Terrain

Notes · contest

Polling

Twenty-eight polls of Victorian voting intention were published in 2026. What do they agree on, where do the houses disagree, and when does the poll average stop telling you the result?

pollingvictoria

A sign at the entrance to a polling place on referendum day in 2023.
Aliceinthealice CC BY-SA 4.0

A sign at the entrance to a polling place on referendum day in 2023.

Where the contest stands

Victoria is a three-cornered contest with a two-party outcome. Two-party preferred means Labor and the Coalition head to head, with every other party’s votes handed on as preferences. The Coalition leads on every published 2026 average that carries the figure: 54.0–46.0 on the Ace Strategies post-change mean of 26 October Ace Strategies , and 52.0–48.0 on BludgerTrack of 19 October Poll Bludger . The Australian Election Tracker publishes no two-party figure at all Election Tracker . Neither old party is near a majority on first preferences. Across the three averages the Coalition sits between 29.3 and 30.4 and Labor between 24.2 and 24.4.

What decides the seat count is where One Nation’s preferences go, and no house has measured that flow at seat level. One Nation holds 22.5% of first preferences on the post-change average Ace Strategies . Ace Strategies puts the whole difference at about three points: moving the flow from 75–25 to 60–40 takes the Coalition from 53.0 to 49.6 on two-party preferred, with not one first preference changing Ace Strategies .

The series, 2023 to October 2026

From 2023 to the middle of 2025 Labor’s primary sat in the high thirties on the back of the 2022 win, then slid into the low thirties, while the Coalition sat between 33 and 40. Resolve put Labor on 39 and the Coalition on 28 in July–August 2023 and RedBridge put them on 37 and 36 in December 2023, with Labor ahead 55.9–44.1 on two-party preferred Wikipedia .

By November 2025 Newspoll had the Coalition ahead 51–49, with Labor on 28 and the Coalition on 36 Newspoll . Then, in early 2026, One Nation entered every house’s questionnaire at once. DemosAU’s 1–10 February wave put it on 21 DemosAU . RedBridge’s 18–27 February wave put it on 24 RedBridge . Roy Morgan’s 13–16 February wave had it on 26.5, ahead of both majors, with Labor on 25.5 and the Coalition on 21.5 Roy Morgan .

One Nation’s high-water mark in the series is 27, in the RedBridge/Accent wave of 14 July RedBridge . Freshwater’s 5–8 June poll gave a second reading of 25 Freshwater . Labor’s low-water mark is 21 in DemosAU’s 7–11 June wave. That is the lowest primary in the series, below the 1964 low that the 2022 result itself had matched DemosAU .

From July, One Nation eases back to the low twenties and Labor steadies between 24 and 27. Through August and September the Coalition’s line barely moves. In the closing month it breaks upward, reaching 33 in Newspoll’s 15–22 October wave, the highest primary of the series, and takes back most of what One Nation gave up Newspoll .

What the two leadership changes did

Labor changed leader on 28 July 2026, when Ben Carroll replaced Jacinta Allan ABC . Comparing each house against itself, which removes the house effect, the change cost Labor almost nothing and One Nation a little. Across the five houses that polled on both sides of it, the moves were Coalition +2.2, Labor −0.1, One Nation −1.8 Ace Strategies . The one poll fielded as the leadership changed was RedBridge’s 28 July–1 August wave, the sharpest of the cycle: One Nation −5, Coalition +4, and the two-party reading moving from 54–46 to 57–43 RedBridge .

On 3 August Warren Pickering became One Nation’s Victorian leader and its candidate for Pakenham ABC . Roy Morgan’s 5–7 August wave, the first after both changes, gave him 35% positive against 58.5% negative, a net of −23.5 and the worst rating of any leader in the dataset Roy Morgan . By Resolve’s October wave his preferred-premier share had risen from 8 in September to 14 Resolve .

Neither change moved the direction of the contest. Resolve has had Labor on 25 for three consecutive months: August, September and October Resolve . Newspoll’s October wave has it on 24, with the Coalition at its highest primary of the year Newspoll .

One poll, four pairings

Roy Morgan’s 5–7 August wave reported the Coalition ahead 51–49 on two-party preferred. The same poll’s three-party preferred figure was Labor 43.5, Coalition 30, One Nation 26.5: Labor first, by a distance Roy Morgan . The two figures answer different questions. The two-party pairing distributes One Nation’s preferences, while the three-party figure keeps One Nation in the race, where the right-of-centre vote is split between two parties. The same release carried a third pairing, Coalition 63.5–36.5 over One Nation, and a fourth, Labor 56–44 over One Nation. Four pairings, four winners. Which party looks like the winner depends entirely on which two are put head to head.

Where the flow goes decides the seat count. If One Nation’s preferences reach the Coalition at even a moderate rate, the Coalition wins on two-party preferred. If the flow is weaker, or One Nation runs ahead of Labor in outer-metropolitan and rural seats, the same votes deny the Coalition the majority that a uniform swing implies Ace Strategies .

House effects

The differences between the pollsters are bigger than the movements most headlines report. On the same fortnight the Coalition’s primary spans 6 points, from 26 to 32, and since the change of premier, ten two-party readings from six houses have ranged from 51–49 to 57–43 — a spread of 6 points. Sampling error on that measure is about 1 Ace Strategies . RedBridge/Accent is the systematic outlier: its 56–44 readings appear twice against other houses’ 52–48 and 53–47 in the same period Poll Bludger .

The houses collect their numbers differently, and that is the likeliest explanation. Roy Morgan surveys by SMS and publishes three party pairings. DemosAU adjusts the 2022 flows and applies a 2025 federal One Nation flow. Newspoll uses the preference flows from the previous election, while most others ask respondents where they would send their preferences. YouGov’s single entry estimates the result from a model. Nobody counted stated preferences for it Ace Strategies .

Seat models

The Coalition needs 45 of the 88 lower-house seats for a majority. That is a uniform swing of 7.6 points on two-party preferred, or a two-party figure of 52.6 Ace Strategies . The post-change average implies 9.0 points, which clears that line comfortably. The published seat models don’t clear it at all. Both mid-2026 models that produce a seat-by-seat count put the Coalition on 39 of 88, six seats short Ace Strategies .

YouGov’s MRP, a model that estimates each seat from a large survey, has fieldwork 16 June–10 July and n=4,003, and gives the Coalition 39, Labor 29, One Nation 17 and the Greens 3 YouGov . RedBridge/Accent’s MRP for the Victorian Trades Hall gives the Coalition 39 and Labor 35, with One Nation on 12 Poll Bludger . Both were fielded before the change of premier, and both have been criticised for their preference assumptions, in opposite directions Ace Strategies .

The more recent Roy Morgan model, March–September and n=7,708, has it closer still and calls the election too close to call. It gives Labor and the Greens 46 seats against the Coalition and One Nation’s 42, with One Nation set to win more than 10 Roy Morgan . The Swing Is On’s simulations put the Coalition on an average 37.3 seats of 88 and give One Nation between 1 and 27 seats in 90% of runs The Swing Is On The Swing Is On .

What could still be wrong

  • No house has tested One Nation’s preferences seat by seat. Where those preferences land is worth about three points of two-party preferred Ace Strategies .
  • House effects are larger than sampling error. A 6-point spread on the same fortnight can’t come from sample size, so don’t read a single poll as movement. Published bands that show sampling error alone understate the uncertainty Ace Strategies Poll Bludger .
  • One Nation’s vote has never been converted into seats in Victoria. Its 24.5% in the May Nepean by-election came in a seat with no Labor candidate, so it doesn’t say what a statewide figure does Ace Strategies .
  • The right bloc is bigger than the left, and has been for months. On the tracker’s 22 October average the Coalition and One Nation hold 52.2% of the primary vote between them, against 38.9% for Labor and the Greens. Labor’s decline didn’t move votes to the Coalition. They went to One Nation Election Tracker .
  • The Greens are the one line still moving in the closing month. Their October readings of 15 and 16 are the highest of the year, and Resolve records a three-point rise since September. Newspoll’s late-October wave has them on 14 Resolve Poll Bludger .

Summary

  • The averages agree on the shape of the contest and on little else. Two of the three carry a two-party figure, and they put the Coalition 2.0 and 4.0 points in front of Labor.
  • One Nation never fell below 11 through the 2026 series and peaked at 27 in the RedBridge wave of July. Labor’s highest primary of the year is 28.
  • The poll average implies a majority and the seat models deny one. The difference is where One Nation’s preferences go next, which no Victorian count has measured seat by seat Ace Strategies .
  • House spread runs to 6 points where sampling error is about 1, so treat a single poll as one reading rather than a movement Poll Bludger .
  • The research names One Nation as the challenger in nine of the 88 seats. The party has never converted its Victorian vote into seats.

The series, and every poll in it

Three 2026 averages and the last counted result
Aggregate As at ALP L-NP ONP GRN Oth Two-party preferred
Ace Strategies post-change mean Ace Strategies 26 Oct 24.4 29.3 22.5 13.8 10.8

Coalition 54.0–46.0
Australian Election Tracker (5-poll, 45-day, half-life 21d) Election Tracker 22 Oct 24.3 30.4 21.8 14.6 8.9

n/a
Poll Bludger BludgerTrack Poll Bludger 19 Oct 24.2 30.2 21.9 14.7 —

Coalition 52.0–48.0
2022 election result Wikipedia 26 Nov 2022 37.0 34.5 0.3 11.5 16.7

ALP 55–45

Two of the three 2026 averages put the Coalition ahead on two-party preferred: by 2.0 points on BludgerTrack and 4.0 on the Ace post-change mean. The election tracker publishes no two-party figure at all. Take the two averages closest in date, 19 Oct and 22 Oct, and they agree on Labor’s primary vote to 0.1 points, and on One Nation’s to 0.1. Ace StrategiesPoll BludgerElection Tracker

Legislative Assembly primary vote, 2026: 28 polls for Labor, the Coalition, One Nation and the Greens A line plot of four primary-vote series across 2026, one point per poll at its fieldwork midpoint, January at the left and October at the right. Labor runs from 28 per cent in late January to 24 in October. The Coalition runs between 21.5 and 33, ending at 33. One Nation runs from 11 in late January to 26.5 in mid-February, peaks at 27 in late June, and stands at 20 in late October. The Greens move between 10 and 16. Five changes in the running order are numbered one to five and listed beneath the chart. Two dated rules mark the leadership changes of 28 July and 3 August 2026. 353025201510 FebMarAprMayJunJulAugSepOct PRIMARY VOTE, % scale runs 9 to 35 per cent, not zero fieldwork midpoints, Jan–Oct 2026 · 28 polls 28 Jul — Allan → Carroll (Labor) 3 Aug — Pickering leads One Nation 1 2 3 4 5
  • Labor
  • Coalition
  • One Nation
  • Greens
Legislative Assembly primary vote, 2026: 28 polls for Labor, the Coalition, One Nation and the Greens A line plot of four primary-vote series across 2026, one point per poll at its fieldwork midpoint, January at the left and October at the right. Labor runs from 28 per cent in late January to 24 in October. The Coalition runs between 21.5 and 33, ending at 33. One Nation runs from 11 in late January to 26.5 in mid-February, peaks at 27 in late June, and stands at 20 in late October. The Greens move between 10 and 16. Five changes in the running order are numbered one to five and listed beneath the chart. Two dated rules mark the leadership changes of 28 July and 3 August 2026. 353025201510 FebAprJunAugOct PRIMARY VOTE, % scale runs 9 to 35 per cent, not zero fieldwork midpoints, Jan–Oct 2026 · 28 polls 28 Jul: Allan → Carroll (ALP) 3 Aug: Pickering (One Nation) 1 2 3 4 5
  • Labor
  • Coalition
  • One Nation
  • Greens

Numbered marks: where the running order changed

  1. 1 10 Feb One Nation passes the Coalition — the first reading in the series with a minor party ahead of a major.
  2. 2 21 Feb The Coalition retakes the lead from Labor, and holds it for the next three months.
  3. 3 3 Jun One Nation passes Labor on its way to the June peak, and stays ahead of it for six weeks.
  4. 4 20 Jun One Nation passes the Coalition to lead the field, briefly, at its June peak.
  5. 5 9 Sep One Nation passes Labor a second time; the 13 Sep Resolve poll puts it back behind.

Each mark is solved from the two polls either side of the crossing, so it sits where the lines cross.

Labor’s primary vote ends October at 24, having drifted down all year; the Coalition’s line rises gently, peaking at 33 in 15–22 Oct; One Nation’s climbs from 11 in late January to 27 in 17–28 Jun, then comes back to 20 by October. The Greens move between 10 and 16 and end at their highest reading of the year.

We don’t join the polls into one trend line. No house polls every fortnight, and the gaps between houses are bigger than the movement they’re measuring.

Every poll, as published: Legislative Assembly primary vote, 2026, oldest first. "Others" is as published by each house and may include independents. A dash means the house published no such figure.
Fieldwork House Client ALP L-NP ONP GRN Oth 2PP Source
12 Jan–14 Feb Resolve The Age 28 30 11 12 18 — Resolve
1–10 Feb DemosAU Premier National 23 29 21 15 12 L-NP 53–47 DemosAU
13–16 Feb Roy Morgan self-published 25.5 21.5 26.5 13.5 13 L-NP 48–52 Roy Morgan
19–23 Feb Freshwater Herald Sun 28 27 23 13 9 L-NP 50–50 Freshwater
18–27 Feb RedBridge/Accent AFR 25 28 24 13 10 L-NP 52–48 RedBridge
19–23 Mar Freshwater Herald Sun 27 30 20 14 9 L-NP 52–48 Freshwater
Mar–Apr Resolve The Age 27 29 21 10 13 — Resolve
22–24 Apr Roy Morgan self-published 25.5 24 24.5 13.5 12.5 ALP 51–49 Roy Morgan Roy Morgan
May–Jun Resolve The Age 26 26 24 12 12 — Resolve
5–8 Jun Freshwater Herald Sun 23 27 25 14 9 L-NP 53–47 Freshwater
7–11 Jun DemosAU Premier National 21 30 23 15 11 L-NP 55–45 DemosAU
17–28 Jun RedBridge/Accent AFR 26 26 27 13 8 L-NP 54–46 RedBridge
16 Jun–10 Jul YouGov MRP Common Threads 23 26 24 13 14 L-NP 53.8–46.2 YouGov
12 Jul Resolve The Age 27 27 22 12 12 — Resolve
14 Jul RedBridge/Accent Victorian Trades Hall 26 26 27 12 8 — RedBridge
23–26 Jul Newspoll The Australian 28 31 19 13 9 L-NP 52–48 Newspoll
28 Jul–1 Aug RedBridge/Accent AFR 23 30 22 14 11 L-NP 57–43 RedBridge
31 Jul–3 Aug Freshwater Herald Sun 25 30 22 14 9 L-NP 52–48 Freshwater
5–7 Aug Roy Morgan self-published 26 26 23.5 12.5 12 L-NP 51–49 Roy Morgan
6–11 Aug DemosAU Premier National 23 32 22 13 10 L-NP 55–45 DemosAU
16 Aug Resolve The Age 25 27 23 12 13 L-NP 53–47 Resolve Resolve
1–14 Sep RedBridge/Accent AFR 24 29 25 15 7 L-NP 56–44 RedBridge
13 Sep Resolve The Age 25 28 23 12 13 — Resolve
10–20 Sep DemosAU Herald Sun 25 30 20 14 11 L-NP 52–48 DemosAU
24–28 Sep Freshwater Herald Sun 24 29 22 14 11 L-NP 53–47 Freshwater
2–8 Oct RedBridge/Accent AFR 24 30 24 16 6 L-NP 56–44 RedBridge
6–10 Oct Resolve The Age 25 28 23 15 9 — Resolve
15–22 Oct Newspoll The Australian 24 33 20 14 9 L-NP 55–45 Newspoll

Two-party preferred is printed as each house published it. A dash means that wave carried no figure, most often Resolve, which doesn’t compute a two-party result at all. Two rows carry their own citation because their figure comes from somewhere other than the poll’s release: the 16 Aug Resolve wave published none Resolve , and the 22–24 Apr row reads 51–49 in the note where Roy Morgan’s release labels that pairing Labor-first, so we show it the way the house did. Roy Morgan The 2022 baseline is the last counted result, not a poll. Wikipedia

The houses, and how each collects its numbers

Seven houses produced Victoria’s 2026 voting-intention polls. Between them they ran 36 waves from June 2025 to October 2026, and the 28 in the 2026 calendar year are the ones plotted above. Ace Strategies Each poll has its own source note with the published figures, the sample and the URL.

One house you might expect is missing. Guardian Essential hasn’t published a Victorian state voting-intention wave this cycle — all its polls have been federal. Essential

Samples run from about 1,000 to 6,500, plus one pooled MRP wave of 7,708. Where a house published a two-party figure, it’s listed beside the primary vote; where it didn’t, that cell is a dash. Ace StrategiesRoy Morgan

Resolve
7 waves · The Age online panel, respondent-allocated preferences
RedBridge/Accent
6 waves · AFR and the Victorian Trades Hall online panel, respondent-allocated preferences
Freshwater Strategy
5 waves · Herald Sun online panel, respondent-allocated preferences
DemosAU / Premier National
4 waves · Herald Sun and self-published online panel, adjusted 2022 flows plus a 2025 federal One Nation flow
Roy Morgan
3 waves · self-published, by SMS SMS survey, respondent-allocated, three party pairings
Newspoll
2 waves · The Australian online panel, previous-election preference flows
YouGov (MRP)
1 wave · Common Threads online panel, model estimate (MRP)

What the polling is about

Voting intention tells you how the vote is splitting. It doesn’t tell you what’s moving it. The research tracks eleven issues, and this is where each one sits against the polling that measures it. Cost of living and crime lead because the polls put them there; the rest follow in order of how directly they feed a measured issue.

Each line below opens a note: what the issue is, where its evidence lands, that evidence sorted by whether it supports, complicates or challenges the claim, and what it means for campaigning. The order is Terrain’s arrangement of the research, not a finding of its own.

  1. 01

    Cost of Living and Energy Prices

    The single most salient issue, and the one that has flipped owner: 68% put it in their top three and it is the top issue for 47% of Labor defectors The Conversation .

  2. 02

    Crime and Community Safety

    Second-most salient at 56% top three; among One Nation converts crime is the leading concern, at 79% The Conversation .

  3. 03

    Health and Ambulance Response Times

    Rated top three by 38% overall and 43% of Labor defectors, with Labor’s ownership “relatively soft” The Conversation .

  4. 04

    Housing

    DemosAU found housing only 5th in Victoria at 8% — a lower-salience finding than in other states — while the rental data is stark Ace Strategies .

  5. 05

    Statewide Treaty

    YouGov (n≈4,000) found 40% support for treaty against 32% opposed, and more than 60% in six inner-metro Liberal seats call a 100-day repeal pledge out of touch The Age .

  6. 06

    SEC and Electricity

    Energy prices arrive as the bills inside the issue the polling ranks first; the Victorian Default Offer fell 5% for households from 1 July 2026 Essential .

  7. 07

    Emergency Services Levy

    A charge on every property owner; PBO figures quoted by the opposition put Victorians at an average $221 a person against a $177 national average ABC .

  8. 08

    Transport and Infrastructure

    Roads, the $33.3bn Suburban Rail Loop and VNI West; VAGO found 48% of state-managed roads in poor or very poor condition The Age .

  9. 09

    ‘Mass Immigration’ and Racism

    One Nation’s migration plan is the only one of the three that opens net-negative, its launch video drew the racism charge it denies, and migration is the top concern for its own voters ABCSMH .

  10. 10

    Education and Teacher Pay

    The AEU agreement ended a year-long dispute; the poll rates schools below cost of living, crime, health, corruption and debt as a top-three concern GuardianThe Conversation .

  11. 11

    Integrity and Anti Corruption

    The CFMEU/Big Build scandal and the royal commission answering it; corruption is top-three for only 15% of Labor defectors and 28% of One Nation converts The Conversation .

See where these numbers land, seat by seat All sources

Newspoll

Two waves this year. In 23–26 July the Coalition led 51–49 and Labor sat on 28, and in 15–22 October, the last poll in the series, the Coalition’s primary hit its 2026 high of 33 and Labor sat on 24.

Newspoll

Notes

  • Newspoll allocates preferences from the previous election’s flows, while most other houses ask respondents where their preferences go. That is one reason its two-party figure can’t be compared row for row.

Resolve

Seven waves for The Age, from 12 January–14 February through 6–10 October. Resolve has had Labor on 25 for August, September and October, and its October wave records Wilson gaining economic traction as Carroll’s popularity fell.

Resolve

Notes

  • Resolve publishes no two-party figure of its own, so all seven rows carry a dash in that column. The 16 August wave’s 53–47 comes from The Conversation’s write-up of it.
  • Its February wave is the first in the series, and it put One Nation on 11 — the party’s lowest reading of the year.

RedBridge/Accent

Six waves, for the AFR and the Victorian Trades Hall. It carries the series’ high readings for One Nation: 24 in late February and 27 in mid-July. It also carries the sharpest single-poll movement of the cycle, its 28 July–1 August wave, fielded as the leadership changed.

RedBridge

Notes

  • RedBridge/Accent is the systematic outlier on two-party preferred. Its 56–44 readings appear twice against other houses’ 52–48 and 53–47 in the same weeks.
  • Its 14 July wave published no two-party figure at all.

Freshwater Strategy

Five waves for the Herald Sun across the year, from 19–23 February to 24–28 September, with the Coalition’s primary between 27 and 30 throughout.

Freshwater

Notes

  • Its 19–23 February wave is the only poll in the series that returned a tied two-party figure, 50–50.

DemosAU / Premier National

Four waves, from 1–10 February to 10–20 September, published by the Herald Sun and self-published. It carries Labor’s lowest primary of the year, 21 on 7–11 June, and the Coalition’s high of 32 on 6–11 August.

DemosAU

Notes

  • DemosAU adjusts the 2022 flows and applies a 2025 federal One Nation flow, so its two-party figure is a modelled allocation rather than the order respondents gave.

Roy Morgan

Three SMS waves and one MRP. The 13–16 February wave put One Nation on 26.5, ahead of both majors. The 5–7 August wave published four party pairings in one release. The March–September MRP of n=7,708 calls the election too close to call.

Roy Morgan

Notes

  • The MRP is a seat model rather than a poll wave, so it is listed here with the house that produced it, not as a fourth wave in the series.
  • The 22–24 April row reads 51–49 in the research note, but Roy Morgan’s own release labels that pairing Labor-first. The table on this page shows it in the orientation the house published.

YouGov (MRP)

One entry, fieldwork 16 June–10 July and n=4,003 for Common Threads: Coalition 39, Labor 29, One Nation 17, Greens 3, with a two-party figure of 53.8–46.2.

YouGov

Notes

  • It is a model estimate rather than a count of stated preferences, so it is the one row in the table that isn’t a survey result.

Ace Strategies, the register and the mean

The poll register with each house’s method, plus a post-change average re-based on 26 October after Newspoll’s 15–22 October wave. It also carries a preference-flow sensitivity and its reading of both published seat models.

Ace Strategies

“Moving the One Nation flow from 75–25 to 60–40 takes the Coalition from 53.0 to 49.6 without a single primary vote changing.”

The sensitivity that separates the poll average from the seat models.

“Seats: post-change average implies a majority swing, but both published seat models leave the Coalition on 39 of 88 — six short of 45.”

Notes

  • The average is an unweighted mean of the post-change polls. Only ten of them published a two-party figure, and the bands shown are sampling error only.
  • The sensitivity is the register’s own projection, not a reading of any ballot: it holds every primary fixed and moves one flow.

Australian Election Tracker

A weighted average of five polls over 45 days with a 21-day half-life, as at 22 October. It publishes primaries and no two-party figure.

Election Tracker

Notes

  • A tracker whose weight halves every 21 days sits closer to the last fortnight than a simple mean. Because it publishes no two-party figure, it can’t be compared with the other two averages on that measure.

The Poll Bludger, BludgerTrack

The Poll Bludger’s own aggregate and seat model, its 19 October reading of 52.0–48.0, and its write-up of the RedBridge and Resolve waves. The house-by-house outlier reading on this page comes from that write-up.

Poll Bludger

Notes

  • BludgerTrack is an aggregate with a model attached. Its seat projection inherits whatever the aggregate assumes about preferences.

The Swing Is On

A published seat model simulating each seat 5,000 times, with its preference calibrations declared and its seat-level probabilities listed.

The Swing Is On

“The site's headline call for Victoria: Coalition minority government if an election were held today (avg. 37.3 / 88 seats).”

“The model gives One Nation between 1 and 27 seats in 90% of all simulations.”

Notes

  • A 1–27 seat range is the honest shape of the uncertainty. The same votes support a One Nation bloc of one seat or twenty-seven, depending on where preferences and three-cornered contests break.

ABC News and Antony Green

The ABC’s pendulum and margins for all 88 seats, its round-up of the minor-party picture and the working preference ranges both blocs use, and Green’s reading of the South Australian flows.

ABCA. Green

“On the ABC's framing, Labor's path runs through preferences (Greens ~82–93% to Labor) and a split conservative vote (One Nation preferences ~65–77% to the Coalition, historically around 60%), while the Coalition's path runs through winning back Melbourne's east and offsetting losses to One Nation in the regions.”

“The lesson from South Australia is that preference flows between One Nation and the Liberal Party are weaker than between Labor and the Greens.”

Notes

  • The ABC’s ranges describe recent polling and party expectations, not counted flows, so the seat models have to declare their own calibrations.

Wikipedia, the 2022 baseline and the pre-2026 series

The polling article the pre-2026 series is read from, and the 2022 result every swing in this note is measured against: Labor 37.0, Coalition 34.5, Greens 11.5, One Nation 0.3, on 55–45 two-party preferred.

Wikipedia

Notes

  • The 2023 and 2024 waves are not in the 28-poll series and not plotted on the chart. No wave before 2026 has a source note of its own in the evidence base.