FORECAST CHECK · 13 SEPTEMBER 2026
Live models and dated MRP polls answer slightly different questions. This page puts their headline seat totals in the same party order, then links to the full method and uncertainty at source.
ROAD TO 326 SCENARIO
30917 short of 326. This is our simple model, not a consensus forecast.
SEAT ESTIMATES
Uniform national change
Each party's change from its 2024 Great Britain vote share is applied equally to every Great Britain constituency. Northern Ireland stays at its 2024 result.
Check publisher ↗Demographic constituency model
The publisher's rendered headline names four parties. The remaining 67 seats are grouped here as other parties rather than guessed party by party.
Check publisher ↗Current General Election Prediction
Central seat estimate based on polls conducted 7-27 July 2026. The publisher also reports wide 90% seat ranges, which should be read at source.
Check publisher ↗The Final MRP of Keir Starmer's Premiership
Fieldwork ran 5-28 June 2026 with 14,358 respondents. The published GB party table accounts for 631 seats; Northern Ireland's 18 seats are shown separately here, leaving one seat unitemised.
Check publisher ↗PLMR MRP with tactical-voting adjustment
Fieldwork ran 23-30 June 2026 with 5,545 GB adults. Restore Britain is shown separately; three minority/independent seats are grouped as Other.
Check publisher ↗Reading the rows: Labour starts at the left; Reform UK finishes at the right. Figures are dated snapshots, not endorsements or voting advice.
SOURCE DIRECTORY
Only concise headline totals needed for comparison are reproduced above. Full constituency tables, ranges, datasets, charts and methodology remain with each publisher.
A rolling seven-poll average translated through a demographic model of Great Britain constituencies.
Open publisher source ↗A poll-of-polls model and separately published large-sample MRP work, each with its own assumptions.
Open publisher source ↗Large-sample constituency estimates published as dated snapshots rather than a live election forecast.
Open publisher source ↗National voting intention is an input signal, not a constituency-by-constituency seat estimate.
Open publisher source ↗Market prices reflect traders, liquidity and contract wording. They are not polls or seat projections.
Open publisher source ↗A separate trading signal whose prices and settlement terms can change. It is not polling or financial advice.
Open publisher source ↗READ CAREFULLY
Polls measure vote intention. They estimate national support during fieldwork and carry sampling and methodological uncertainty.
Seat models transform assumptions. Geography, tactical voting and local effects can produce different totals from similar polls.
Markets measure trading. Prices reflect contract wording, liquidity and participants; they are neither polls nor financial advice.