still unrepresentative they are starting to look more like the overall count.  37 divisions are now included with Granville over 10%.  The main count has had only a trivial update but the check count refines estimates for BTLs (including that Latham's share has gone down a little bit, so the multiplier there is "only" 1.088 now).

These are my new live-count estimates (again using the overall count and scaling all the Others upwards):

L-NP 34.79% (7.654 Q), ALP 29.23 (6.431), Green 9.54 (2.099), ON 6.83 (1.503), SFF 5.31 (1.168), CDP 2.21 (0.486), AJP 1.81 (0.398) Others 10.28 (2.26)

Leading others: LDP 2.63 (0.578), KSO 1.75 (0.385), SUSA 1.67 (0.367), VEP 1.46 (0.320)

I'm not bothering with anyone who doesn't project to over 1%.

So improvements have been made here by Labor and KSO, while LDP and VEP are down.  The question is whether KSO can keep improving at this rate.  With the Green vote in the check count still slumming it at 6.3%, it's quite plausible that they can.

Tuesday: Check Count Starts!

The check count has started early.  At this stage the votes counted are unrepresentative, both within divisions and in terms of the booths chosen within divisions - for instance the Greens are underrepresented in the current count while One Nation are overrepresented.  However this brings a lot more clarity to the final contest.

Firstly it allows me to refine multipliers for below-the-line voting in the Others pile, and the big news here is that Mark Latham has a huge below the line vote.  Currently multipliers for the seven counted parties (ATL+BTL divided by ATL) are: Coalition 1.014, Labor 1.016, Greens 1.047, One Nation 1.094, Shooters 1.031, Christian Democrats 1.057, Animal Justice 1.039.  All these are higher than last time but it's not obvious that anything in the selection of seats counted (only 33 electorates have anything counted at all) is driving it.  So pending any later reduction I'm going to assume for now that it's a real thing and a result of more voters getting used to numbering multiple boxes c/- the Senate election.

Feeding these new multipliers into the revised primary figures for the select seven I now get these estimates of their live primary tally:

Coalition 34.79% (7.654 Q)
ALP 29.16 (6.415 Q)
Greens 9.53 (2.097 Q)
One Nation 6.88 (1.514 Q)
Shooters 5.11 (1.159 Q)
Christian Democrats 2.19 (0.482 Q)
Animal Justice 1.80 (0.396 Q)
Others 10.54 (2.319 Q)

Now we actually have an early check count breakdown for Others, but bear in mind this is from a skewed sample of just 25,747 votes in which the Greens (especially) and Shooters are doing unrepresentatively poorly, while both majors, CDP and One Nation appear to be doing unusually well.  The sample will therefore underrepresent both urban and distant bush electorates and probably overestimates seats in between.  But these underrepresentations and overrepresentations are not going to save the bacon of any party polling way less than 1% in the sample, unless its vote is very regionally concentrated.

The Others total in the Check Count sample is only 8.36% and I project that to rise once all votes are counted.  Applying a uniform multiplier of 1.261 to Others on this basis, this is where the best of the Others are projected

Liberal Democrats 2.94 (0.65 Q)
Voluntary Euthanasia 1.74 (0.38 Q)
Sustainable Australia 1.66 (0.37 Q)
Keep Sydney Open 1.31 (0.29 Q)
Small Business 0.83 (0.18 Q)
Conservatives 0.72 (0.16 Q)

... and I think we can write off everyone else at this point, including Buckingham (0.26 (0.06 Q)

So a live projected primary order in the race for the final four seats is:

1. Liberal Democrats 2. Coalition #8 (these two very close together), 3.One Nation #2, 4. Christian Democrats (these four currently appear best placed)
5. Labor #7, 6. AJP, 7. VEP, 8. SA, 9. KSO (these five chasing)

However, given that the Green vote is undercounted (I project it is only 58% as high as it should be) I would expect that some of the micros in the list above - especially KSO and SA and possibly VEP as well - will rise by more than my 1.26 multiplier, while others (especially the Liberal Democrats) will therefore probably rise by less.  This could bring SA and VEP up towards 2% and KSO up towards 1.5%.  Even so, unless these micros have been even more underrepresented than the Greens, it is hard to get them up to level with One Nation and the CDP.

One Nation will lose votes off Mark Latham's BTLs because of leakage.  With 8.6% of their vote below the line it could be that, say, half of that leaks on Latham's surplus.  However, currently I only project One Nation's surplus to be about 2.3%, meaning that the loss on leakage could be about  0.1% (.02 quotas).  Currently they would still be fighting out third and fourth with the Christian Democrats.

The other thing to bear in mind here is preferencing.  Currently the four parties that are projected to be best placed on primaries are all right-wing, while the five chasing are various forms of centre-to-left.  (Sustainable Australia has right-wing aspects in its anti-immigration stance and its status as a left party is dubious, but its voters tend to preference left-wing parties and can be assumed to be left-leaning themselves - suckers or otherwise.)  The Greens will also have a substantial spare vote and Greens voters are strong users of preferencing.  So there must be a realistic chance that even if things finish as per the above projection, one or two of the chasing pack can get over either One Nation or the Christian Dems (or LDP or Coalition should their vote fall).

For those following preferences, keep an eye on this Tally Room thread re HTV recommendations, but remember that because NSW allows just a 1 above the line, most preferences exhaust, and those that don't frequently scatter.  If Labor are excluded, their HTV card preferences generally go to the Greens (who will very probably be out by then), then Animal Justice, then Keep Sydney Open, and in some seats to VEP.

In 2015, Greens preferences were distributed with 64.8% exhausting, 21.3% to Labor, 11.5% to Animal Justice, 2.9% to Liberal, 1.5% to No Land Tax and 1.7% to Christian Democrats (these were the only parties still in the count).  At that rate on current projected votes, Labor would gain about 0.02 of a quota - which helps, but is not nearly enough by itself.  Perhaps the preferencing rate will be higher this time - the other thing that can help Labor is that the Greens often lift in late counting.


Monday night: I've seen a comment indirectly from Greens scrutineers that Buckingham is "no hope" and I think this figures given that the Green vote appears at this stage to have held up particularly well.

Monday 5 pm:

A large number of votes have been added.  I now have the following estimates of live primaries:

Coalition 34.46 (7.58 Q)
Labor 28.76 (6.33 Q)
Green 9.77 (2.15 Q)
PHON 6.39 (1.41 Q)
SF+F 5.15 (1.13 Q)
CDP 2.14 (0.47 Q)
AJP 1.83 (0.40 Q)
Others 11.5 (2.53 Q)

So the order in the race for the final seat on live primaries is still Coalition, CDP, PHON, AJP, Labor.  However both majors have dropped back and the Greens have come up (not unusual).  The minimum target for other parties is still around Animal Justice's vote.

If precisely one Others party has enough vote to get in, then we could see a contest between Animal Justice and the second One Nation candidate , which Animal Justice would be well placed to win with help from Green preferences.  However depending on how the Others pan out, both or neither could win.

Nick Casmirri has noted that the Others vote is high in the city and especially high in the Sydney electorate, which seems to bode well for Keep Sydney Open.

Intro (Sunday 3:30 pm)

It's still early days in the Legislative Council count but here's a post to try to start making sense of what incomplete information we have.

Let me start by saying that this count is an absolute mess and there are at least three seats we are nowhere near being able to call and will not be near being able to call for a long time.

Concerned about OHS risks associated with working long hours, the NSWEC came up with the following decision: In the on-the-night count, votes counted for the Legislative Council would be classified to:

1. Above the line number 1 votes for Liberal-National, Labor, Greens, One Nation, Shooters Fishers and Farmers, Christian Democrats and Animal Justice.

2. Blank informal votes

3. All other votes.  The others pile includes all votes for the remaining 13 grouped tickets as well as all the below the line votes for individual candidates.  It also includes informal votes that are not blanks.

While I understand the OHS aspect as concerns work on the night, what I don't get is why we have to wait til the check count from Wednesday on for this "others" pile to be sorted even roughly to party.  Especially now when there are fewer close Legislative Assembly seats than would have been expected.

Here's what we know.  2.57 million ballot papers have so far been counted; this is 48.8% of enrolment.  Current primaries excluding blank informals are:

Liberal/National 34.16%
Labor  28.18%
Green 9.04%
One Nation 6.11%
Shooters Fishers and Farmers 4.89%
Christian Democrats 2.09%
Animal Justice 1.71%
Others 13.82%

In 2015, non-blank informals were about 1.5% of all votes that were not blank informals, so on that basis the Others vote contains about 1.5 points of informals.  After removing these from the count, all the party totals including the remaining Others will increase (ignoring further vote movements based on which seats and vote types have been counted).

The Others votes also include below-the-lines, some of which will be for the seven separately counted parties.  Here the 2015 below-the-line rates can be used to determine how much the party totals for these seven should be expanded by to account for below-the-lines.  Thus I multiply Animal Justice by 1.03, Liberal-National by 1.008, Labor by 1.012, Shooters by 1.022 and Greens by 1.037.  For Christian Democrats their multiplier was 1.039 in 2015, but 1.028 in 2011 when Paul Green rather than Fred Nile was the candidate.  I will use the Paul Green figure.

Unfortunately the party I don't have a multiplier for is One Nation.  In 2011 One Nation ran as an unnamed ticket with Hanson herself as head candidate, and the multiplier was a whopping 1.258.  However that seems to have been down to the lack of a party name.  If we look at the 2016 Senate election, the multiplier for One Nation in Queensland where Hanson was the candidate was slightly lower than for the Christian Democrats and slightly higher than for the Shooters.  In NSW the multiplier for One Nation was only very slightly higher than for the Shooters.  Mark Latham is a prominent candidate in his own right so I've decided to use a multiplier of 1.03 for One Nation in this case.

It's possible below the line voting could have become more common following the Senate changes, but there are not many reasons to suspect this.  There are two minor special reasons it might be the case though - firstly, faction-fighting within the Greens over their ticket order, and secondly, the minor cult phenomenon (as seen on Twitter) of voting below the line all the way to put Mark Latham 346th.  However I doubt either of these will have a significant impact.

After adding all this in, this is what the count looks like:

Liberal/National 34.96% (7.69 quotas)
Labor  28.85% (6.35)
Greens 9.52% (2.09)
One Nation 6.39% (1.41)
Shooters Fishers and Farmers 5.07% (1.12)
Christian Democrats 2.18% (0.48)
Animal Justice 1.79% (0.39)
Others 11.15% (2.45)

Bear in mind that these are estimates of the live count and do not account for whether where the votes are from might be unrepresentative.   These totals will move around substantially as more votes are counted.

On raw quotas the current seats are seven L-NP, six Labor, two Green, one ON, one SF+F, leaving four undecided.  These are then decided by the remainders after preferences - and usually in NSW there are not that many of those.

At present of the listed parties the order for the remaining four seats is L-NP, CDP, One Nation, Animal Justice, Labor.  But these last four are all rather close together and the order will shift around as more votes are counted.

The first question then is whether, among the 11.15% split between 13 remaining parties, there is anyone who has polled enough to get into this mix.  Currently, the target figure based on just primaries is to beat Animal Justice, which means getting about 1.8%.

The parties/groups whose votes we do not know are:

Socialist Alliance
Sustainable Australia
Advance Australia
Group G (Seniors United)
Group H (Monaghans)
Group L (Jeremy Buckingham)
Australian Conservatives
Keep Sydney Open
Liberal Democrats
Voluntary Euthanasia Party
Small Business Party
Group S (James Jansson)
plus ungrouped independents

We can get some clues about the competitiveness of Sustainable Australia and Keep Sydney Open.  Both these parties ran in many seats in the Lower House, as did Animal Justice.  We know that Animal Justice is a contender in the Legislative Council.  The current ABC-projected state lower house votes for these parties are:

Sustainable Australia 1.48% (55 seats - average 2.50)
Animal Justice 1.46% (48 seats - average 2.83)
Keep Sydney Open 1.38% (42 seats - average 2.92)

Small parties do not poll as well in the Upper House as in the seats they contest, because there are more small parties, and because they tend to contest favourable seats.  This at least suggests that Keep Sydney Open are competitive with AJP for a possible seat, and that SA could be vaguely competitive, but if Labor lifts there may not necessarily be a seat for any of them.

Preferences could also come into play here, with an attempt by Glenn Druery to network how-to-vote card preferences between these groups that will probably not have much impact - but in a close race any impact at all could make a difference.  It may be that KSO are more successful than AJP in outperforming their Lower House vote in the upper house.  On the other hand, their vote could be negligible outside Greater Sydney (see comments).  On the other other hand, they didn't run at all in the lower house in their single most prospective seat of Sydney.

The Australian Conservatives are also attracting an average Lower House vote to suggest they might be in the mix - in the 19 seats they're contesting they are on an average 2.59%.  However there may be an aspect of cherry-picking in their seat selection, and also their average is obviously skewed upwards by a few seats where there were only 4 or 5 candidates.

I suspect that SA, KSO and CON between them account for something like 4.5-5% out of the missing 11.15%, leaving about 6.5% for the remaining parties.  Socialist Alliance, Monaghans and Jansson are likely to get very little, Flux are never big scorers, VEP got nearly 1% last time.  There might be something approaching 5% between Advance, Seniors, Small Business, Buckingham and Liberal Democrats.  It seems plausible based on all this that the Liberal Democrats might have enough votes hiding in this mysterious others pile to get David Leyonhjelm over the line, but we will have to wait - a long time - to see!

2019 New South Wales Postcount: Lismore

(Link to main postcount thread)

LISMORE (2015 margin Nat 2.9% vs Green 2CP, Nat 0.2% vs Country Labor 2PP)

Key questions:

1. Who will be second after preferences between Janelle Saffin (Country Labor) and Sue Higginson (Greens)

Outlook: Very likely to be Saffin and assumed to be so given that both Nats and Greens have conceded.

2. If Saffin is second after preferences, who wins between Saffin (Country Labor) and Austin Curtin (Nationals)?

Outlook: Saffin

3. If Higginson is second after preferences, who wins between Higginson (Green) and Curtin (Nationals)?

Outlook: Probably Curtin, but there are no direct data on this

Overall outlook: Saffin (Country Labor) strongly expected to win

2019 New South Wales Wrap And Lower House Postcount

The Berejilklian Government has been returned with a very small majority

Expected final result Coalition 48 Labor 36 Green 3 Ind 3 SF+F 3 

Seats changing hands
Barwon (Nat to SF+F)
Murray (Nat to SF+F)
Coogee (Liberal to Labor)
Lismore (Nat to ALP)  link to Lismore thread


This thread will give some summary remarks about the NSW Lower House count (which may be updated if necessary) and will also follow the postcount in a few undecided seats where only two candidates are in contention.  Lismore has thrown up another three-cornered contest that is of special interest and it will get its own thread.

Saturday, March 23, 2019

2019 NSW Election Live Comments


Coalition has won the election - probably with a small majority 

Apparent wins Coalition 45 Labor 35 Green 3 Ind 3 SF+F 3, 4 in doubt
(Apparent wins are not necessarily all confirmed)

Apparent seat changes:

Barwon (Nat to SF+F) - but need to wait for Broken Hill booths
Murray (Nat to SF+F)
Coogee (Liberal to Labor)

In serious doubt:

Lismore (Nat vs Green vs ALP - long postcount ahead)
Dubbo (Nat vs IND - close)

In some doubt:

East Hills (Liberal leading Labor)
Upper Hunter (Liberal leading Labor)

Assumed won but at low doubt levels: 
Wollondilly (Liberal vs IND - no 2CP count available)
Penrith (Liberal leading Labor)

Commentary appears below the double line with latest comments scrolling to the top - refresh frequently for new comments once count starts.  Comment clearance during commentary will be slow and comments may not be replied to until very late.

Friday, March 22, 2019

NSW 2019: Final Day Roundup

SUMMARY: Polls imply likely Coalition victory, but if so, majority status is still touch and go.

This thread will cover anything I think is of interest re Saturday's election, starting with a pre-Newspoll look at where things might stand (but there is some weird stuff going on, so who knows) and followed up with anything I want to add once Newspoll (or any other late polls) come out.  I'll be working in the afternoon, so don't expect updates between about 11 am - 6 pm.

There will be live coverage here on election night from 6:00 pm and going through til very late, and the live thread may be started earlier in the day if there is anything of special interest going on.

Tuesday, March 12, 2019

Poll Roundup: Medevac Fails To Shift The Dial

2PP Aggregate: 53.4 to Labor (+0.4 since last week)
52.9% with One Nation adjustment
Labor would easily win election "held now" with about 90 seats.

In the wake of the Coalition's 50th consecutive Newspoll 2PP loss, here's another roundup of the state of federal polling, and I also include some comments about the state of seat betting, which I track in the approach to each federal election.

A few weeks back, just after the last roundup, there was a lot of hot air about a possible "Tampa moment" for the Coalition in the form of a close Ipsos poll immediately following the passing of "medevac" legislation by Labor and the crossbench.  Since that Ipsos 51-49 (with the Coalition lucky to get 49 on the published primaries anyway) we've seen Newspoll at 53-47 and 54-46, Essential at 52-48 and 53-47 and a commissioned ReachTEL at 53-47.  There was also a Queensland-only YouGov-Galaxy with a 6.1% swing to Labor in that state.

Saturday, March 9, 2019

A Token Post About Modelling The 2019 NSW Lower House Election

Recent state polling suggests a hung parliament (approx 44 Coalition 41 ALP 3 Green 3-4 Ind 1-2 Shooters) - but there's hardly been any of it!

Update Monday 11/3: Polls over the long weekend (a uComms/ReachTEL at 51-49 to Labor and a Newspoll at 50-50) have been completely consistent with the assessments below.  

Rinse and repeat ... another state election is only weeks away and there's been virtually no public polling.  At this stage in the 2015 election cycle there had been five statewide polls, but so far this year there have been just two.  Perhaps they will come thick and fast in the next two weeks, but I have so little hope of that that I think the best I can do is write an article complaining about the lack of polling in the hope that my article becomes out of date as soon as possible.

Anyway, to briefly poke my head through a gap in what is often a very busy few weeks of the year for me (the dreaded "AGM season"), I thought I'd post some comments about where things might be at if the very limited public polling we've seen this year is anywhere near accurate - which it may not be.