Should Manifold Markets implement this feature: Resolution Bonds
2
Ṁ36
Feb 1
64%
chance
This feature would allow market creators to post a bond when creating a market. If the market it resolved poorly (more on that later), the market creator loses the bond. The rationale for this feature is that, as a market creator, I want to encourage people to participate in my markets by showing that I'll resolve it honestly. Or, to look at it from the trader's perspective, I might see a very lopsided market where I believe I can make a lot of M$, but only if the market creator resolve the market honestly. I think reputation is the first line of defense against "bad" market resolutions. If you resolve a market in a way that is obviously wrong, or self-serving, you'll lose your reputation and people will stop participating in your markets. I see the bond idea as a complement to this, because it allows a market to creator to pledge M$ in addition to their reputation. It also allows potential market creators to "bootstrap" their reputation: to start with, they have no reputation (because they're unknown), so they can post large bonds with their markets. After they've resolved many markets, they'll develop a good reputation, and that will presumably keep them honest. The obvious thorny problem, and largest barrier I see to implementing this feature, is how to judge when a market has been resolved poorly. I see two options so far: 1. Allow market participants to judge the resolution after the market has closed. Only allowing people who had participated in the market before it closed would be allowed to vote, which would somewhat mitigate the brigading problem (e.g., if a market is resolved against you, you ask your friends to vote against the resolution). There would still be a risk that any market with a bond is always resolved in favour of the majority opinion (because they can outvote the market creator) 2. The bond specifies who the appellate judge is. That is, if market participants contest a resolution after the fact, there is a pre-selected user who will be asked to decide if the market was resolved fairly or not 3. Some kind of market-based mechanism, where contesting a market creates a new market to judge whether or no the first market was resolved fairly. See here for a feature request for proposing questions for someone else to judge: https://manifold.markets/JiaobeiMandos/should-manifold-markets-implement-t-d5da24049b2ee 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:
Keep in mind that if your market is flexible enough that other weird side-solutions will naturally develop. For instance, nothings tops me from starting "insurance" bets on other bets (in fact we've seen a few already). My vote is always to keep things as simple as possible and just admit where we're relying on trust, so that people can make informed decisions about whose trust to accept, rather than pretending we can eliminate the need for it entirely. But it's also interesting to see what systems are organically developing just from the natural capabilities of being able to make bets on things. Derivative bets could open up a host of things for better and for worse.
Interesting! We've been thinking on the problem of "badly resolved markets", and one of the proposals was about delegating to another judge (your option 2.). Another was a market based mechanism (option 3.): if Jiaobei is unhappy with how Austin resolved "Will the sun come up", then Jiaobei "sues" Austin by setting up a market "Did Austin fairly resolve 'Will the sun come up'" and commit some amount of money to NO; and the market gets decided by the judge in 2. James views many proposals about smoothening the resolution process as convoluted or unnecessary; I'm curious if a bonds-based approach will appeal to him!
I think this is a great idea and will become necessary as the platform scales and we start seeing more activity from bots and brigades. I personally am not averse to having Manifold Markets calling balls and strikes; I think most human systems run on trust and there's nothing wrong with that so long as it's properly earned, but others might have a different opinion. Sometimes if you try to design a system of checks and balances that's too convoluted in its operation, you just introduce more vulnerabilities rather than less. Just as big part of establishing trust in a system is not just being assured that the process is fair, but also that people can understand what the process is and how it works.

Related questions