Conviction voting

How conviction accumulates

Voters continuously express preference by staking tokens behind proposals they want approved, and the weight of that vote grows over time. Collective conviction accumulates until it reaches a threshold the proposal specifies according to the funds requested, at which point it passes and the funds are released.¹ The reference model describes the behavior as a capacitor charging up, with an alpha parameter setting the half life of the decay when support is withdrawn.¹

Two consequences fall out immediately. Participants can weigh in at any time, so no consensus has to be assembled per proposal and the quorum bottleneck disappears.¹ And stake is rival: tokens supporting one proposal are not supporting another, so backing five proposals splits your conviction across all five. A yes or no ballot does not force that choice. Conviction voting does, which makes it a prioritization mechanism as much as a decision one.

The deployment we can confirm

1Hive has been developing conviction voting contracts with BlockScience and the Commons Stack since early 2019, and runs a DAO live on the xDai network that uses its native token Honey to allocate funds to proposals through the mechanism.¹ 1Hive's own documentation describes the same thing from the other side: Honey funds a common pool, and public goods that grow the 1Hive economy are collectively funded with conviction voting.²

That is the record we could verify. Other deployments appear in ecosystem writing, but we found no primary documentation with dates for them, so this page names one live implementation rather than implying a category. That thinness is itself information: this is a mechanism with a small evidence base, not a widely stress-tested standard.

What it buys you

The stated case is that continuous, time-weighted support mitigates control by majority holders, last-minute vote swings, and the low participation that comes from short voting windows.³ The middle item is the strongest. A vote whose weight is a function of duration cannot be manufactured inside one transaction, which takes borrowed-capital governance attacks off the table without a snapshot rule or a flash-loan check.

The second thing it buys is a live signal instead of a periodic one. At any moment you can read how much conviction sits behind every open request and how that has moved, which is a richer picture of what a community wants than a series of pass or fail results with weeks of silence between them.

What it costs

Speed, first. A mechanism built so nothing passes quickly cannot be used for anything that has to pass quickly. Conviction voting is unsuited to security patches, emergency parameter changes and contract upgrades, all of which still need a conventional path with a timelock behind it. Adopting it does not retire your governor. It sits alongside one.

Then the parameters. The alpha decay and the shape of the threshold curve determine everything about how the system behaves, and no setting is correct across treasuries.¹ Set decay too slow and stale support keeps passing proposals nobody currently wants. Set the threshold curve too flat and large requests clear as easily as small ones. Simulate both against your own proposal volume before launch, which is why the reference implementation ships as a simulation model rather than only as contracts.

And be clear about what time-weighting does not fix. It blunts a large position arriving late. It does nothing about a large position that arrives early and waits. A patient holder with size still decides outcomes.

Where it actually fits

Continuous allocation from a common pool, not binary protocol decisions. The natural use is a treasury with a steady flow of funding requests, where the real question is which of forty proposals deserve money this quarter and in what order, and where waiting a few weeks costs nothing. That is the shape of the confirmed deployment.²

Which points at the underlying business question. Conviction voting distributes a pool that already exists. It does not create the revenue filling that pool, and a community carefully prioritizing an emptying treasury is running a good process toward a bad end. Get the funding source right first, then pick the allocation mechanism that suits it.