[Phase-3] HIP-63 Quadratic Delegations on the DAO Snapshot

I don’t understand how this has to do with voter suppression in any way. If any, it follows the usual more democratic ways of representing both the opinion of the minorities and majorities in a group.

The proposal literally mentions it is a reaction to the burmese delegation.

He liked the idea. I also love the idea, but in practice I believe it will be net negative. He also mentions he doesn’t know if it will work in practice or not.

For example, here Santi mentions it is completely OK to vote with the keys of 7 family members:

That’s equivalent to someone that has 49 delegations.

If we move forward with this HIP, I believe we need to vote on strong sybil-resistance first to avoid the situations above, otherwise this HIP will punish honest players that delegate vs people that hold multiple wallets.

1 Like

I am against it:

  • Not aligned with 1 human 1 vote, as the delegator’s voting power is greatly reduced.
  • Not fool-proof: it can be worked around easily by finding more delegates who are aligned. That just makes UX worse for the same outcome, with extra steps.

Avsa’s idea of quadratic decay is much more appealing to me but I agree that it should be a different HIP. Also bearing in mind that quadratic decay does still infringe on the 1 human 1 vote idea but only as an incentive to renew the delegation choice, and decay should be calibrated to kick in only after an initial period of say 6 months.


I like this and voted for it, but I do think that square-root fall-off may be a little too steep.
voting power = (d)^(1/2) gives a marginal voting power of 0.4 at 2, 0.15 at 10, 0.05 at 100 and 0.015 at 1000.

There is nothing magical about 1/2 power (square root). We could also use a power closer to 1.
For instance:
voting power = (d)^(3/4) gives a marginal voting power of 0.7 at 2, 0.4 at 10, 0.24 at 100 and 0.13 at 1000.

I realize that this might be technically more difficult to implement, but I don’t see why we are getting so hung up on the square root function.

There’s a reason why Vitalik, Gitcoin, Glen Weyl, etc. insist on the quadratic formula. I don’t thing it’s wise to modify a very roadtested formula with improvisation that would break the credible neutrality held by square roots. Modifying the quadratic formula breaks the incentives and equilibrium of power that aims to make an environment where minorities can form while avoiding a tyranny of the majority and bipartidism. There’s a good paper here: [1409.0264] Nash Equilbria for Quadratic Voting

Delgations on Snapshot are an on-chain smart contract that informs the system who has delegated to whom. You cannot bypass or modify that smart contract in any way.

If you want to make people follow the way you vote with an alternative system, you still need to get everyone to sign a message with their private key for every new vote.

Signatures on Snapshot take as an input the space, the proposal ID and choice ID. These are mandatory in order to cast a vote … You cannot automate those actions since it requires that each voter signs a message with their private key… if you think about it, it’s really more convenient for you to ask people to vote directly rather than building a shadowy nonsense alternative delegator tool. Which is precisely what Quadratic Voting incentivizes among many other things: :clap: With it, direct votes weight more than delegated votes :clap: — which makes sense since no one believes the 130+ or so farmed delegations are coming from people really aware about our DAO.

Should be noted that 30+ more farmed delegations happened throughout this Phase 2 vote that just got approved.

Proposal is now on Phase 3 voting on Snapshot

1 Like

Votes from farmed accounts and votes from people that control the wallets of their friends and family weight more than delegated votes.

We should at least push for EXPLICIT sybil-resistance before passing this.

Ask the people here instead of accusing and basing a proposal on this.

1 Like

Glad to see you are not disagreeing with the fact that writing an alternative delegator is nonsense.

I agree that that problem you describe relates with explicit sybil resistance of the protocol itself more than governance rules. I suggest you check the recommendations @donosonaumczuk gave @green regarding that proposal in the governance channels in order to improve the writing on phase 3 for it.


It’s not because I didn’t mention that I agree with it. This proposal only adds an UX barrier, it doesn’t actually generate any of the effects promised in its text. And you know it.

One month ago:

1 Like

Yeah, you don’t seem to understand I am not against delegations. I am in favor of making them quadratic to prevent the effect of delegate whales that can hijack the governance vía clientelism or gerrymandering.

That plus all the other benefits described in the proposal, specially incentivizing minority voices over bipartidism.

Yeah, I thought that through and it turns out you can’t do it because each vote requires a private key signature with data from each proposal. For this proposal I actually paused and thought that through and realized Clement was utterly wrong about that.

I find it utterly funny you copied Clément’s opinion without doing your own research in the first place. :joy: There are many ways to automatize copy-voting, maybe you just haven’t thought this through enough?

The difference between removing delegations and quadratic delegations is pretty minimal. No one will look for delegations through snapshot, people will campaign and use use other methods with worse UX instead.

1 Like

Isn’t that what you do for a living? Specially taking into consideration you are repeating the same things yet again on that post while being exposed to arguments you can reason about.

We had this argument with Clement when we debated about delegations on UBI DAO. Since we applied token voting there, I was against them, he was in favor. I ended up agreeing with the consideration that the UBI DAO should inherit the laws of PoH DAO and we had a clarification vote.

Again: the UX argument is absurd. People have to sign with their keys for every new vote. People are not idiotic robots and usually like to know what they are signing for.

Thanks for your concern, Santi. Knowing what it takes to make a vote was the first thing I reasoned about a couple of months ago when pondering on delegations. I’m baffled that you only realized it now, and that it’s a discovery for you.

It is absurd, until it isn’t. It’s great to think on measures in advance and be consistent instead of being reactive and taking ad-hoc measures (like the MB resolution that was an overreach :man_facepalming:) when you feel it’s convenient for you.

1 Like

Yet another chance to offer a rebuttal without any argument exposed other than “I thot about that” (which is not an argument at all). ¯\ _ (ツ)_/¯

ps: If you actually thought about it, why you never mentioned it? Would’ve been a big plus to the debate.

@santisiri it’s clear that you are an expert in quadratic voting for all your comments here and in the telegram chat, but I’ve only hear the advantages of the quadratic delegation.

Can you point out the disadvantages so all the people have the complete information to vote?.

Thank you


Fair enough. The main disadvantage I could mention is that it’s still a concept that requires education, most people will hear “quadratic something” and it might not come naturally what it means, how it works or why is it beneficial to minorities over bipartisanship.

Similar to “liquid democracy”, it’s a concept that is still novel to most people out there… but my take is that these ideas are precisely what makes digital democracies incredibly superior to legacy paper ballot ones where tyranny of the majority prevails.

As an old political rival I used to have always said, “to do politics is to teach”.

If we become the first DAO that combines both Liquid Democracy and the Quadratic formula, our DAO itself will become an educational tool about some of the most profound innovations in democratic governance we have seen in the last decade.

Quadratic delegations are in my view a great feature where we have more to gain than to lose as a democratic community.

A couple of useful articles that can help bring more insight on the nature of the quadratic formula:


ps: making zealotry expensive as bloomberg describes it might be a nice antidote to the toxic politics we’ve been having.

Thanks for the explanation and links, but the links that you have provided are related to quadratic voting, not delegation, so I’m not sure they apply 100%. From your explanation, there is only one disadvantage: “educatiion the DAO members”, there are no others?. It’s all profit if we perform the change?

Other thing that I’m seeing, correct me if I’m wrong please, is that there are no antecedents of quadratic delegation, but there are some demos and cases in quadratic voting, so if this HIP is approved, will be a democratic experiment, I’m right?. Even more, doing a small research about the topic it seems that it’s an old discussion inside Democracy Earth the quadratic delegation that has not achieve a closure. Can you elaborate a little more on why is still an open debate inside DE?, and, What have you make change your opinion from this comment a few years ago?

I think the angle on…

Revoke as a control mechanism
Ability to not be conditioned if you really want to delegate power to someone.

… are strong arguments.

Thanks again for your time

1 Like

The thing that seems to be argued on the “other side” of the table is that people will organize around groups to vote directly on issues to bypass the delegating aspect… But to be quite honest, I see that as a net positive since it will push all parties to engage people more with the DAO rather than having a lazy attitude of “just delegate me and trust my will” which is what happens right now.

The quadratic formula applied to delegations is a novel idea in itself. Similar to how we applied quadratic balance to UBI token voting last year with the governance of the UBI DAO. There aren’t many precedents.

I think it’s desirable to use the quadratic formula when you have many choices rather than a simple binary choice in a ballot for example. Mainly because the magic of QV is something that actually helps capture not only the preference of a choice, but also the intensity of those preferences… that’s why long multiple choice lists work very well with QV. (fun fact: we actually picked the UBI logo that way back in the day!)

In the case of delegations, the novel incentive I see is that you might be better of delegating to someone with few delegates rather than going for power concentration in a single whale (Clement or myself). It incentivizes the emergence of many leaders and not just an duopoly or potential monopoly from whoever is more popular (I’m certainly cutting my legs with this).

You just blew my mind :exploding_head: with that one! I have completely forgotten about that thread and debate. At least helps show how far back our debates on these ideas goes… that was very prior even to the pilot we held in the USA :us: with the folks of radicalxchange.org (whom @paulaberman is now actually the COO of that organization).

I ask for your view about the disadvantages of using quadratic vote, but like you are avoiding the question, it’s like there are no disadvantages, or you don’t want to tell about them. Hope that is the first one.

Quadratic voting != quadratic delegation as for my understanding, that’s why I’m asking and asking to convince my self that it’s a good option to implement this. Right now, I’m still doubting it’s a good approach and still thinking this will be just an democratic experiment, not an improvement for our DAO.

You are talking about the benefits of the new incentives, but in my experience, every time you change an incentive, there are drawbacks in other side. That’s why I’m asking and asking about the disadvantages.

PS: Nice that you mentioned @paulaberman, because in that github issue, she has some concerns about the quadratic delegation and point out the benefits of the delegation that we are using right now. Quoting her words: “I also am uncomfortable with the idea of taxing users unless its a necessary fee for the maintenance of the system.”
And in a following comment:
“No one gets a blank check to do whatever they want, delegates can have their sovereigns revoke their delegation at any time. So the ability to revoke can also be considered a mechanism against the formation of monopolies.”