When deciding on how to shortlist projects, our northstar should be collusion resistance: preventing wealthy individuals & their cartels from self-upvoting or bribing their community to vote for them en masse.
Vitalik’s 2019 article on this topic was really interesting; i’m adapting the mathematical formulas he wrote to fit Octants use case.
-
Wealthy project owner & their community acquire n GLM tokens
-
each allocation window lets them allocate k to their own project (k = reward from staking n GLM)
-
these give them a reward of k x q (q being the multiplier)
The public goods funding system thus becomes one where projects simply get an interest rate of k x q on their n holdings
Octant solves the collusion issue in a way i haven’t seen other mechanisms use: careful curation of only 24 participating projects, which are mostly well-intentioned and follow the spirit of the system.
But how do we create a system that doesn’t rely on the goodness of participating projects?
One way would be removing predictability of a project being able to participate in every single epoch, reducing incentive for system capture
This approach (no project gets to participate in all epochs for a given year) actually fits in well with this user feedback; if you’ve fulfilled your funding requirements for a year in a single epoch, there’s no reason to re-apply in 3 months.
We can optionally couple this with theme based epochs to make things more interesting for GLM stakers: careful curation of projects within a theme so as a community we collectively learn everything there is to know about that part of the crypto ecosystem during an epoch.
Based on our discord chat I know some of these experiments are going on with mini-rounds, I hope the lessons from there can help us out.
This lengthy preamble was to only underline how important the selection procedure of participating projects is, as curating only 20-24 projects per epoch is octant’s unique answer to the collusion issue laid out by Vitalik, which is prevalent in any decentralized public good funding system.
His 2017 article on blockchain governance advocates multifactorial consensus where ultimate decision depends on the collective result of a plurality of mechanisms. For octant, the different groups at hand are;
- Octant team core contributors
- Coin holder votes (currently in use for epoch 2 selection)
- user votes with some form of sybil resistance
- norms, as they develop over time
- proposed governance council
I don’t yet have an answer on how these can all fit together to ensure a good roster of projects in an epoch. Tightly coupled coin-voting as we are currently doing on snapshot is probably not a good idea in future rounds due to relatively low participation in relation to the larger eligible pool, voter fatigue, unequal wealth distribution among holders drastically swinging votes and bribery attacks.