One problem with prediction markets is that often things turn out to be an edge-case that doesn't fit neatly into the original resolution criteria. The way to resolve those edge-cases could often be clarified ahead of time, but often this doesn't sufficiently happen.
Recently I've been thinking: imagine you come up with an edge-case and ask the creator for clarification. If the creator first tells you about it, and gives you time to trade on the info, then that would give you profit, and thereby incentivize you to bring likely edge-cases to the creators' attention. Then after you've traded, you could reveal the creator's answer to cash out.
But the obvious counterargument is this could be disadvantageous to other participants, making it feel unfair. So I think it should be taken up for vote and discussion to see how people feel about it.
Maybe there are also other pros/cons that I haven't thought of.
1,000
3.00I would take a more collectivist approach and attempt to incentivize information aggregation in the comments, paying individuals for clarifying market edge cases. Insider trading doesn't incentivize information sharing. I don't get a lot out of just seeing, "line go up," on Manifold, I really do not care about that. I care about learning new things, which helps me achieve other objectives in life. However I do understand that most people do care very much about line go up so I may have an anti-perspective on what's needed.
@Joshua Last time I did that the creator traded on it before answering me ๐. Lesson learned: bet first, ask questions later.
There are a bunch of different kinds of edge cases
"someone bets on the market and then takes illegitimate action to push the market to resolve that way" - the money dupe market, and a few more I mention here. There shouldn't be any incentive to do that at all, markets should just resolve as if the manipulation didn't happen, or if that's impossible manifold should punish the person who did the manipulation.
"Clearly not what they meant" - saying a market about if "twitter.com" has >100M users resolves no because it's X now and the grand doge decided to entirely take down the website "twitter.com" for whatever reason. In this case, there should just be a universal norm of not doing that and not resolving markets based on that.
"Not what the title/description imply, but clearly implied by the specific resolution criteria" - Market called 'will Destiny get 500k subscribers before September', market description says 'At August 31 at 11:59PM UTC, I will go to youtube.com/destiny, and look at the number. YES if it is at or above 500k.' Someone notices on Aug 15 he has 500k, so bets a ton of yes, but then it goes back down. These just kinda suck, there's not much you can do about that, should probably just resolve to the more specific description in some cases (like that destiny market) and resolve to the general description in other cases (youtube.com was down from 11:50 to 12 so there wasn't a number of subscribers when he checked so it resolve NO even when it'd yes otherwise).
Genuine ambiguity - Market on how many DAUs twitter will have in Dec 2023. Twitter starts publishing two DAU measurements, one 'adjusted using proprietary algorithms but may undercount' and one using 'industry standard methods, unadjusted'. Which one do you use? Not much you can do here either.
(there's also "description says one thing, market creator clarified 100 comments down he means a different thing", which sucks)
I don't really see a case here where this kind of incentive is useful.
And market creators aren't really going to use whatever new edge cases they learn to protect future markets - click on any new market, they don't have any protections for edge-cases in their resolution criteria, even though all sorts of edge-cases have happened in the past, becuase people use manifold pretty casually and lots of different uncoordinated people create markets.
@jacksonpolack All of your examples were cases where there was an already quantified measure of the market outcome. Meanwhile, I more had open-ended markets in mind. Think e.g. these markets below, where there sorta is an objective answer but we can't just look it up as a number, and also think of the even more abstract markets where there is no objective answer.