This market resolves YES if someone starts running a Manifold plays chess (or another game) series of markets based on using conditional markets to pick moves, as described below, before June 2023. I intend to be pretty broad with what counts here.

Background

Manifold plays chess 1 chose each move with a prediction market that predicted which move would be chosen - this is just a Keynesian beauty contest (predict what other people will predict) and failed because someone forced a blunder and made a profit on a big bet that Manifold would lose.

Manifold plays chess 2 chose each move by having people vote on it, weighted by how many YES shares they held in Manifold winning the game. Think of it like people can buy (or short) shares in "Manifold chess corp", and the YES shareholders vote their shares on what to do. This is still playing out, but there are some obvious flaws that have already been seen.

Neither of those really leverages the true power of prediction markets to make decisions.

Proposal

If I were designing a way for Manifold to play and win chess with prediction markets (aka a futarchy), my first instinct would be to use conditional markets: for each move decision to be a set of binary markets, each one being "If we play move X, will we win?" Then you pick the move with highest probability, and resolve at the end of the game.

I think this sort of approach works reasonably well (it's still exploitable, but harder).

One flaw with this very initial design is that people's mana will stay locked up for the entire duration of the game - it's very capital inefficient. (Note, margin or loans aren't a suitable solution to that because those markets are all 100% correlated - they all resolve the same based on whether we win.) To address that, you could try to strike a balance by evaluating the position after N moves. Examples could be, after N moves resolve to some simple evaluation of the board position (say, will we gain/lose material, or the evaluation of a weaker stockfish), or to manifold's evaluation of the win probability from another market.

So here's my ruleset proposal I currently have in mind for a futarchical Manifold plays chess: For each move, we generate 5 candidate moves by some method (e.g. ask a weak-stockfish to pick some moves, or by popular vote). We create a market for each: If this move is chosen, what will be the win probability after 5 moves, as evaluated by weak-stockfish (or, as evaluated by a prediction market)? The one with the top probability is chosen (the others resolve N/A). Of course, there are lots of variations on this theme that you could imagine - e.g. you could resolve to Manifold's predicted win probability instead of weak-stockfishes, or let Manifold generate the proposed moves.

The markets would probably benefit from typical anti-snipe rules - e.g. use the average probability over the last 5 hours, or allow the market to extend if there's excessive volatility just before closing.

This models real-world scenarios where you have some imperfect metrics to give an evaluation of the state of the world, and some method to suggest proposals that may or may not be good, and then you use prediction markets to evaluate how good each proposal is - i.e. a futarchy

I don't intend to run this myself because I'm not that familiar with chess. But I'd also be interested in Manifold playing another game that doesn't have powerful computer engines. That might be a more realistic model of real-world approaches to futarchy design.