Close seats – Friday morning update

We’ve made some more progress yesterday, and I believe we can narrow this list from nine to seven seats today.

As a reminder, the first four columns of data in this spreadsheet represent the AEC’s official data on how many votes are left to be processed in each category (bearing in mind that pre-poll and absent numbers are expected to grow).

Seat summary

I’ve called Chisholm and Cowper for the Coalition today. Gilmore is now clearly leaning towards the Coalition while the Labor lead in Hindmarsh is slim enough to treat it as one of the five very close seats.

So the seat categories are:

  • 72 Coalition
  • 66 Labor
  • 1 Coalition leaning (Gilmore)
  • 1 Labor leaning (Cowan)
  • 5 very close (Capricornia, Flynn, Forde, Herbert, Hindmarsh)
  • 5 others

Assuming the Coalition wins Gilmore, the Coalition needs to win three out of five of the close seats.

Seat Absent Provisional Pre-poll Postal Current Labor lead Projected Labor lead
Capricornia 2004 626 575 6123 476 -428
Chisholm 128 1314 335 1875 -2026 -2626
Cowan 2333 1355 50 740 534 713
Flynn 1729 733 633 6421 646 -1226
Forde 3156 1104 870 4765 -687 -587
Gilmore 0 1425 322 211 -1316 -937
Herbert 1356 1136 465 5404 449 -459
Hindmarsh 0 1599 0 1501 68 578


We now have preferences in 32 out of 70 booths and the trend is very clear. Nationals MP Luke Hartsuyker has a substantial lead over independent Rob Oakeshott, and I’m going to call this as a win for the Nationals.


Labor’s lead has narrowed by about 250 votes, and the model suggests the LNP will take the lead.


Yesterday’s projection had the Liberal lead growing to 2000 votes, and that has happened today, and the model suggests the lead will grow further. I’m calling this seat for the Liberal Party.


While the Labor lead has shrunk slightly, the model suggests they are likely to hold on. We’ve almost finished counting the postal votes here.


The Labor lead was cut by another 400 votes yesterday, and the remaining 6400 postal votes are projected to push the LNP in front.


The Liberal Party has a 687-vote lead, and are projected to hold on to most of this.


The Liberal Party is now leading by 1316 votes, and is likely to win the seat, but we’ve almost run out of postal votes, and my model suggests Labor will recover some ground.


Labor is still in front in Herbert by 449 votes – the LNP looks likely to gain a slim lead.


Labor came within 8 votes of falling behind in Hindmarsh today but have now regained a lead of 68 votes, and the model suggests that lead will grow.


Below the lines in Tasmania show potential for upset

The Senate data published on election night was very messy.

For a start, it didn’t make any effort to distinguish above-the-line and below-the-line votes, instead throwing all votes for each group into one pile.

Secondly, a lot of votes that will turn out to be formal were treated as informal and thrown into that pile, to be later resolved. Apparently this applied particularly to people who just voted ‘1’ above the line, but it’s not clear how consistent the AEC was with this rule – it’s possible to imagine that this happened more in busier booths and in states with bigger ballots.

In this post I want to focus on the first two problems: incorrect informality and no below-the-line data.

TLDR: There is a lot of below-the-line votes reporting for Lisa Singh and Richard Colbeck, who have the potential to be the first candidates to win seats off the back of below-the-line votes, potentially defeating their party’s preselected ticket. This data also suggests that Labor and the Greens may only get six seats, whereas the party totals suggest seven. Read the rest of this entry »


Close seats – Thursday morning update

Quite a bit of counting took place yesterday.

Yesterday we narrowed the list of seats to watch from 17 to 12, and today I think we can narrow that list to nine.

Seat summary

I’m ready to call Melbourne Ports for Labor and Grey and Dunkley for the Coalition. At the moment there are five seats which are extremely tight, and the rest look likely to break 72-68-5.

The seat categories are:

  • Coalition leading: Chisholm, Cowper
  • Labor leading: Cowan, Hindmarsh
  • Extremely tight: Capricornia, Flynn, Forde, Gilmore, Herbert
  • Called for Labor: Melbourne Ports
  • Called for Coalition: Grey, Dunkley

This brings the total to:

  • 70 – Coalition
  • 66 – Labor
  • 2 – Coalition leading
  • 2 – Labor leading
  • 5 – Others
  • 5 – Extremely tight

Of the five extremely tight seats, Labor currently leads in three and the Coalition leads in two. The Coalition needs to win four of these seats to form a majority government.

Cowper and Grey

In Cowper we now have preference counts from 24 out of 70 booths. The Nationals lead with 53.9%. My model suggests that lead will widen, but I’d like to wait for more results before calling Cowper.

We have about 60% of booths with preference counts in Grey, and Liberal MP Rowan Ramsey is leading with over 56% of the vote after preferences – so I’m happy to call Grey for the Liberal Party.

Melbourne Ports

I’m going to call this seat for Labor, with the first batch of postal votes being very bad for the Greens, as expected. If this trend continues, the Greens will fall thousands of votes behind Labor.

The other races

There is still frustratingly little data on how many absent and pre-poll votes are left to be counted, so we are still mostly working off postal vote data, and assuming the same numbers of absent and pre-poll votes in 2016 as in 2013.

Seat Absent Provisional Pre-poll Postal Current Labor lead Projected Labor lead
Capricornia 0 626 126 7493 732 -37
Chisholm 831 1314 335 4848 -1394 -2008
Cowan 0 1399 25 2116 701 784
Dunkley 0 1619 196 8279 -1387 -2777
Flynn 0 733 166 8131 1065 -1450
Forde 1052 1104 181 8622 -265 -657
Gilmore 0 1425 322 1823 -991 -856
Herbert 0 1136 122 6286 620 -362
Hindmarsh 0 1599 0 2951 151 541

(The first four columns of data in this table represent the number of outstanding declaration votes the AEC has announced – clearly these absent and pre-poll numbers are too small.)


The Labor lead has been cut by 260 votes in the last day, and now the LNP is projected to win, but still by a small margin.


The Liberal Party had the slimmest of leads yesterday morning, but now leads by around 1400 votes. The projection suggests this would grow. Not calling it now but likely will do so soon.


Labor’s lead dropped, but they are projected to remain steady. Still need to wait for more information.


The Liberal Party’s lead has expanded out to almost 1400 in Dunkley and is expected to blow out further, and I’m happy to call Dunkley.


We projected the Labor lead in Flynn would collapse and this is well under way. Labor’s lead has dropped from about 2000 votes to about 1000 votes. All the same, the projection expects Labor to lose another 2500 votes relative to the LNP, which is due to their very poor performance in postal votes – 35.8% in the votes counted so far.


The LNP has taken the lead in Forde and the projection is expected to grow a bit more.


The Liberal lead in Gilmore has increased from 353 votes to almost 1000 votes, and this looks set to stay steady.


Labor’s lead in Herbert has been cut back by about 300 votes, but the projection remains fairly steady, expecting an LNP win by less than 50 votes.


Labor’s lead has been cut back severely, but they are still expected to win.


Close seat update – here’s the model

We saw a small amount of counting yesterday, but we should be able to narrow the list of undecided seats.

When I last posted, I listed 17 seats as in some doubt. This included 13 conventional Labor-Coalition races, plus Grey and Cowper, where we were waiting on a two-candidate-preferred count, Melbourne Ports where Labor is at a small risk of coming third and losing the seat, and Batman which is a reasonably conventional Labor-Greens race.

Seat summary

I still haven’t called the following seats:

  • Coalition leading: Cowper, Grey, Chisholm, Dunkley
  • Labor leading: Melbourne Ports, Cowan, Hindmarsh
  • Coalition likely to take lead: Flynn
  • Extremely tight: Forde, Herbert, Gilmore, Capricornia

And I have called these

  • Called for Labor: Batman,
  • Called for Coalition: Dickson, La Trobe, Petrie, Robertson

So this brings the total to:

  • 68 – Coalition
  • 65 – Labor
  • 5 – Coalition likely to win
  • 3 – Labor likely to win
  • 5 – Others
  • 4 – Extremely tight

In this case, if the Coalition wins all the seats they are leading in, they need to win three out of four of Forde, Herbert, Gilmore and Capricornia to get a majority – they are currently leading in one.

Grey and Cowper

The most interesting counting undertaken yesterday started late in the afternoon, when we saw a large number of booths in these two seats count their preferences. In Grey the count is between the Liberal Party and the Nick Xenophon Team. In Cowper the count is between the Nationals and independent candidate Rob Oakeshott.

Preferences have been distributed in 14 out of 70 booths in Cowper and 35 out of 124 booths in Grey. These all happened quite late in the day, and presumably this count will conclude today.

In Grey, the Liberal MP Rowan Ramsay is ahead with 55.5% after preferences. The sample of booths is slightly favourable to him, with 45% of votes going to the Liberal Party in those booths where preferences have been distributed, and 41% in the rest of the seat. The Nick Xenophon Team candidate polled 1.7% better on primary votes in the booths yet to report, and Labor’s candidate polled almost 3% better. My projection suggests the Liberal two-candidate-preferred vote will drop to 53.9% with these votes, but I’ll wait for more votes before calling this result.

In Cowper, Nationals MP Luke Hartsuyker is ahead with 51.9% after preferences. Those booths reporting are 4% less good for the Nationals than the remaining booths. The current sample is very good for the Greens: 10.8% compared to 5.7% in the rest of the seat. This is primarily due to the inclusion of Bellingen.

I expect Hartsuyker’s lead to grow to 54% in the remaining counting, but again I’m happy to wait for tomorrow to call this seat.

Melbourne Ports

I haven’t built a model for this seat, although I may in coming days.

Currently there have been no special votes counted. Labor’s Michael Danby is almost 2% ahead of the Greens candidate, which is almost enough. Labor underperformed on special votes in 2013. While the Greens slightly outperformed on absent and pre-poll votes, they massively underperformed on postal votes, a much larger bundle of votes. I would like to wait and see some postal votes counted before calling this seat.

The other 14 races

In these seats most preferences have been counted and we are just waiting for the special votes to resolve a slim race.

At this point we are still waiting for hard figures on how many absent and pre-poll votes are to be counted. The following table shows how many votes are reported as waiting to be counted in each category.

Seat Absent Provisional Pre-poll Postal Current Labor lead Projected Labor lead
Batman 0 1553 192 7934 2330 2238
Capricornia 0 626 114 9217 994 -493
Chisholm 0 1312 178 10897 -76 -1197
Cowan 0 1399 0 5724 959 895
Dickson 0 706 45 10885 -1258 -2666
Dunkley 0 1619 196 13146 -421 -2208
Flynn 0 650 0 10238 2058 -973
Forde 0 1104 54 11061 104 -316
Gilmore 0 1425 322 5859 -353 -612
Herbert 0 1136 122 6540 913 -10
Hindmarsh 0 0 0 4293 645 1284
La Trobe 0 1481 239 12909 -1350 -3013
Petrie 0 701 88 10761 -1602 -2710
Robertson 0 1622 264 6317 -1183 -2085

We’ve seen the first batches of postal votes counted in Batman, Chisholm, Cowan, Flynn, Herbert, Hindmarsh and Petrie.


The Greens did about as well on postal votes in 2013 as they did on other votes, but the same is no longer true. Labor won over 60% of the first batch of postal votes, compared to just over 50% of the ordinary votes. Under my model, this indicates that Labor is on track to maintain their current large lead, so I am calling Batman for Labor.


Labor polled 5% worse on postal votes in 2013 than on ordinary votes – on the first batch in 2016 that gap is 7%. At the moment my model suggests that the thin Coalition lead will blow out, but I won’t call it yet.


Labor performed 5% worse in 2013 on postal votes and the first batch matches that pattern. Labor looks on track for a win, but not comfortable enough to call it.


Labor did extremely poorly on postal votes in Flynn in 2013, which is why a seemingly unassailable lead wasn’t enough to call it for Labor. Indeed Labor managed less than 35% after preferences in the first postal vote batch of 800 votes. My model is currently projecting a 3000-vote turnaround giving the seat to the Coalition. We’ll need to see more votes to be confident of this.


Labor also did poorly in the first batch of postal votes here, and my model has this 913-vote lead dropping to a 10-vote lead for the Coalition.


Labor did lose the first batch of postals in Hindmarsh, but by the same margin as in 2013, and my model predicts an increase in the Labor lead.

Calling seats

In addition to Batman, I’m calling Coalition wins in Dickson, La Trobe, Petrie, Robertson. So I won’t be tracking these seats any further.


Pauline Hanson and the old Senate system

There’s been a lot of arguments coming out suggesting that Pauline Hanson and the rest of her team are only contenders for seats thanks to the Senate voting reforms, and the double dissolution.

Clearly a double dissolution made it easier for all small parties, including One Nation, and you can argue that a double dissolution was made possible thanks to Senate voting reform (although it wasn’t necessary to do both).

There are a number of points to challenge in this argument and I’m planning to cover them in an op-ed I’m working on, but I just wanted to lay out the facts about why I believe Pauline Hanson would have been elected regardless of the voting system, and would have won under a half-Senate election.

There’s a theory that preferences have hurt Pauline Hanson – parties organised to ensure preferences didn’t flow to her, locking her out of Parliament. This was definitely true in 1998 and 2001, in both houses. The major parties, the Democrats and Greens all preferenced against her, and it helped prevent her from winning in Blair in 1998, and meant that One Nation only won a single senate seat in those two elections instead of winning a bunch.

However that is no longer the case. For whatever reason there are now a bunch of parties which poll a significant proportion of the vote who have shown a willingness to preference One Nation very highly.

In 2013, One Nation was a party that received a lot of preferences far above other rivals for seats, including the Greens, the major parties and even a bunch of other larger minor parties.

Let’s play a little hypothetical here. Let’s take the primary votes for the Senate in Queensland as they currently stand, but imagine that it’s a half-Senate election, so the quota is higher. The leading tickets are:

  • 2.36 quotas – Liberal National Party
  • 1.90 – Labor
  • 0.64 – One Nation
  • 0.53 – Greens
  • 0.18 – Liberal Democrats
  • 0.14 – Family First

I did try to do a full preference distribution using Antony Green’s 2013 calculator, but that’s not possible due to the presence of parties belonging to Glenn Lazarus, Nick Xenophon, Derryn Hinch and Jacqui Lambie that don’t have a 2013 parallel.

So instead let me quickly run through all the parties that give their preference to One Nation before any of the other parties contending for seats.

  • Online Direct Democracy (previously Senator Online)
  • Katter’s Australian Party
  • Democratic Labour Party
  • Shooters, Fishers and Farmers

These three add up to 0.28 quotas, bringing One Nation up to 0.92 quotas.

While this is going on, Family First has also gained votes from Sustainable Australia, Animal Justice and Australian Christians who flow next to One Nation after Family First, and then presumably knock out the LDP who flow to Family First.

This puts One Nation on 0.92 quotas, and Family First on 0.45 quotas – all of which are votes that flow to One Nation next if Family First is knocked out.

There’s a lot of preferences we can’t predict – who knows where Lazarus, Xenophon, Lambie preferences go. But at this point One Nation is close to a quota, and Family First is still behind the Greens primary vote. So let’s assume Family First is knocked out.

At this point, all of these parties’ preferences flow to One Nation, electing their candidate:

  • Animal Justice Party
  • Sustainable Australia
  • Australian Christians
  • Liberal Democrats
  • Family First

Easily electing One Nation.

Thus you can see how One Nation, polling the kind of vote they did on Saturday, would have likely won a Senate seat in Queensland regardless of system or number of senators elected.


Looking forward in the close seats

I have a list of sixteen seats which I think are close enough to pay further attention over the next few days as more votes flow in.

In this article I want to explain what data we will be receiving which will help us project forward in each electorate and hopefully narrow the number of undecided seats.

These seats are Batman, Capricornia, Chisholm, Cowan, Dickson, Dunkley, Flynn, Forde, Gilmore, Grey, Herbert, Hindmarsh, La Trobe, Melbourne Ports, Petrie and Robertson. In thirteen of these seats the race is a conventional Labor-Coalition race, and it’s just a matter of tracking how close the vote is as extra votes report. In Batman, the race is between Labor and the Greens, but the same principle applies.

In Melbourne Ports, the question is whether the Greens can get ahead of Labor, which could see Greens or (more likely) Liberal win the seat.

In Grey, we are waiting for a two-candidate-preferred count between Liberal and Nick Xenophon Team to determine the result – that count may well make it clear who will win.

The remaining 134 seats split 65 Coalition, 64 Labor, 2 independent, and one each for the Greens, Katter’s Australian Party and the Nick Xenophon Team.

Firstly I want to look at the preference count in Grey. Preferences began to be distributed this afternoon, and these figures have been reported for four out of 124 booths. So far NXT candidate Andrea Proudfoot leads with just over 56% of the vote, but it’s a small sample.

These booths are mostly in the Port Augusta area, which was relatively strong for NXT. When you take the preference flows in these four booths and apply them to the rest of the seat, the Liberal Party comes out on top with 51.2%.

Of course different areas can flow differently, and it’s hard to extrapolate – but NXT probably needs to get stronger preference flows in the remaining booths to win.

In the other fifteen seats, we will be concentrating on the incoming declaration votes. These votes fall into four categories:

  • Postal votes
  • Provisional votes
  • Absent votes
  • Declaration pre-poll votes (effectively absent pre-poll)

Counting has paused for today and tomorrow to allow the AEC to securely swap around those absent and pre-poll votes that need to move between electorates. A majority of postal votes have already been returned to each electorate, but there is still time for more to arrive.

It’s possible to track the number of votes in each category which are yet to be counted, which is what we’ll be doing over the next few days. We can also extrapolate how each bundle of votes is likely to break based on how postal, absent and pre-poll votes flowed in 2013, which I’ll be doing over coming days.

For now I just leave you with this summary of the number of declaration votes which have been returned to each seat.

Seat Absent Provisional Pre-poll Postal Postal yet to return
Batman 0 0 0 8088 4104
Capricornia 0 0 0 7968 3680
Chisholm 0 0 0 11767 4411
Cowan 0 0 0 5276 2116
Dickson 0 0 0 9802 3714
Dunkley 0 1617 195 12318 4032
Flynn 0 0 0 9518 5422
Forde 0 0 0 8966 4610
Gilmore 0 0 0 5373 2478
Herbert 0 0 0 6126 5194
Hindmarsh 0 0 0 7843 3684
La Trobe 0 1479 0 11711 4460
Petrie 0 0 0 10574 4095
Robertson 0 0 0 5855 2513
Melbourne Ports 0 25 0 11762 6436

Counting of these votes will commence on Tuesday.


The day after – quick update

Firstly, apologies for the technical problems last night. I hope that readers were able to follow my commentary as part of the Guardian liveblog. In future if I’m participating in a liveblog somewhere else I’m not going to try and keep my own liveblog going here – I’ll just refer people to that better service. But I will look at technical improvements.

In addition to the liveblog, you can watch a video of my assessment of the result last night (below), and I also participated in a late-night podcast summing up the results.

I’ll be doing some writing for the Guardian this afternoon about various elements of the campaign, so keep an eye on my Guardian profile to catch those.

I will be returning to this blog over the next few days to do analysis of the key House and Senate races, looking at the votes still expected to flow in.

For now, this is a quick summary of the votes that are yet to come in:

  • It appears that all ordinary votes have been counted in lower house races. This covers in-electorate pre-poll votes, in-electorate election-day votes, special hospital votes and some other small batches of votes.
  • In the lower house, we are still waiting for declaration votes, including: postal votes, absent votes and pre-poll out-of-electorate votes. Apparently these will begin to be counted on Tuesday, with today and Monday used to ensure that declaration votes get to where they need to be.
  • In the Senate, in addition to declaration votes, we are also waiting for in-electorate pre-poll votes to be counted. These should be counted on Monday.
  • It’s also worth noting that there are likely to be a lot of formal votes currently sitting in the ‘informal’ pile for the Senate. Apparently booth workers were instructed to put just ‘1’ above the line votes in that pile, and there is a history of below-the-line votes being thrown in the informal pile if formality is not immediately clear.

That’s it for now, but there’s a lot more to say.


Federal election live

In addition to updating this liveblog here, you can also follow my commentary on Twitter or at the Guardian liveblog.

9:26pm – Sorry guys, I’m going to cut this short. Follow the Guardian liveblog for my coverage, I haven’t got the space to do both. I may return tomorrow with some kind of summary, although I’ll also be doing some writing for the Guardian.

7:45pm – The Greens are doing very well in Batman and Wills, which they could be winning. The Greens have also gained a swing of almost 10% in Richmond, but will still come third. The Greens have not gained much if any swing in Grayndler and Sydney.

7:40pm – Sorry guys, been busy sending snippets to the Guardian and making calls on seats for their model. It looks pretty clear that Labor has gained three Tasmanian seats plus Macarthur, looking good in Macquarie and well and truly holding the notional Liberal seats of Dobell, Paterson and Barton.

Things aren’t looking great for Tony Windsor in New England.

6pm: Polls have just closed on the east coast. We should start to see data in about half an hour.


Election eve open thread

Polls open on the east coast in fifteen hours. I thought I might open up a quick post to allow commenters to post general comments about the election.

If you have predictions feel free to post them below.

I have had a number of articles up at the Guardian in the last few days.

On Monday I wrote about the state of the polls.

Yesterday I wrote about the massive surge in early voting.

And my latest article about the ‘sophomore surge’ phenomenon has just gone up.

I will be posting another quick post tomorrow morning for election-day comments, and tomorrow evening I will be posting on a liveblog here, as well as contributing the Guardian’s liveblogging. I’m hopeful that the improved hosting systems for this blog will mean the website holds up better than it did in 2013, but if you can’t access it please just follow the Guardian liveblog to read up on my thoughts.


Seat of the day #54: Petrie

Petrie1-2PPPetrie is a marginal LNP seat in the northern suburbs of Brisbane, including Bracken Ridge and parts of Aspley.

The LNP’s Luke Howarth won the seat in 2013. He holds Petrie by a 0.8% margin, which makes it the most marginal Coalition seat in Australia.

Read more