Should Manifold Markets implement this feature: Propose a question for someone else to judge
2
Ṁ7
Feb 1
49%
chance
The way this feature would work is that I create a market, but instead of judging it myself, I want someone else to judge it. Once the market is created, it would work sort of like a pull-request, in that it remains in-active until and unless the proposed judge accepts it. (If the judge rejects it, it just gets deleted). There would probably need to a mechanism to prevent a malicious user from spamming people they doesn't like with spurious or obscene questions. One option is to allow judges to block users from proposing questions, but I think a better option would be to require question proposers to use M$ in order to propose a question. This M$ could be treated as a bond, such that they get it back if the judge accepts the question, but they lose it if the judge does not. I'm not sure who should get the reward for resolving the market. The simple / easy options are that it goes to the judge, or gets split between the judge and the proposer. A more complex variation is that the proposer and judge can agree on the split as part of the process for accepting the question. The rationale for this feature is that I might have markets that I want to create which I don't think I'm qualified to judge, or which I don't have enough credibility to judge well (that is, people might not trust me to resolve it properly). But I know of a user who I believe is qualified to judge it, or who has a lot more credibility. I could ask them to make the market, but I can set up the market for them, they're likelier to say yet. Resolution: I commit to resolving this question according to whichever option gets the most support (i.e., "no" if %chance is less than 50, "yes" if % chance is greater than 50, and "N/A" if % chance is exactly 50). See here for a request to make this kind of thing a feature: https://manifold.markets/JiaobeiMandos/should-manifold-markets-implement-t) #ManifoldMarkets #FeatureRequest
Get Ṁ1,000 play money
3 Comments
Sort by:
My thought process on the resolution commitment was as follows: 1. I think this feature request is a good idea. 2. It's obvious that I think this is a good idea, since I wouldn't have made the market if I didn't want the feature to be adopted. (It occurs to me now that my being in favour of this feature maybe wasn't obvious, and maybe I should make a market for a bad feature just to see what will happen). 3. If everyone can predict how I'll resolve the market, the incentive will be to make trades in the same direction, which will distort the result. 4. Committing not to do that gives people more incentive vote according to their actual opinion, or to their best guess of what the community as a whole would favour.
Object level: I'm not sure this would be worth the additional complexity in the short term (next month or so), mostly because the core desire of "ask a question and have someone else resolve" is achievable without in-platform support. You could just DM the market creator and suggest that they set up the market for you. (We might need to support DMs soon...) Of course, we should beware trivial inconveniences (https://www.lesswrong.com/posts/reitXJgJXFzKpdKyd/beware-trivial-inconveniences), and I'm curious to hear from other users about whether this feature would be exciting for them!
Meta level: This is a super interesting approach to a market! One of my long-term goals behind Manifold is to explore futarchy, or decisionmaking via prediction markets. So far, though, futarchy has always been phrased using multiple conditional questions, e.g. "If Manifold allows resolution delegation, will they have 1000 users in Feb?" vs "If Manifold supports crypto, will they have 1000 users in Feb?" What you're doing here is just allowing people to directly vote with their dollars. I'm not sure that resolving to "highest % option" is the best idea -- does it introduce weird dynamics where you vote according to what's popular instead of what you personally prefer? You're only financially incentivized to bet once you can swing the probability down to 50%... Maybe you should commit to resolve to N/A instead (so in effect locking in your money as a way of voicing support for an outcome)? I really have no idea, this is such a cool experiment! Another loosely related thought: For this kind of market, maybe quadratic funding would be more appropriate than out standard dynamic parimutuel (https://vitalik.ca/general/2019/12/07/quadratic.html)?

Related questions