United Kingdom General electionby 15 Aug 2029in 997 days Majority for any party 36% Labour largest party 93% Calendar Track record Research
What-if tool · polls to 20 November 2026

What if the UK polls are wrong? You decide by how much.

Other forecasts start by assuming the polls are right. Here you set the polling error yourself. Move a party's slider right if you think the polls are overstating it, left if they are understating it. Or start from a real miss: the presets replay what actually happened to the final polls in 2015, 2017, 2019 and 2024.

NewTactical votingWhat if voters gang up on Reform or the Tories? off

Seat projections from national figures assume each party's voters stay put. Many don't: they back whoever can beat the party they like least. Set how many switch, on top of your numbers.

Tested on 2024: from the 2019 result with the real 2024 national votes, plain swing called 479 of 631 seats and gave Labour 329. Adding "Stop the Conservatives" at 65% called 544 and gave Labour 402 (actual 411). A right-wing pact made the 2024 test worse, as there was none. Our published projection does not include tactical voting.

Your polling error

Points by which the England polls are wrong about each party. The same error is applied to these parties in Scotland and Wales; the SNP and Plaid Cymru stay as polled.

Loading…
Loading…

Seats that change hands

Compared with the polls as they stand, closest first.

Loading…

How big is a realistic polling error?

Election CON LAB LD REF GRN What happened
2015 -5.7 +1.5 +1.8 +0.8 +0.9 Polls had a dead heat; the Conservatives won a majority.
2017 +0.6 -3.9 +0.7 +2.1 +0.6 Polls expected a Conservative landslide; Labour surged.
2019 -1.6 +0.9 -0.6 +0.8 +0.1 Polls were close to right on the big two.
2024 -3.5 +4.3 -0.9 +2.0 -0.6 Polls overstated Labour's landslide and understated the Conservatives.

Final England polls minus the England result, in points. Positive = overstated. Red: overstated by half a point or more; blue: understated. Pollster-by-pollster record →

How the swingometer works

Polls are never exact. At the last four general elections the final polls missed each main party's vote by between a fraction of a point and several points, and in 2015 and 2024 the misses changed the story. This swingometer lets you set your own polling error for each party, or replay a real miss, and see how many of the 650 seats change hands from today's polls.

Positive numbers mean the polls overstate a party, so its vote comes down; negative numbers mean they understate it. The seats are worked out with the same uniform swing as our projection, and the tactical voting settings add switching on top.

Our projection for the next general election · Set your own vote shares · How accurate the UK polls have been

If the polls miss like a past election

Today's polls in the UK with each past election's polling error applied, through our seat model. Our forecast: Labour 319 of 650, 7 short of a majority.

If the polls miss likeLargest partySeatsMajority?Seats change hands
The average past miss Labour 321 5 short of a majority 19
2015 Labour 285 41 short of a majority 56
2017 Labour 396 a majority of 142 85
2019 Labour 319 7 short of a majority 1
2024 Labour 276 50 short of a majority 69

Figures as of . Published by Polling Forecast.

You may also like

This page in other countries France Germany Spain Italy Netherlands Poland Canada Australia