Resolves early to market prob if the price has been consistently trading within a 6%-wide band for one week.
If it does not resolve early, it resolves to whether AI kills all humans by 2100.
The amount of mana someone is willing to bet on this represents combination of their confidence on this Q, their track record of making good predictions on other Qs, and their expectations of the first two vis a vis all the other traders. So it seems like a better way of aggregating the actual epistemic status of the crowd than either the market that only resolves if we survive, or the market where all the trades get rolled back at some future date.
@Shelvacu I think a pure self resolving market with these rules would stabilise at 50%. So if this one stabilises at other value, that might be because a bias towards actual peoples' predictions.
@roma Nah in practice this type of market would resolve far from 50% e.g https://manifold.markets/JonathanRay/will-gravity-still-exist-on-jan-1-2?r=Sm9uYXRoYW5SYXk
@JonathanRay Also see this one with the same self resolving structure and 33 traders which resolved 35% https://manifold.markets/JonathanRay/is-it-bad-manners-to-bet-against-yo?r=Sm9uYXRoYW5SYXk
@JonathanRay Yeah, but these are not pure in the sense I meant. I was thinking of a market that has no meaning attached, like "this resolves to current % after it stabilises" and that's it.
@JonathanRay But I think we agree. I also think that when there's a meaning attached it is not going to be 50%. But still, how well this measures actual opinions, not sure.