Our tool gives an overview of predicted results in each constituency (in the left tab) and the overall results (in the right tab). You can adjust these results in a couple of ways:

Changing the date: you may want to see what our model predicted at an earlier point in time e.g. how would an election have gone before and after Liz Truss’ short tenure. To do this, simply slide the date slider left and right to find the dates you’re after. The tool will give you results between the start of 2022 and the present.

Changing the swing: the results shown by default by the tool are Bradshaw’s ‘best guess,’ based on a number of simulations of the data. However, you might have a hunch that the Tories might have a better day than some pundits are expecting or perhaps that Labour’s potential winnings are being undersold. To swing results in the Tories’ favour, simply move the swingometer’s slider to the right and to swing them in Labour’s favour, swing them to the left.

Bradshaw Advisory is a blend between an economics and data consultancy, public affairs agency and a policy think tank.

Our team have worked as civil servants, advisers to Labour and Conservative politicians or in leading think tanks.

Based on the latest poll of polls and our proprietary modelling using econometric techniques (brief methodology below), Bradshaw’s Election Prediction Model gives an up-to-date prediction of how an election would pan out in each constituency in Great Britain and by extension, who will be the next government. The model is:

Responsive to a fast changing political environment. The model does not require a large poll to be undertaken - instead using poll of polls. Bradshaw’s model will be refreshed to reflect changing conditions fortnightly.

Reliable - Bradshaw’s model was 92% accurate when run over the last three general elections - a similar performance to YouGov’s leading MRP modelling.

Savvy, as the underlying data has been processed with the shape of British politics in mind, with results changing to reflect shifting constituency boundaries, by-elections and narrowing poll leads in the run-up to elections.

In broad strokes, Bradshaw Advisory’s Election Prediction Model provides estimates of the likelihood of parliamentary constituencies changing hands were a general election called immediately. The model takes into account the national-level polling performance of the parties in each seat, as well as data regarding the past behaviour of that seat, for example, how many times in the last four elections has the given seat changed hands? These figures help build a picture of how likely the seat is to change hands.

Once each seat has had its change probability estimated, those figures are run through a number of scenarios which provide us with a picture of how a general election would play out at different points in time. We provide a ‘central estimate’ of what the model suggests is likely to happen were an election called tomorrow, as well as a range of other results which allow users to ‘swing’ the election slightly, depending on their own views of how things could change in the near future.

You can find a full methodology in the appendix of our report, here.

You can find an in-depth analysis of the model’s predictions across 2022 and early 2023 in our report linked here. The report also contains detailed information on the model’s methodology and potential applications.

Alongside the model, we have produced a suite of other tools which can be used by anyone with a keen interest in Britain’s electoral landscape. These include our:

Voter Spread Index: while any party would be pleased when they see their popularity go up in the polls, where that popularity is being gained is massively important for electoral outcomes. If a party were to gain voters in areas it is likely to win anyway, this is not particularly helpful when it comes to winning elections. Our Voter Spread Index measures how dispersed voters of each party are in each region of the UK, highlighting those areas in which parties run the highest risk of seeing their vote too highly concentrated.

Outlier MPs: in each region of the UK we have identified those MPs who, at the last election, over-performed or underperformed relative to the others in their party in their given region, giving an indication of where the political high-flyers are and where electoral contests may buck regional trends in the future.

Vulnerable constituencies: we have identified the constituencies which have the highest probability of changing hands across Great Britain, providing a hint to incumbent parties and to potential challengers as to where they should focus their resources.

If you would like to make use of these further tools, you can download them here.

Predicting the future is hard. As with all models they are a ‘best guess’ of what may be likely to happen and not set in stone.

This is the first iteration of our Election Prediction Model (EPM) and it will be updated and improved in the future.

The main shortcoming of our model is its assumption that if a given seat is likely to change hands, it will normally go to the party that came second in the seat’s last contest, be that in 2019 or in a recent by-election. Given that most seats across the UK have historically been two-horse races, this assumption is reasonable in many cases. However, in some cases there are multiple challengers who have a chance to take the seat, which makes the assumption less likely to hold.

Polling in Scotland and Wales is also less frequent so the results there may not be as responsive as they are for English constituencies.