202510151147 - convincing-voting
Main Topic
Q: What is conviction voting, and why would a DAO use it instead of one-off token voting?
Conviction voting is a continuous allocation mechanism where supporters stake voting power on proposals over time. The longer stake stays allocated to a proposal, the more “conviction” accumulates (usually with a decay model so conviction grows toward a limit). A proposal passes once its accumulated conviction exceeds a threshold.
The core design goal is to let participants express strength of preference without requiring synchronous, one-shot elections. In typical token voting, every decision is a discrete event, and participants must show up at the right time. Conviction voting instead turns governance into a background process: people signal ongoing support, and proposals that gather sustained support eventually pass.
A useful intuition is the “slow cooker” model:
- You allocate tokens to a proposal (turn on heat).
- Over time, conviction increases as long as tokens remain staked.
- If you remove tokens, conviction decays (the pot cools).
- Bigger asks (more treasury spend) can require higher conviction thresholds, so they need broader and/or longer support.
This mechanism is often used for allocating treasury funds among many competing proposals, where you want continuous prioritization rather than winner-take-all voting rounds.
🌲 Branching Questions
Q: How is conviction typically computed (conceptually), and what parameters matter?
Most implementations use a time-weighted function with decay, so conviction at time t depends on prior conviction and current staked amount. The practical reason for decay is to avoid “infinite accumulation” and to make it possible for support to fade if attention shifts.
Key parameters:
- Decay rate: how quickly old support fades after tokens move away.
- Time constant / half-life: controls how long it takes for conviction to build.
- Threshold function: how much conviction is required to pass; often depends on requested funding size and available treasury.
- Minimum stake / spam resistance: prevents many tiny proposals from cluttering the system.
The parameters encode political values: responsiveness versus stability, and how expensive it is to change direction.
Q: What are the main benefits and failure modes in practice?
Benefits:
- Reduced coordination overhead: no need for everyone to vote within a fixed window.
- Better expression of preference intensity: sustained support is rewarded.
- Natural prioritization: proposals compete for a limited pool of staked support.
Failure modes and risks:
- Voter apathy can still dominate: if few people stake at all, outcomes can be skewed.
- Whales can dominate if token distribution is concentrated.
- Parameter tuning is hard: thresholds that are too low create “easy pass” dynamics; too high stalls everything.
- UI/UX complexity: participants must understand continuous staking, which can be harder than snapshot voting.
Conviction voting works best when the DAO has good proposal curation, clear treasury policies, and a simple interface for stake allocation.
Q: When should a DAO prefer conviction voting vs. quadratic voting or snapshot-style voting?
Conviction voting is a good fit when decisions are about ongoing prioritization and funding allocation across multiple proposals. Snapshot-style voting is better when decisions are discrete, high-stakes, and need explicit legitimacy at a specific time (constitutional changes, parameter upgrades).
Quadratic voting can express preference intensity in a single round, but it introduces identity / sybil assumptions (or requires strong anti-sybil measures) and different game dynamics. Conviction voting instead assumes token-weighted power but uses time as an additional signal.
A common pattern is hybrid governance:
- Conviction voting for routine treasury allocations.
- One-off governance votes for protocol upgrades and policy decisions.
References
- https://github.com/1Hive/conviction-voting-cadcad
- https://1hive.org/ (project context; references pending for a canonical paper-style spec)