We’ve been having many brainstorming sessions here at the Funding the Commons residency around Octants mechanism, with @revmiller & @artem
Something we were discussing is how epoch rounds are more resilient than the snapshot one, since epochs have quadratic voting while snapshot has one wallet one vote.
Since octant has removed minimum thresholds, projects are scrambling to try and simply clear the eligibility bar to get some money.
In epoch 5 snapshot, just 8 votes are enough to get you into the round. This creates unhealthy dynamics where dishonest projects who farm 5-6 wallets can be assued of entry into the epoch.
Snapshot has a quadratic voting mechanism, so we were wondering why this is not the strategy for eligibility. This would reduce incentive to farm wallets because its not one wallet one wallet, but also weighted by how much glm the wallet has locked into it.
I’ve been thinking about this more. How would QF in snapshot solve for this issue?
Can you share why you think this? The reason we have sybil defense in our rounds in the first place is because there is a direct incentive to farm wallets.
The only thing I think your proposal would change, is that it would give whales more voting power. Am I missing something here?
Yeah i should rephrase, it would reduce the incentive to farm but not eliminate it.
Right now one eligible wallet one vote is crazy high incentive to farm
Whereas weighing it by amount kept in the wallet would partially reduce the incentive
The prior question we should try answering is: how do we give experts more weight in decision making?
The easiest way is capital: those with more money are more of an expert than those with less. While obviously flawed, it is sometimes better than a direct democracy.
The harder way is reputation based: those who have onchain history get more voting power than those without. This runs into the sybil issue pretty quickly, but imo its not an issue if we weight voting power by amount (quadratic), reputation (onchain history) and diversity of projects (COCM) they support (the new gitcoin approach).
Snapshot is harder obviously to implement this approach, but i do feel like the current approach of 1 wallet 1 vote is horribly broken and favoring projects that are dishonestly trying to get as many eligible wallets as possible. Since even 8 votes is enough to get you into the epoch and receive some amount of funds.
I am not very familiar with your mechanisms, and this might not directly address your problem.
I have two observations though.
1. Separate validation, qualification, and weighting.
You vote for weighting, but I would suggest there are three components to consider to get funding. Validiation is about meeting the criteria to be a valid project (which you have, but might want to consider if there is a missing component). For a validated project, it then needs qualification, and that might be the missing component here. Here that might involve reaching a threshold of support (community votes) of being a worthwhile project. The third step is weighting, of projects that are validated and qualify, they can get a share in proportion to weight.
2. Allocation based on percentile weight.
I am not sure this is relevant here and will not go into depth. I will start from the observation though that while quadritic voting is an interesting mechanism that helps to solve some scenarios where other voting systems can lead to perverse outcomes, and there are no doubt valid applications of it, I think that in general for many scenarios it is a bad voting paradigm.
An alternative that can work for some scenarios is that voters can specificy how much of the pool they think should go to various projects. This can combine to provide both a threshold level of support (qualification), and how much of the pool (weight). I will not go deep into mechanisms here unless there is interest.
By using a percentile measure rather than a direct weight, voters are not incentivised to vote for more than they think is appropriate hoping to average out, and it is a key part of such a mechanism to get broad support rather than local spikes of support.
You will need voters to allocate to a number of projects rather than their favourite. This can be a disadvantage if it is hard to get that level of participation, but could work well if there is good participation (and encourage a better form of participation judging across projects).
Thanks Rem, I appreciate your message. You’ve caught my interest! Just to be clear though, @devanshmehta is talking about a stage of our process before the actual grant round with QF.
If we are indeed talking about the QF grants rounds that we run. the problem I see with what you are proposing is the way that Octant is setup. Reward streams are split into individual rewards that a user earns for locking GLM in Octant, and then the matching pool. The matching pool is allocated based upon QF which is a function of the donations from individual rewards.
I’m not clear on how your idea would work in our setup, because everyone has different amounts of rewards to potentially donate, so choosing how much of the pool should go to each project seems to be in conflict with our setup, the difference being nobody is voting, they are donating.
Okay, so you were talking about getting into a round, and the donations can only be made if they have been cleared to be in the round?
I was talking about both being cleared and allocating from the matching pool. If it is applicable here is not clear to me. Maybe not. I think it would require a change in paradigm, although I suspect it could work.
I am not convinced by the premises of quadratic funding in practice. If you are convinced by them, you probably lack the motivation to change.
A core reason I am not convinced is that once it becomes an established norm that lots of people is the way to unlock value, then people will be farmed at scale. Currently you consider it meaningful diversity if a hundred people act a certain way, but you are creating a system where that will no longer true and million can used. Actual independence will be swamped by sheer numbers of real used people.
I think a more meaningful measure is actual locked tokens. It is a temporary blip that the number of people mitigates the bad outcomes you seek to avoid.
We have changed a fairly large, to very large, aspect of our round nearly every round since we began in 2023. So we are definitely willing to change if it makes sense
I don’t argue this.
This was actually our first mechanism. But the problem lies within the imbalance of token holders. We had a whale come in at the end of a round with 90 minutes left, and essentially “flipped the table upside down.” Whereby donating a massive amount of value to a few projects, essentially left many projects that were going to receive very supportive funding before that 1 vote, now out left to dry. Why would you want to participate in something that you hold no weight in decision making processes?
I know, although I did not know why you changed it.
I can see why that would dissapoint. I do not know the scale here, but when you say ‘whale’, is this more like a big fish in a little pond?
Is the problem that the pond is so small that it does not take much to distort it?
Having ‘no weight’ is an exageration though? At scale, you are expecting individual participants to have very little weight. A large proportion of us vote in elections, despite our individual weight being questionable.
Conversely, why would you want to contribute a lot if you do not get weight proportional to your contribution?
I suspect, particularly if you consider the scale you seek, QF has more of a negative impact on the second than a positive impact on the first.
You only need influence in proportion to what you contribute, that is fair. Not getting even that is discouraging. You could well have an unintended consequence of keeping the pond small.
Because of the event mentioned above. Most of the community was not happy with it.
10s of millions of dollars locked into the protocol where the average user has a few hundred. I’d say a whale, for our pond at least.
It’s not. When you have a user that has the ability to donate 45,000 dollars to the average users donation of 2-10 dollars, you can see the disparity. Also, plutocracy’s in general are highly frowned upon. Why do we want to repeat the same old systems that give weight to the rich?
I’m not advocating that QF is the best mechanism out there, but for right now with the tools we have, its a better mechanism than plutocracy if we value the community who cares about funding public goods. The majority of them don’t have capital, so we would be boxing them out, and thus our highest engaging userbase.
