conviction voting

Conviction voting replaces the snapshot election with something more like a slow, continuous tug-of-war. There is no fixed voting day; instead, supporters stake their tokens on the proposals they favour, and the weight behind a proposal builds up the longer that support is held in place. A proposal passes only when its accumulated conviction crosses a threshold. The effect is that fleeting, last-minute, or borrowed support counts for almost nothing — to move the organization, you must keep your tokens committed to a cause over real time, which is hard to fake and hard to flash-loan.

Mechanically, each proposal has a conviction value that updates every block as a decaying accumulator: new conviction equals old conviction times a decay factor alpha (a number just under 1) plus the tokens currently staked. Stake tokens and leave them, and conviction rises asymptotically toward a ceiling; remove them, and it decays away. The pass threshold is not fixed either — it scales with the share of the treasury a proposal requests, so a request for 1% of funds clears easily while a request for 50% demands overwhelming, sustained backing. Voters can also stake on several proposals at once and reallocate at any time.

The design targets specific failures of one-shot token voting. It is naturally resistant to flash-loan and bribery attacks because conviction must be earned over many blocks, not bought in one transaction. It removes the artificial deadline that forces rushed decisions, letting funding flow continuously to whatever the community sustainably supports. The trade-offs are latency (good proposals take time to build conviction) and parameter sensitivity (alpha and the threshold curve must be tuned, or the system feels either sluggish or twitchy). It is used by communities such as 1Hive and tooling from the Commons Stack and Aragon.