SEAT ESTIMATE ALP 94 L-NP 43 IND 10 GRN 1 CA 1 KAP 1
SEATS APPARENTLY CHANGING (defections disregarded)
L-NP to LABOR: Banks, Bass, Bonner, Braddon, Deakin, Dickson, Hughes, Leichhardt, Moore, Sturt, Forde, Petrie, Menzies (see below)
GREENS to LABOR: Brisbane, Griffith, Melbourne
L-NP to IND: Calare
IND to L-NP: Goldstein
SEATS IN SIGNIFICANT DOUBT - seats shown in bold have now been discusssed in detail either here or on one of the linked pages:
ALP Occupied vs IND: Calwell (likely Labor? ... come back in a few weeks)
L-NP Occupied vs IND: Bradfield (IND ahead, extremely close, recount expected)
Seats I am not currently considering in doubt that were in the lists above (minor doubt re those marked *)
Flinders, Monash, Forrest, Grey, Fisher, Longman: expected Liberal retains
Wills, Fremantle, Bendigo, Bullwinkel, Bean: expected Labor retains
Goldstein: expected Liberal gain
Melbourne: ALP gain
Ryan: expected Green retain
Kooyong: expected IND retain
-----------------------------------------------------------------------------------------------------------------
Because of the complexity of the House of Reps postcount and my limited time to cover it, for starters I am putting up a hub page for links to postcount threads with a summary of the state of play and of seats I am keeping my eye on.
There are some seats where the AEC chose the wrong two-candidate pair on the night and is now already in the process of realigning the count to the correct two-candidate pair. During this process the 2CP figure on the AEC and probably also ABC site will jump about considerably - wait for it to settle down or for a projection to be posted here or elsewhere. Examples that might be competitive include Bean. Fremantle, Bendigo and Melbourne. Franklin is among several others being realigned but Julie Collins will win Franklin easily.
There are many seats where it is clear who has won the two-party preferred count but an independent is in third and might in theory make the top two, then might in theory win. Probably most of these will fall over - the independent may fall back on primaries, may fail to get into second after preferences, or may make the final two but lose anyway.
A seat tally estimate will be posted when I have had time to assess some of these seats in more detail.
List of links to non-classic postcount pages:
Calwell, Fisher, Flinders, Forrest, Grey
For Senate counts see my Senate count page. A Tasmanian Senate thread has been posted separately.
Classic major-party seats will be covered on this page. The following will be unrolled at the bottom of this page as time permits (others may be added). It's a sign of the times that these are a minority of the sites I am watching.
Bendigo (ALP 11.2%)
Monday: A bizarre outcome in an election that Labor has won very easily is that a seat well into its safe zone is at risk of flipping to the Nationals. The AEC pre-set this seat as ALP vs Liberal based on its past form and the count is now being reset to ALP vs Nationals.
The booths that have been reset so far are Bendigo and Junortoun prepolls, four small special hospital booths and a mobile booth, and part of the postal count. All these booths are strong for the Coalition parties compared to the overall count so the Nationals' current 53-47 lead is unreliable and will decrease. The preference flow to Labor in the substantial booths is varying wildly from 31% to 49% (and even more so if including the smaller booths).
In this case the most compelling relationship is between the preference flow to Labor and how much of the available preference pool is Liberal. In booths where a lot of the preferences come from the Liberals, Labor has a worse preference share.
Based off the booths so far I project the preference share for Labor to be 45.5% in the current live count, though I expect this to fall and should caution that these booths are unreliable in various different ways.
In the current live primaries Lisa Chesters (ALP) has 33.89, Andrew Lethlean (Nats) 30.57 Greens 11.15 Liberals 10.08 ON 4.66 LCP 3.21 FF 2.85 VS 1.7 a minor IND 0.98 Libertarian 0.91. The total preference pool is 35.54% at the moment and on my projection above the Nationals gain at 0.09 votes per preference. Therefore I project that the Nationals are currently gaining 3.20% of a current 3.23% lead, which puts Labor only precariously in front by 0.03%. There could be 5000 postals to go but in 2022 postals were overall not that harmful to Labor compared to ordinary votes in this seat and I expect that to be the case again. While my current projection is impossibly close (even closer than the ABC's 50.3 to Labor), it has a large error margin because of the nature of booths in the count and what will most advance the projection is seeing more booths realigned in the seat. It may also be that the flow is affected by the relative Labor to National primaries as well as the flow from that.
Monday 10 pm: More postals caused the ABC's projection of the live count to flip to Lethlean but caused my projection of the Labor preference flow to increase so that I now have Labor on live 50.33%(because I treat postals as a single dot point rather than let them drown the sample, though that has its limits when some of the competing booths are tiny!)
Tuesday 5:40: More realignment booths have come in and my model has put Lethlean ahead now with a wafer thin 50.02 2CP (that is after I remove tiny booths; with them in Chesters still leads).
Wednesday 2:00: Some more realignment booths that have come in are mostly better for Chesters. I am now running two projections, one off Labor's primary vote and one off the Liberal share of the preference pool. The former is on 51.7 for Chesters, the latter 50.4. They're both equally explanatory so perhaps the truth is in the middle. Lethlean is leading the live count on 50.55 but it is very prepoll/postal heavy compared to booth votes.
Thursday night: Watching the realignments I am now satisfied that Chesters wins. I base this on the following. The live realignment has Chesters 1391 ahead but is missing many booths still. However all of these are now day booths. In the day booths realigned so far Chesters has 32.8% Lethlean 26.0% Liberals 8.0%, and Chesters has 53.6% of preferences and 54.9% 2CP. In the day booths still to be realigned Chesters has 34.0% primary Lethlean 27.9% Liberals 9.5%. This gives Chesters a further 1054 votes in gains just off the primaries in these booths that Lethlean would have to pull back on preferences together with Chesters' existing 1391 lead - but there's no reason to think he can pull back more than a very small amount (because of their higher Liberal primary) based on the general similarity of these votes with the other votes where Chesters is gaining on preferences. To get even within 500 and have some sort of hail mary shot on remaining votes Lethlean would have to pull back 2000 votes which would require this large set of booths to have a preference flow around 20 points different from the other day booths despite having only slightly more favourable primary votes. What's more, even though these day booths are not yet rechecked, all of them are small, so random errors of 50 or 100 votes here cannot dislodge the projection in a way that a 500 vote bundle in the wrong pile in an urban prepoll can. I am satisfied it is appropriate therefore to call this seat.
Bullwinkel (ALP notional 3.4%)
Tuesday 80.0% counted: I finally get around to a first look at Bullwinkel, the new notionally Labor seat on Perth's eastern fringe, by population in large part ex-Hasluck but including conservative hinterland. This is a three-cornered contest where former WA state Nationals leader Mia Davies has polled solidly but won't reach the final two. Presently Labor's Trish Cook leads the Liberals' Matt Moran by 50 votes. It's hard to project anything reliably with this seat because it is a hybrid of other seats but it is notable that the mail is mostly in in Bullwinkel; only 2865 postals were sent out and not yet counted, and the bulk of those will probably never return. The postals are pretty toothless anyway, currently running only one point better for Labor than ordinaries. Unless Labor does badly on out of electorate prepolls I would expect Labor's usual strong performance on absents to give them the advantage here, especially in a seat with a decent Green vote. So I think Labor are slightly better placed here pending rechecks.
Thursday night: Especially well placed after corrections worth about 200 votes in their favour in rechecking - one that added 112 votes to Labor's lead in the Toodyay booth and a raft of little corrections across many booths. Labor also gained on the first absents and leads by 333.
Friday: Cook now leads by 386. I am waiting for the recheck of the large Midland PPVC (around 9000 votes) and if Cook gets through that one unscathed then I think it will be appropriate to call this seat.
Friday 7:30: Not only has Labor got through that but they've moved to a 634 vote lead after more strong absents. More than enough to take this one off the watch list.
Longman (LNP 3.1%)
Tuesday 79.5% counted: Longman popped up for a few mentions as an LNP Queensland seat that might fall if things went well for Labor. At the moment the LNP are in the mix to save it while having seemingly lost four seats further up the tree. This is a straightforward two-party race where Terry Young (LNP) leads Rhiannyn Douglas (Labor), currently 50.20-49.80 (439 vote lead). Checking for booth issues I did notice something odd with a 17.6% vote for Trumpet of Patriots candidate Benjamin Wood in the Wamuran booth and a large swing away from Labor in that booth - the UAP polled only 6.6% there previously, it might be a home booth effect or maybe an error of some kind (any information welcome). The 2PP swing in this booth is normal but the preference flow to ALP is surprisingly high if the primaries really are correct (228 votes TOP 189 votes ALP).
Postals have so far broken only a few points better than ordinary votes to Young, which also happened in 2022 there may be 4000 or so left. In 2022 there were 2655 absents which broke 5.5 points better for Labor than ordinaries, 226 provisionals which broke 16.5 points better and 4187 out of division prepolls which broke 2.0 points better. Based on the numbers of envelopes issued it is looking like the absent and prepoll votes could increase. Accounting for some possible weakening of the already weak postal flow in late postals, it seems to me that there could be enough juice in the other categories to overturn Young's current lead and then a little bit more. The primary votes will also all need to be rechecked before we can be too confident about the exact status of the 2PP here; several including the above sleighted Wamuran have seen no action since Saturday night.
Thursday 5:30 Young now leads by 320 after absents broke 383-280 for Douglas. It's unreliable to project off absents to other absents without knowing where they're from but a start for Labor pretty much in line with the past.
Friday 2:40 Young leads by 290. The first out of division prepolls broke 361-320 to Douglas and the first provisionals (most of them in fact) broke 127-87. Young must have gained a few dozen from somewhere too. There are still 16 booths awaiting rechecks. If the patterns in the current vote types continue Douglas will win by about 350, but it is unreliable to project off absents or out of div prepolls without knowing what boundary of the electorate they were from.
Friday 7:15 A second lot of absents were slightly weaker breaking 268-214 to Labor. Young leads by 231 but I still have Douglas by 340 on projection (and that's ignoring Waruman, correction of which may benefit them.)
Saturday 3:10 A big shift in outlook for now as Young does well 509-459 on out of division prepolls. These may be from somewhere not representative but he is now ahead on this category! If he can neutralise out of division prepolls he should at least be very close to holding on though past form suggests this won't be easy. Young leads by 249. Seven booths are awaiting rechecking including two very large prepolls. Note that the Waruman issue was fixed (50 TOP primaries were actually Labor and 100 didn't exist) but it didn't help Labor's 2CP.
Saturday 5:30 Young lead down to 219, I believe on rechecks. Still five PPVC centres to be rechecked.
Monday 2:20 Four now, and Young leads by 221.
Monday 4:45 Two more booths rechecked, Young leads by 236.
Monday 5:15 Young's fighting hard on these rechecks, or his scrutiners are fighting hard, or something. He's now 256 ahead. One recheck to go.
Tuesday 12:30 The rechecks are finished and now Young leads by 193. At the rate of votes so far I project Young to hold on by about 100. However the main issue now is what happens with the remaining out of division prepolls. They have been slightly good for Young so far compared to ordinary votes, but that might be because of where they're from. If they finish up the same relative to ordinary votes at the end of the count then Young would lose by about 30 votes.
Tuesday 5:10 Young now leads by 162. Potential votes remaining: absents 1387 (expect 1297 formal), provisionals 270 (expect 129 formal), dec prepolls 2675 (expect 2457 formal), known postals 538 (expect 512 formal), unknown postals a few hundred. Not sure where the 31 vote shift came from yet. Projection now has Young by only 23 ignoring remaining postals which can at this stage do anything, so there is a high chance of a recount and Douglas still has real chances to win here.
Wednesday: Young lead down to 121 and still maybe 2500 to go, very much a live contest
Wednesday 3:35: Young lead now 128 after more postals. The bulk of what is left now (maybe c. 2100 formal) is out of electorate prepolls. It is expected 900 of these will be counted today.
Wednesday 4:20: And suddenly Terry Young has stepped up to the plate and SMASHED IT OUT OF THE PARK on today's out of electorate prepolls. He is 335 ahead and won't be caught on the remaining 755 votes worth of scraps and whatever postals are to come. Sorry Longman not even a recount for you.
Saturday: Longman is the first seat in Australia to have "gone to zero" as I call it, meaning there are no votes left to count in the initial count, with Young leading by 311. The formal distribution of preferences is the next stage in which only very minor changes will occur.
Monday 19th: I spoke too soon, another 17 were found on the last plane from Nairobi or wherever (and also it seems Bullwinkel was first), so those have still to be counted.
Menzies (Lib -0.4%, notional ALP)
Sunday 76.1% counted: Adding this one first because of the interest in the fate of an MP touted (including by me if anyone asks) as a potential future leader and because the ABC has moved it back to ALP Likely. In Menzies in 2022, Keith Wolahan's final 2PP was 1.00 points higher than his 2PP in ordinary votes (which includes booth votes and within-division prepolls). This time he needs to double that gap as Labor's Gabriel Ng starts with an ordinary vote of 51.98% 2PP. Is there any hope at all?
Firstly, better postals. In 2022 Wolahan's postal 2PP was 6 points higher than ordinary votes. In the postals counted so far it is 10 points higher. But that's probably just the normal conservative behaviour of early-counted postals. So that probably won't do it. Secondly, more postals. In 2022 Menzies took 19734 postals. In 2025 it has issued a staggering 30340 postals; maybe 25000 or so could get returned and admitted, so that does look like a substantial increase in postals. Thirdly there are a couple of small outstanding PPVCs, Heidelberg and Ringwood, and the latter was slightly better than average for Wolahan in 2022, so the ordinary vote count might come down a little. Fourth, given the increase in prepoll and decline in day voting, there may well be slightly more out of division prepolls and slightly fewer absents.
Counteracting that, there is what looks like a 194 vote error that when corrected favours Ng in the booth Wattle Park. Here the Liberals have been credited with a gain of 176 preferences and Labor with 79; it appears those numbers should be the other way around (a common error in early counting, generally fixed in checking).
All this taken into account I find it hard to get the final margin below about 1000 in Ng's favour and expect that Labor has indeed gained Menzies.
Hi Kevin, I've heard some people on social media claim that overall turnout is down significantly. Are they just misinterpreting the current live count?
ReplyDeleteWe always get this. The time for talking about overall turnout is when all the votes are counted which takes weeks. People look at an incomplete turnout and say it's down when it's not finished yet. Have not tried to project where it is going to finish yet. Some seats at least have healthy %s counted for this time.
DeleteRe: "the normal conservative behaviour of early-counted postals" - do you have an explanation of why early-counted postals skew more heavily conservative than late-counted postals? I'm trying assess Zoe Daniel's chances of holding on in Goldstein. (They don't look good.)
ReplyDeleteFirstly older postal voters (who skew conservative) are more likely to vote quickly on getting their ballot and put it in the mail, whereas working-age postal voters are more likely to take time to do it (there's also a stereotype about lefties being more disorganised though they may also be wanting to see more of the campaign before voting). Secondly late arriving postals are more often travellers posting in from remote areas, these tend to skew left also by virtue of age among other things.
DeleteThank you very much for your analysis and keeping us informed. Sorry if this is a silly question, but why is it that the first batch of postals counted tend to be more conservative and the later-counted ones less so?
ReplyDeleteWhat about Ryan? What is going on there? I see votes shrinking between Labor and Greens but there is not a lot of postals left to count so surely Greens have this seat and it should be called soon?
ReplyDeleteGiven the high informal vote in high-ESL seats with large numbers of candidates, would it be worth considering a 6-preference savings provision for the house to account for people confusing it with ATL senate. Seems like a good compromise to enfranchise as many as possible without going to full OPV, your thoughts?
ReplyDeleteI have an old article about the unintended informal voting problem here. https://kevinbonham.blogspot.com/2020/07/how-should-we-solve-problem-of.html I actually quite like semi-OPV (1,2,3 required). Full OPV I am increasingly strongly against. However anything that allows votes to exhaust creates issues for election statistics so it's possibly worth looking at the best solutions that keep everything as a full vote (somehow)
DeleteConspiracy theories abound: Where did all those postal votes for Wilson come from? A lot of private nursing homes in Goldstein. Plus Advance, religious groups claimed to be up to some kind of postal tricks. Sans proof, it's all just nonsense. But would it be possible to detect any such anomalies in the postal vote flows, if they did exist. Or prove that such claims are baseless? I couldn't, but maybe you could.
ReplyDeleteIs it me or the Bendigo re throw of preferences really slow. Bean has re did their preference count while Bendigo struggles to do many each day. Is it transportation issues or something else?
ReplyDeleteBendigo is slow - one reason Bean is fast is that it only has four candidates. There may be other factors.
DeleteRandom question: how is the statistical 2pp for the national vote calculated for seats with 2 coalition candidates?
ReplyDeleteIt is calculated based on the two-candidate-preferred result between Labor and the highest finishing Coalition candidate, whichever party they belong to.
Delete