CFM - design updates

Governance Designs
1

In this thread, we will regularly post updates to Butter’s Conditional Funding Markets (CFM) design.

2024-10-09 CFM update

Recap of discussions

In https://ggresear.ch/t/conditional-funding-markets/27:

  • distbit triggered a discussion on the optimal reward versus liquidity budget allocation of the DAO. We acknowledged that the mechanism looks like a composition of two mechanisms, curation followed by cfm with curation needed because of limitations on liquidity. At the moment we can’t expect to reverse-engineer the right allocation from the definition of the two sub-mechanisms. So the best shot would be trying different curation mechanisms, and think about, in the permissionless case (which is the goal), making parameters adapt to market conditions through a feedback mechanism.
  • distbit also suggested bid/funding-ordering as a better proxy for project success. An identified limitation is that this doesn’t account for the fact that the existing auction mechanism incentivizes bidders to bid more if their project is more predictable.
  • We mentioned Sebastien Zany’s remark on markets degrading into markets for illicit forms of sabotage. We pointed out that a typical scenario would require enough coordination, making it relatively improbable. We also remarked that legal forms of sabotage have a higher chance of being an issue, but that complementary incentives to align internal contributors can be derived.

Review with Robin Hanson:

  • Play money markets: if we want to use play money instead of real money, we must make sure play money is considered valuable to all potential traders, e.g., the value of social rewards is limited to traders who intend to remain in the ecosystem.
  • Market mechanism: LMSR is not more capital efficient if outcome tokens correspond to independent predictions.
  • Sabotage: this can be counteracted by making sure project contributors have positive stakes (can be produced by the project itself) and can’t hold negative positions (which would require identifying these actors, which is hard to enforce).
  • Incentivize trader participation: what about incentivizing project leaders (submitters) to trade in other projects markets.

Updates

Answers:

  • short term: which AMM to use?
    • for now, CPMM as we don’t expect to profit from LMSR combinatorial nature
      might be re-assessed
  • short term: how to incentivize trading?
    • incentivize project leaders to trade in other projects’ markets
  • long term: which auction mechanism to use?
    • it looks like we might want to try different ordering of winners, like bid/funding

Questions:

  • How to design the optimal auction mechanism?
    • we would ideally like a closed-form solution: is this achievable?
    • can we otherwise design a feedback mechanism that updates parameters dynamically?
      • can be first governed by admin keys (centralized)
      • then made governed by an automated feedback mechanism
  • Can we use alternative market mechanisms like Ordered Prediction Markets Ordered Prediction Markets - Pepe Blog (suggested by Sebastien Zany)?
    • appears valuable in terms of UX
    • the author claims this would make liquidity better utilized - to validate
  • How to leverage incentives to project leaders to have them participate in other projects’ markets?
    • sounds valuable as project leaders have actual knowledge about metrics and projects
    • might be worth updating the mechanism.

Upgrade

Version Bump: none → v0.1.2.

Changelog: ø.

1 Like