Try a scenario for your area and see how far past elections say it could be wrong
This is a scenario tool, not a forecast. It works on general-election results (2013–2024) for places, not people. It shows how an area leaned toward MDP against the Yameen/Muizzu bloc (PPM/PNC) and how that could move if the national picture moves. It says nothing about who will vote in a party election such as the MDP Qaumee Majlis, and it knows nothing about individual voters.
Scenario
Saved areas come from the Campaign Area page (2024 constituency lines).
MDP's share of votes for MDP and PPM/PNC combined. This is the big unknown: it has moved 7–12 points between elections.
Changes how many votes the area casts, not how they split.
Points added to MDP's share in this area only. A hunch, not a measurement. If you have a poll, use the box below instead of this slider.
Your own poll of this area
Count only people who picked one of the two sides. The poll is blended with the model, each weighted by how reliable it is. A poll through your own contacts counts for much less, because the people you reach lean your way.
How wrong has this been?
Free for everyone. How accurate it has been is public; the interactive projections are free with an account. The method itself is not published.
We pretended each past election had not happened yet, gave the model only the elections before it, and told it the true national result (the number you set with the slider). Then we compared its place-by-place guess with what happened.
What it can't tell you
Party electorates. The MDP Qaumee Majlis vote is decided by party members, not by general-election voters. Treat these results as a picture of where MDP-leaning voters live, not as a prediction of that vote.
Turnout as a percentage. The public spreadsheets give votes cast per place but not how many people were registered there, so this models how many votes, not what percentage turned out.
Sentiment. There is no public polling in the Maldives to learn from. Your own poll goes in the "Your own poll" box, where it is blended with the model by reliability. The blend treats the poll as a picture of general-election voters in the area, so a poll of party members says something different.
Candidates. It ignores who is standing. In 2024 parliamentary seats the error is the largest (about 7 points) because local candidates matter more.
Independents and other parties are left out of the share: it compares MDP with PPM/PNC only.
Seat-level differences. How a constituency leans compared with the island or city it votes in did not carry over from 2019 to 2024 (a correlation of 0.07 across 70 paired seats), so the projection barely applies it. An area that is only part of a place, such as one Malé constituency, inherits that place's lean and carries an extra error of about 8 points.
Pooled places. Malé's wards, Hulhumalé, Vilimalé, overseas, resorts and jails are pooled because their ballot-box names change from one election to the next.