Multi-stage Governance Proposals, Starting w/ Stage 0 and Stage 1

Here’s a crazy idea I’ll just throw out there, in case we need another way to disincentivize spam: if you adopt a stage 0 proposal, you stand to lose some ICP if the stage 1 proposal is rejected.

“Sort by Stage 0” is the new “sort by new.”

1 Like

I like this idea. Stage 0 still costs the proposer 1 ICP for rejection but nobody gains from voting in either direction at this point. Could Stage 0 be configured to require manual votes so a neuron with lots of followers can’t steer the outcome?

1 Like

I assume that could be done, but perhaps the threshold should be reduced if we did that.

Love most of the original proposal, not a fan of this spin off - we should use carrots not sticks to incentivize voting.

What happens to the minority in this case? I don’t think we should penalize voters for supporting minority opinions - this would just make centralized parties more powerful.

2 Likes

I think this comment is a very valid concern: Proposal to restrict rewards qualification to a threshold - #51 by Manu

I think right now people pretty unanimously vote no on spam proposals, because they’ll get the increased rewards anyway, and additionally the 10 ICP of the proposer will be burned if they reject. With your proposed change, the voters are now actually incentivized to vote yes on spam proposals, because if it reaches at least 3% yes, they will get boosted rewards. So I expect that many will vote with their own financial gains in mind, and that therefore all spam proposals will reach 3% yes, and the change you proposed will not have any effect.

2 Likes

Yes, I’m very confident a 1% threshold will be met if it is required to achieve stage 1 where voting rewards are offered. Here is my high level assessment of current voting power…

Typical has been 46% total voting power participation on governance proposals when Dfinity votes

22% VP owned by Dfinity and/or ICA
Up to 3% VP following Dfinity
6% VP following cycledao (when they vote first and independently)
2% VP following ICDevs (when they vote first and independently)
2% VP following ICPMN (when they vote first and independently)
Up to 1% VP following some combination of cycledao, ICDevs, and ICPMN
(Also important to note that Cycledao and ICDevs follow each other and/or ICPMN, so most of the time when ICPMN votes No first it cast No votes for all 3 public neurons and when ICDevs votes Yes first it casts Yes votes for both ICDevs and Cycledao)
The remaining 10% are not public neurons and it’s not possible to know who is voting and what is their agenda, but max voting rewards does seem to be a priority.

Also, when a proposal is submitted that has impact on voting rewards, a lot of voting power (maybe 2-6%) will vote manually before their Followee votes.

The spam has become routine enough now that people just let their followers vote for them instead of trying to make the spam pass.

My best guess is that there could be up to 15% of total voting power that would be willing to vote for spam in order to get more voting rewards. Hence a system like this would probably just incentivize voting for spam unless you first remove the incentive for spam for the purpose of voting rewards. If it takes all the voting power behind public neurons to vote against spam then the proposal might as well be stage 1 already.

I could end up being wrong about all this, but it’s my best guesstimate after watching voting patterns for the last couple of months.

4 Likes

I think you’re right, I think the threshold would be met quickly if stage 1 still provides outsized rewards. What do you think of this?

Here’s a crazy idea I’ll just throw out there, in case we need another way to disincentivize spam: if you adopt a stage 0 proposal, you stand to lose some ICP if the stage 1 proposal is rejected.

@justmythoughts I know you don’t like the quoted idea, but how do we solve this problem then?

Love the naming of “Stage0” vs. “Stage1” - it comes through much clearer than “incubation period” :laughing:.

From the UI side, this could be as easy as having a different tab within the voting page that the user has to click on like “Browse Stage0 proposals”, or a search bar where you have to put in the proposalID.

2 Likes

You are right it needs the right forum and appeal to a wide range of stakers.

Well wait…we just need enough people in stage 0 to vote reject, even if the threshold to adopt is passed. Do we all think that there won’t be enough people rejecting stage 0 proposals to counteract those who adopt for the rewards at stage 1?

We need a separate mechanism to handle financial incentives. I believe that actually stems from how governance rewards are weighted coupled with how NNS voting rewards currently work. I detail a potential solution and the issue here - fix the governance rewards so it doesn’t matter if 0 for 1000 governance proposals are submitted in a given time period - the rewards remain the same

1 Like

I think the plan is good, but I will like to suggest a slight modification.

  1. I think stage 0 should not be for a particular location, but it should be for those that vote manually only. This way the network can ensure its decentralised and everyone can be a gate keeper as long as you decide to vote manually. Also, this will ensure that more informed people, and people that are dedicated to the network will be involved in the decision process.

  2. As for the UI, they can just include a button (maybe called it ‘council member’) can be included in the current UI, and this button will automatically appear immediately you stake, but disappears immediately you become a followee to any body on governance.

  3. If the individual decides to start voting manually again by unfollowing any of those known neurons, the button should have a time delay(maybe 14days) before it becomes accessible to them. This way, you can’t make a proposal, then unfollow and immediately go and vote on it.

4.As regards the reward, I think there should be something, but it should be very small.

  1. As regards the my issue with assigning a particular location, it will not be effective if most of the neurons in that location are followee of a particular known neuron. That will result in some form of centralisation.

I’m not sure what you mean when you use the word “location” here. Can you elaborate?

think stage 0 should not be for a particular location, but it should be for those that vote manually only. This way the network can ensure its decentralised and everyone can be a gate keeper as long as you decide to vote manually. Also, this will ensure that more informed people, and people that are dedicated to the network will be involved in the decision process.

I think requiring manual voting on stage 0 could be a good idea. I’m curious to know what others think.

2 Likes

I would agree that stage0 voting should be manual. It ensures there’s enough organic/grassroots support from multiple neurons for a proposal, and also makes it more difficult for a proposal creator to reach the threshold simply by gaining favor with a single neuron.

It’s the closest thing we have to asking for ballot proposition signatures imo (since no proof of personhood).

2 Likes

I was just making reference to this.

1 Like

@justmythoughts @lastmjs

I haven’t commented on the idea of fixed rewards for governance proposals in a while, but there is an aspect of it that I think could become a problem for further decentralization later. As we move toward Dfinity and the IC community wanting more proposal topics to become decentralized, I think the only way it will happen is if those proposal topics can have high proposal weights like Governance does today (there are other requirements, but proposal weights seems fundamental). Hence, I’m currently a bit skeptical that fixed Governance rewards is a good idea.

I don’t think proposal weights is the root cause of spam. I think the root cause is the imbalance in voting participation that is caused by default configurations of Followees. Default Followees has lead to an expectation that people are entitled to voting rewards because of staking, but entitlement to voting rewards is actually granted only if voting occurs. Hence, we really need to focus on solving the voting participation problem across different proposal topics as the solution and I don’t think fixed governance rewards accomplish that goal (even if it does disincentive spam for financial gain). Proposal weights moves people to act very quickly and we need that tool available when it is needed again.

1 Like

The risk with making stage 0 manual is that whales would fully control the stage 0 outcome and they are the most likely participants to want to allow spam through to stage 1. Right now ICPMN and ICDevs are accountable to the public because the people behind those organizations care about what the public thinks and they are active in these governance discussions. Hence, there is incentive to not support spam and to want to do something about it including vote No.

1 Like

Wenzel,

If I submit 0 governance proposals then governance rewards make up 0% of rewards for the day.

If I submit 1 governance proposal then governance rewards make up 75% of rewards for the day.

If I submit 50 proposals then governance rewards make up 99% of rewards for the day.

That’s the problem, clear and simple. Fix the rewards at 75% per day (which is more than enough to incentivize participation).

If there are 10 proposals in that day and you vote on 5/10, you miss out on 50% of the governance rewards you would have received if you had voted on all 10/10 proposals, and those missed rewards are distributed to those who did vote on those proposals.

I think the problem is pretty clear and has nothing to do with decentralization or default followees - see the code for yourself.

2 Likes

From my observations over the last few days, whales will 100% promote at least one proposal a day to get max rewards.

My 3% proposal that basically did this was halted unless the weights were readjusted.

With out adjusting the weights, very little works except jacking with peoples neurons(from other experiences they really, really hate this)

6 Likes