Who forms government on Spring Street?

A probabilistic forecast of the 88-seat Legislative Assembly. Forty-five seats is a majority. Every number below comes from simulated elections that combine the polling average, a fundamentals prior, seat-level primary-vote projections and full preference counts.

Outcome

Seat map

Seat distributions

Probability of each seat total. The dashed line marks 45. Coalition = Liberal + Nationals.

Crossbench size

Probability pendulum

All 88 seats ordered by the Coalition's chance of winning, least likely on the left. Colour shows the favourite; paler means less certain. The dashed line marks off the Coalition's 45 most winnable seats. Select a seat to find it in the table.

Seats most likely to decide government

Share of simulations in which the seat delivers the 45th seat for the Coalition or for Labor, averaged.

What would it take?

Coalition two-party-preferred vote (vs Labor) at which its chance of a majority reaches 50% and 80%, from a logistic fit across simulations. One Nation's vote also matters, so the threshold is shown by One Nation's statewide level.

Polling average

Every seat

Margin 2022 is the winner's two-candidate share minus 50. "UNS" is the classic pendulum: 2022 Labor 2PP plus a uniform statewide swing, shown for comparison only.

How this compares with other forecasts

Seat projections published by others, beside this model's median and 80% range (shaded). Hollow dots are this model rerun under other assumptions, which reproduce most of the disagreement. The dashed line marks 45.

Markets

Why they differ

    No published Victorian projection from Anglo Election Insights was found; its analysis appears to be in videos that could not be reviewed. Dates shown as reported by each source; blank where none was given.

    How the method did before

    The two-party core of the model rerun on 2018 and 2022 using only the final pre-election polls. Seat noise for each year is calibrated on the other year.

    Read this before quoting a number

    Configure assumptions

    Change any setting and rerun the full model in your browser. Defaults are the published forecast. Nothing you change here is saved or shared.