In the upcoming UK 2024 General Election, there are 650 seats in contention. This prediction market revolves around the accuracy of Manifold's market predictions for each of these seats. We're interested in finding out how many of these seat predictions Manifold will get right.
Criteria for a Correct Prediction:
- A prediction is deemed correct if the party or independent individual with the highest YES percentage in the Manifold market for a particular seat is the actual winner of that seat.
- No minimum YES percentage threshold is required for a prediction to be valid.
Handling Ties:
- In the event of a tie for the highest YES% in the Manifold market for a seat, if one of the tied options is the actual winner, this will count as 0.5 points towards the total correct predictions.
- Similarly, a tie in the actual election results for a seat, regardless of how it is resolved, will also count as 0.5 points.
Resolution Method:
- At approximately 10pm on the election day, I will collect data from each market within the group UK 2024 General Election Constituency Winning Party. The party or individual with the highest YES percentage at this time will be considered the predicted winner for each seat.
- The initial election results, as declared by the returning officer and broadcast on major UK media organizations, will be used for resolving the market.
Resolution Criteria:
- The final score will be calculated as follows: 650 (total seats) minus the number of incorrect predictions. Tied calls, both in the market predictions and actual results, will be valued at 0.5 points each.
- Each option in this market will resolve as YES if the final tally aligns with the statement: 650 - incorrect_calls - ROUNDUP((tied_calls / 2), 0) equals or exceeds the specified number in the option.
Related questions
🏅 Top traders
| # | Name | Total profit |
|---|---|---|
| 1 | Ṁ82 | |
| 2 | Ṁ20 | |
| 3 | Ṁ17 | |
| 4 | Ṁ14 | |
| 5 | Ṁ2 |
@Fion I’m going to try and do it as close to bang on 10pm as I can but making no promises. Dependent on when the date ends up being I might live stream the whole evening on twitch.
@Noit would there be an argument in favour of doing it at like 5 minutes to 10 just to be sure? I don't think there's going to be any relevant last minute predicting in the last five minutes, but if you accidentally close after the exit poll, there is a risk you capture a bit of "post-diction" that we probably don't want to include.