DevolvedElections.co.uk is a new tool developed by Ayushi Gupta and Scottish Election Study researcher Dr. Eoghan Kelly which allows anyone to project the results of the 2026 Scottish Parliamentary Election, 2026 Welsh Senedd election and 2027 Northern Irish Assembly election. The tool works by converting user-input vote shares into projected seat counts for each legislature.
Users can project a wide range of potential outcomes to get a realistic set of results under their chosen scenario. The tool aims to make election outcomes more intuitive as the implications of votes cast using Scotland’s dual-ballot AMS/MMP system can be confusing, since the apportionment of regional list seats is dependent in part on the number of constituencies each party has won in that area. The site lets anyone play around with the numbers to better visualise which seats can change and why.
The 73 constituency seats are counted first and are awarded to whichever party has the most votes – they don’t need to win a majority. List seats are decided afterwards and try to ensure that parties which struggled to win constituency seats are proportionally represented across 8 broad regions – 7 of these seats are available in each.
Preset Example Scenarios
The website lets users play around with hypothetical scenarios or choose between some preset examples. Currently the presets include the 2021 results (on the new boundaries), 2024 vote shares and a ‘Nowcast’ which is a simple average of recent opinion polls. The Nowcast is not weighted in any way and is not intended to represent a definitive prediction. If the polls are wrong this will be also wrong – so definitely don’t bet based on it!
The Nowcast on 21/05/25
We’ve also included hypothetical scenarios in which people of pro- and anti-independence persuasions strategically split their ballot to maximise the representation of their side of the constitutional divide. These maximally efficient pro-union and pro-independence scenarios are highly unlikely to happen in reality, but they are included to illustrate the potential power of ‘split-ticket’ voting.
Scottish Methodology
The Scottish model is based on the new constituency boundaries proposed this month by the Scottish Boundary Commission. The model uses hypothetical ‘notional’ vote shares for each constituency which were created by Ballot Box Scotland to make a baseline understanding of which parties are more popular in each constituency.
The projection models the next election using adjusted proportional swing from those 2021 notionals. If a user doubles the vote shares of a party, let’s say the Lib Dems, it doubles their vote share across Scotland – but not by exactly double everywhere because that would give them over 120% in Orkney! The system adjusts down their increase in places like Orkney which requires giving them a larger increase in other areas.
The Greens, Alba and Reform are all included in every constituency – if any of them announce that they are not standing in a seat we will amend the projection match real-life. If they didn’t stand a candidate last time they are projected out of the ‘leftover’ vote share – basically add up the parties that did stand and subtract that from 100%. This is then readjusted nationally and within each region to make more realistic vote shares for each which match the user’s input.
The projection will work with numbers totalling 95-100% on both ballots, but the closer the user input is to 100% the more accurate it will be. There are some very, very unlikely scenarios that it struggles to deal with and if you find something especially strange just let us know: devolvedelections@gmail.com
Happy forecasting!