2020 Queensland Election Updates

In this post I will provide running updates on results as they report.

I am running this post as an adjunct to the ABC’s election results site. Using the computer that drives the site, I can change settings in the system to deal with the variation in the data coming from the Electoral Commission. Refer to the ABC site for overall percentage votes, seats won and seats remaining in doubt.

Sunday 15 November

Apologies for the lack of updates over the weekend. The re-count results were announced after I had left the ABC on Friday, and for technical reasons the site can’t be updated until I’m back in the office on Monday. Hopefully by then the final figures will have been published in all seats. That means for the moment that both Bundaberg are shown as in doubt, but they have been decided based on the yet to be updated re-counts.

Friday 13 November

8:45pm – the result of the re-counts is in, with Labor reported as winning Bundaberg by 9 votes and Nicklin by 85. The ABC site won’t be updated with these numbers until Monday.

4:00pm – I’ve manually updated about three-quarters of districts with final figures are after preferences. The rest are yet to be published and I’ll deal with them next week. I think I’ve earned the weekend off after three elections in a month.

Thursday 12 November

The ECQ has started to publish preference counts for decided seats. The ABC election site can’t publish them until I flick a few control switches on Friday.

Wednesday 11 November

Final update for the election. The LNP won Currumbin by 310 votes, Labor won Nicklin by 79 votes, and also crawled back in front to win Bundaberg by 11 votes. The distribution of preferences have not yet been published by the ECQ so the ABC website does not currently show the final figures.

Friday 6 November

6:50pm – again remembering these numbers are preferences estimates (see below), in Bundaberg the LNP’s David Batt leads Labor’s Tom Smith by 95 votes with 88.2% of the vote counted.
6:50pm – again remembering these numbers are preferences estimates (see below), in Bundaberg the LNP’s David Batt leads Labor’s Tom Smith by 95 votes with 88.2% of the vote counted.

6:05pm – based on the preference estimates set out below, the LNP lead in Currumbin is 142 votes with 85.7% counted. In Nicklin, the estimated Labor lead is 259 votes with 87.0% counted.

2:50pm – again remembering these numbers are preferences estimates (see below), in Bundaberg the LNP’s David Batt leads Labor’s Tom Smith by just 56 votes with 87.4% of the vote counted.

2:30pm – I haven’t seen any updates yet for the three in doubt seats yet today, but though I should re-stress that the two-candidate preferred counts being published on the ABC website are estimates as no official counts are available.

The estimated preferences I am using are derived from the final ECQ datafile that contained preference data. The estimates are –

  • Bundaberg – Legalise Cannabis (48% Labor, 52% LNP), Greens (66% Labor, 34% LNP), UAP (30% Labor, 70% LNP) and One Nation (30% Labor, 70% LNP)
  • Currumbin – Takacs-Thorne (IND) (43% Labor, 57% LNP), Stuckey (IND) (65% Labor, 35% LNP), Greens (81% Labor, 19% LNP), One Nation (31% Labor, 69% LNP), UAP (25% Labor, 75% LNP), Logan (IND) (46% Labor, 54% LNP_)
  • Nicklin – Greens (81% Labor, 19% LNP), One Nation (32% Labor, 68% LNP), Independent (35% Labor, 65% LNP), Informed Medical Options (54% Labor, 46% LNP)

Thursday 5 November

12:30pm – official word is the ECQ has decided to stop doing preference counts to enable staff to concentrate on finalising the check count to ensure everything is in place for the official preference count to be done.

11:30am – As I said in my last update yesterday, the ECQ has stopped providing preference count totals for Bundaberg, Currumbin and Nicklin. The ABC site is now publishing estimated 2CP counts based on the preference count details published until yesterday.

Wednesday 4 November

8:20pm – My reported figures below may be completely wrong. I just noticed that the figures are not actual preference counts. The ECQ has removed the 2CP totals from the feeds, so the margins below are all based on preference flow data taken from the feed on Monday. I’m not sure why the 2CP counts have been removed from the feed.

8:00pm – with today’s other distraction finally out of the way, time for an update on the close seats in Queensland.

  • Bundaberg – with 84.8% counted, LNP sitting member David Batt has a 48 vote lead over Labor’s Tom Smith with 84.8% counted.
  • Currumbin – with 83.0% counted, the LNP’s Laura Gerber leads Labor’s Kaylee Campradt by 43 votes, the count having narrowed substantially today.
  • Nicklin – with 81.9% counted, Labor’s Robert Skelton leads the LNP’s Marty Hunt by 112 votes.

It has been a long and tiring day so haven’t got much else to say other than counting continues tomorrow.

7:00am – updates here will be few and far between today as I am involved in US election coverage. However, the ABC’s election website will continue to update with results over here. At the end of last night, the state of play in the three seats remaining in doubt are –

  • Bundaberg – with 83.9% counted, LNP sitting member David Batt has a 4 vote lead over Labor’s Tom Smith, ahead 14,240 to 14,236. My estimate is that including about 800 first preferences not currently counted for preferences increases the LNP lead to 50 votes.
  • Currumbin – with 82.2% of the vote counted, the LNP’s Laura Gerber leads Labor’s Kaylee Campradt by 302 votes, 13,626 (50.6%) to 13,324 (49.4%). There are some 1700 primary votes not yet included in the preference count. Once these are added the LNP lead should narrow to around 200 votes.
  • Nicklin – with 73.6% of the primary vote counted, Labor’s Robert Skelton leads the LNP’s Marty Hunt by 67 votes, 12,681 to 12,614. Once some votes not currently counted for preferences are included, I estimate the Skelton leads by 168 votes.

Just a reminder that there are only five seats where the ECQ is continuing to conduct indicative preference counts – Bundaberg, Burleigh, Currumbin, Hervey and Nicklin. In all other electorates, the ABC is calculating two-candidate preferred counts based on preference throw data from election night. Actual preference counts in seats won’t be available until the formal distribution of preferences takes place starting after 13 November, the final date for the return of postal votes.

After reviewing results last night, I decided it was safe to give Burleigh, Coomera and Glass House to the LNP.

Tuesday 3 November

8:00pm – I thought I should add a little something to the debate on who was responsible for the defeat of Jackie Trad, passing over for the moment whether she played an important part in her own defeat. Was it the switch of Labor voters to the Greens or the change in the Liberal HTV that defeated Trad?

After preferences, the swing from Labor to the Greens in South Brisbane is a little under 10%. That swing consists of two components. The first is the switch in first preference support between Labor and the Greens. The second is the LNP’s change in how-to-vote preference recommendation.

By my estimates, the first preference switch from Labor to the Greens explains roughly a quarter of the swing. The change in LNP preference recommendation accounts for the other three-quarters. Labor may have scrambled to a narrow victory last Saturday had the LNP had made the same anti-Green preference recommendation as in 2017, and had the flow of preferences to Labor again been in the lows 60%s.

However, blaming the LNP for defeat overlooks that it was Labor’s loss of first preference support to the Greens that left the door for defeat ajar. The LNP’s decision on preferences simply pushed the door wide open.

Since 2013 the Federal Liberal Party has chosen to recommend preferences to Labor ahead of the Greens on grounds of ideology. The LNP decision to recommend preferences for the Greens made little ideological sense for a conservative party that normally demonises the Greens. But it made perfect strategic sense in maximising damage the LNP’s main political opponent in the race for government – Labor.

7:45pm – another update in Nicklin increases Labor’s lead from 56 to 67 votes.

5:40pm – Glass House remains in doubt on the ABC website, but only because the count has not been updated in some time and sits at 67.7%.

5:30pm – two late updates. In Currumbin, a new preference count update has put the LNP a smidgen further ahead. Laura Gerber now has a 310 vote lead, 50.6% up from 50.5% last night. Further postal vote counting in Cooper has also widened the lead of the second placed Liberal candidate over the third placed Green, ensuring that the Green candidate will be excluded and Labor’s Jonty Bush will win. Yet to be counted Absent votes will help the Greens, but there are certain to be fewer of them in 2020 in line with the lower proportion of on-the-day voting. Earlier today the LNP’s 2-party vote in Burleigh slipped to to 50.9%, but with 78% counted, Labor pulling ahead from here seems unlikely.

3:50pm – thwe second update from a key seat today. In Bundaberg, with 83.7% of the vote counted, the LNP’s David batt now leads by 4 votes.

10:20am – Updates might be a bit sparse today as I am involved in US election rehearsals. One update of note is that Nicklin now has Labor ahead by 56 votes. The 2CP count here is now ahead of the primary vote count as the preference count has been done on the initial count but those first preferences have yet to be check counted, and until the check count is done, the number don’t end up in the ABC’s first preference tally. Choosing to use the check count totals has worked fine for polling places, but is proving less accurate for accumulating count centres such as pre-polls and postals.

Monday 2 November

8:00pm – see comments below. Main developments today were Bundaberg getting very close and the Liberal lead over the Greens widening in the race for second place in McConnel. I think the two-candidate count for Currumbin is narrower than is being reported, and Labor Party scrutineers think they are in the lead. But it is hard to say anything definite without actually seeing a further preference count. Coomera, Glass House and Burleigh currently sit in the list of ABC doubtful seats, as does Burleigh, but I’m expecting all three to be won by the LNP. Further counting will see those three seats automatically drop off the in doubt list.

5:05pm – the addition of the preference count for postal votes in Bundaberg has narrowed Labor’s lead to just 24 votes. Labor did gain votes on absent votes in 2017, but there are many more postals to come. The swing to Labor on postals is just 0.6%, much weaker than with on the day votes, producing the narrowing of margin.

4:55pm – more postal votes have been added to the initial count in McConnel. These won’t appear on the ABC site until the check count is completed. This increases the lead of the second placed Liberal over third placed Green from 2.8% to 3.2%. The lead last night was 31.0% to 28.2% last night, now out to 31.3% to 27.9%. The lead is 878 votes, and with votes for lower placed candidates totaling 1,329, there would need to be a dramatic turnaround on absent votes to narrow the gap.

2:45pm – in Currumbin, the LNP’s Laura Gerber is on 50.5% on the preference count, but this trails the updated primary count. My projected estimate is that the LNP is on 50.2%, a lead of 106 votes.

2:10pm – after altering the settings for Bundaberg to accept preference counts, the Labor 2PP percentage increased from 50.2% to 50.6%. This is because my estimated preference count applied to the postal vote that has not yet been counted out for preferences. The LNP polled well on first preferences with postal votes so these votes will narrow the Labor lead.

2:00pm – the ECQ has confirmed that they will continue to conduct indicative preference counts in five electorates, Bundaberg, Burleigh, Currumbin, Hervey Bay and Nicklin. The ABC website has been configured to accept preference counts for these five seats, but all other results on the ABC website are now based on estimated preferences.

1:45pm – while preparing my preference tables last night, I noticed what I thought was a transposition error in the indicative preference count for Lytton. I reported it to the ECQ who’ve confirmed the transposition error and their website has been updated. This error would have been spotted when the official distribution of preferences was done and it is a reminder that the indicative preference count is done for information purposes only, not to determine the winning candidate.

12:45pm – the election database has been re-built and all remotely close electorates are now using preference estimates derived from the preference count data released over the weekend. The safe seats listed in my noon update are using preference estimates. The ECQ is currently deciding whether to continue with preference count updates in a small number of electorates where doubt remains over the result.

Last night I also published a blog post looking at party preferences flows as revealed by the indicative preference counts released on the weekend. You can find the post here.

Noon – I’m currently re-jigging the ABC election system to deal with changes in the way the ECQ is publishing its data. I’m also converting a number of seats to Labor v LNP contests where on election night the contest was Labor v ONP or LNP v ONP. This will change the swing in 15 electorates once the re-build is finished. The electorates are Bundamba, Condamine, Cook, Gympie, Ipswich, Ipswich West, Jordan, Lockyer, Logan, Maryborough, Nanango, Rockhampton, Scenic Rim, Southern Downs and Warrego. This conversion had also been done for Gregory, Keppel and Waterford before election day.

You will also note differences in the tally from the two totals published on the ECQ’s website. The ECQ are publishing two sets of totals, a preliminary count from Saturday night, and a check count conducted after Saturday. The ABC is merging these two totals polling place by polling place, as well as for count centres. So you may find variations in the count compared to the ECQ.

1 November

At the end of counting on Sunday, Labor has a definite 50 seats and four seats have changed party, Labor gaining Caloundra, Hervey Bay and Pumicestone from the LNP while losing South Brisbane to the Greens. The LNP have a definite 31 seats with five seats it holds remaining in doubt. The LNP lead in Burleigh, Coomera and Currumbin and is probably favoured in all three. The LNP trail in Bundaberg and Nicklin, the two seats over which the greatest doubt remains. The balance of the Assembly is three KAP, two Greens, one One Nation and one Independent.

7:30pm – In Cook I’ve turned off receipt of preference counts from the ECQ as the result is clearly not a Labor v KAP contest. On the counts to date, the ALP:KAP split of preferences for each candidate is Batzke (IND) 39:61, LNP 15:85, NQF 38:62, ONP 20:80, GRN 78:20 and UAP 41:59. I’ve used these to create an ALP:LNP preference formula, except that I have split KAP preferences 75:25, that is slightly weaker to the LNP than LNP to KAP. Using that formula, Labor’s lead in Cook increases and Cynthia Lui holds the seat.

4:15 pm – As in Hervey Bay, in Clayfield I’m using a preference estimate based on the election day preference flows because the Clayfield Early Voting Centre and the Postal votes have only been counted for first preferences. Both these count centres heavily favour Liberal Tim Nicholls. The preferences estimate based on more progressed primaries is a better picture of the result than the lagging two-candidate count. I am currently doing the same in Bundaberg where there is no 2CP for postal votes.

3:55 – I’ve had a report of a correction in Barron River that will cut Labor’s lead by 440 votes. That won’t change the result.

3:50 pm – The ABC computer has given Hervey Bay to Labor and I need to explain why. Neither of the two early voting centres have reported 2CP counts, but both have reported first preferences. Labor currently leads on first prefernces, and based on Labor receiving the majority of preferences on election day, then projecting that flow on to the current first preference tally delivers a comfortable Labor victory.

3:20 pm – I’m trying to resolve a discrepancy with the count for Currumbin. The ECQ is providing two sets of count at this election, a preliminary count from election night, and a check count from after the election. Both are by polling place, and we merge the two to create a best possible total. However, in Currumbin we are getting a discrepancy between the two counts. It doesn’t show up on the ECQ site as their totals are independent of each other, but when you merge the two you get a 99% turnout. I’ve pointed it out to the ECQ so hopefully it will be fixed. On a 99% turnout the LNP hold Currumbin, but there is clearly an error.

2:10pm I don’t have precise details so this is an observation, but it appears that while Labor’s vote was lower with postal and pre-polls than on the day, the swing to Labor was larger. That means the concerns I had early in the count about seeing an inflated swing to Labor have not proved to be wrong. In fact, Labor’s result firmed up with the arrival of pre-polls and postal votes. At this stage Labor is a strong chance of picking up Hervey Bay and is still ahead in Bundaberg and Nicklin. The ABC election computer has moved Cook back into doubtful, but there are some strong Labor booth yet to have their two-party preferred to be included. To me Labor has a certain 49 seats and likely 50.

1:15pm The ECQ has now discontinued and stopped releasing the Labor v Green preference count in McConnel as it is clearly a Labor versus LNP contest.

12:50pm A few updates on support for minor parties.

In the 60 seats that One Nation contested at both the 2017 and 2020 elections, One Nation’s vote was 8.4%, down 12.4 on the 20.8% it polled in the same seats in 2017. Statewide the party polled 6.9%, half of the 13.7% it polled in 2017, despite standing an extra 29 candidates

Green support was very geographically concentrated in central Brisbane. The party polled 16.5% (+0.1) in the Brisbane City Council area, 12.2% (-0.5) in Greater Brisbane, 11.3% (-1.0) in South-East Queensland overall, and 4.8% (-0.8) in the rest of the state. The over all figure with 64.1% counted in 9.2% (-0.8). This will probably tick up a point or two with remaining counting.

Again with 64.1% counted, Clive Palmer’s United Australia Party stood 55 candidates and has polled 11,853 votes (0.57%). This compare to 12,479 votes (0.60%) for Informed Medical Options Party (31 candidates) and 18,352 (0.89%) for Legalise Cannabis Queensland (23 candidates). Whose was the better value for money campaign?

12:15 – In McConnell, the ECQ continues to provide a Labor versus Greens count, but at this stage it appears more likely that the Greens will finish third. I expect the ECQ to make a decision on further preference counting in the seat at the end of today’s counting. In the ABC’s computer, I am now declining the ECQ’s count and estimating a Labor-LNP finish. There is no doubt Labor wins if the Greens finish third. It seems unlikely with further postal voting to come that the Greens can reach second place.

Midday: My first update. The ECQ has decided to stop conducting indicative preference counts for 11 seats, Condamine, Gympie, Ipswick, Ipswich West, Jordan, Lockyer, Logan, Maryborough, Nanango, Scenic Rim and Southern Downs. Preference counts for these seats are no longer being published. In each of these seats the ECQ had selected to conduct LNP versus One Nation counts, It is a sign of One Nation’s bad result that all of the seats have finished as LNP versus Labor contests. Noe are in doubt so there is no need to re-do and continue preference counting. The preference counts on the ABC’s website are all estimates.

1 am – It’s very late at night, or rather early in the morning, after the Queensland election coverage.

I’ve re-jigged the ABC Election computer to turn off predictions based on polling place comparison. It is now displaying raw numbers and percentages.

When I wake up I will update the site and track updates through the day. But first I’ll flag a couple of changes I’ve made to a couple of seats to overcome postal votes being reported for first preferences but not for two-candidate preferred.

In the following electorates, Bundaberg, Chatsworth, Clayfield and Keppel, I have used the preference flow data from election day to provide a better two-candidate count that now projects preferences results for the postal vote counts.

There are also a few seats including Ipswich, Ipswich West, Jordan, Logan, Nanango and Scenic Rim where the ECQ has turned off its preference count and the ABC system has reverted to using estimated flows of preferences.

I’ll do more posting on the results tomorrow after some sleep.

9 thoughts on “2020 Queensland Election Updates”

  1. Noticed you said you reckon the Greens vote will tick up “a point or two” on counts still to come, can you elaborate on why that is? Thanks for the coverage, excellent as always.

    COMMENT: The Greens always do well on absent votes.

  2. Discrepancy in Currumbin between the preliminary v. official count at Tugan Early Voting Centre. Considerable changes made between the primary contestants, together with substantial increase in informal votes!

    COMMENT: The ABC merges the two counts, replacing the preliminary count with the official count for a count centre where available. The ECQ had some errors in the entry of their check counta earlier today which I raised with the ECQ and they have been corrected.

  3. Hi Antony – i notice on the ABC the difference between Labor and LNP in Currumbin is only 139 votes with 72% counted. Is there a situation where the close count would trigger a recount or only if the challenger requested it? Thanks!

    COMMENT: As I understand it, Queensland has no minimum number. It is a decision made on request.

    1. Some more postals, polling day absent votes, pre-poll absent votes, and a smaller group of sundry declaration votes. The number varies from seat to seat. I don;t know what the numbers are.

  4. Hi Antony – in Cooper, given the Greens first preference of 31%, the LNP first preference of 33% and the LNP referencing Greens ahead of Labor, how is this seat not ‘in doubt’ awaiting distribution of preferences between Greens and Labor?

    I would have thought the Green would end up with 55-60% after preferences. Is it because Greens candidate will be eliminated first before the LNP preferences are distributed to them?

    COMMENT: Yes, it is because the Greens are in third place. They can’t benefit from Liberal preferences unless they first pass the Liberal candidate. That’s looking tough to achieve.

  5. Hi Antony,
    I’ve been following Qld Elections since 1983 and this one has been one of the hardest to follow as it was often not clear if the 2PP votes were actual or projected. Also, on election night it took forever to check the results as the fake 2PP (projection) appeared first and then it was necessary to to click on a link to obtain the primary results and then sometimes the actual? 2PP.

    Just this morning, figures came up for Bundaberg, reducing the ALP lead from 277 to 181 votes. From reading your article above, it now appears that this was only a projection? At the time I saw it though, I interpreted it to mean the LNP would likely lose the seat as their gain on this batch of postals was not enough. If that was a projection, it should have clearly been marked “PROJECTION”

    It appears that some seats show projections and others don’t and that the projections are switched on and of at various stages. If they are not clearly marked it is impossible to keep track of and therefore creates considerable confusion in interpreting the results. Perhaps your IT people find a way of flagging this better?

    1. You make some good points, and some things could be labelled better. One of the basic things was the the “Projection” setting for election night had not been turned off. That has been done now and the site now says current count. I’m also getting a code change to turn the polling place count off because it is not updating because the ECQ has stopped counting preferences.

      But even then that may not be entirely accurate and there is little I can do about it. There are only 5 electorates doing progressive preference counts. Every other electorate is now a preference estimate based on Saturday’s preference flows. The alternative is to leave the preference counts as they were in election night for the next two weeks until the count is finalised.

      As for Bundaberg, as I tried to explain in this blog, I was using a preference estimate on Monday morning as there were 3,000 primary votes in the count favouring the LNP that had not been counted out for preferences. An estimate of preferences was a more accurate reflection of the count than the actual count. Once the ECQ confirmed they would continue counting preferences for Bundaberg, I changed the flag to again accept preferences counts.

    2. There may have also been some manual setting lost when converting the database to use preference flow data derived from election night counts. I initially switched Bundaberg and the other four preference count seats to estimates before the ECQ confirmed they would start counting preferences again yesterday afternoon. There are also problems in resolving differences between the two primary vote counts being released by the ECQ, and the actual 2CP’s being released. The entire download frontend had to be re-written because of the ECQ’s count changes.

      The ABC’s computer system is now a decade old and was designed to support a television coverage, not as a website publishing tool. As more than 80% of traffic to the website is from mobile devices, the website has also be designed around mobile usage, cutting back more detailed elements. The new system was run in parallel for the Queensland election and will become the sole election analysis system at the WA election. It has been designed with more thought for web publishing and hopefully will clarify some of the issues over which you have complained.

Leave a Reply