Aggregating Predictions for the General Election

This site aggregates constituency-level predictions for the 2024 British general election.

Thrown together by Peter Inglesby. Drop me a line: peter.inglesby@gmail.com!

View the source code on GitHub; download the data here.

Notional 2019 results are from Colin Rallings and Michael Thrasher; click links in summary table headings for more about each model's data.

Scroll down for seat-by-seat predictions; compare tactical voting recommendations here.

Summary

This summary shows, for each model and each party, the number of seats where that party is predicted to get the most votes. For more detail, here's a breakdown by margin.

Note that these totals are indicative of the parties' level of support from each model's prediction, but won't match the number of seats predicted by each model.

To understand why, imagine two parties and ten constituencies, where a model predicts that party X will get more votes than party Y, and assigns party X a 90% chance of winning. In this case, the model may predict that X will win nine seats and Y one. However, in this summary, we will instead show X with ten and Y with none.

Electoral Calculus appears twice: they release a regularly-updated forecast based on their Strong Transition Model, and they have published a single forecast based on a Multilevel Regression and Post-stratification (MRP) model. Other models using MRPs are Focaldata, Ipsos, More in Common, Savanta, Survation, WeThink, and YouGov.

Constituencies

Showing predicted/actual winner for all constituencies