What kind of voting system do we want?

In the first governance proposal we have a choice and a possible voting, but what if 51% wants it and 49% doesn’t? Should we have an old democratic voting system that may lead to brexit-like outcomes?
I’d like to discuss with you about voting systems. How many voting systems do you know? What is the one you prefer? Do you know about the Arrow impossibility theorem regarding votes?

In abracadabra we use Snapshot, and you can read the different voting mechanisms here: Voting types - Snapshot

but once you voted, and you got 33% option 1, 33% option 2, 33% option 3 and 1% option 4, what should the community do?

How do we limit the scope of every decision into being a yes/no choice but still maintain efficient decision making? That’s a great question. These are the fundamental aspects of decentralized decision making we will need to figure out if these channels are to be effective means of organized decision making.

I have generated a proposal for standardizing the proposal format to try and ensure the discussions keep a relatively slim scope as a means to avoid creating situations where decisions being made are not based on fundamental concepts. Our goal as a body of governance should be to try and get to the root cause and effect of each decision we make and part of that seems to be effectively framing the decision/choice in a way that is conducive to this idea.

Part of my formatting proposal includes something you have included here! a link or reference section which is interesting to see.

can you explain, in general terms, what the Arrow of impossibility theorem is with respect to voting?

1 Like

Hi, granularization is a good way to make the decision process nice. Sometimes it overloads the voters with too much voting by the way.

From investopedia: “Arrow’s impossibility theorem is a social-choice paradox illustrating the flaws of ranked voting systems. It states that a clear order of preferences cannot be determined while adhering to mandatory principles of fair voting procedures…”

Easily put: you may have Yes/No voting and less votes in 26 states than those in the remaining 24 and win, because you count states and not votes, or you can have in ranked voting A B C and end up with both a majority of preferences saying A>C and have C>A at the same time. And if you add one single vote suddenly B overcomes both, even if you add a majority for A or C.

But this is partially important. In fact voting can be inspected in a granular way:

  1. First, the proposal: who can propose a voting?
  2. Voting allowance (everyone votes? or just those who participate to something or have certain characteristics?) or geographic/area/subject voting allowance (only people from Brazil or only medical doctors)
  3. Then the voting technique, for example approval voting is Arrow impossibility free, it doesn’t generate a false or tricky “win”

But it is at the end that the challenge is the hardest. What do we do with votes? Winner takes all? What kind of winner? What happens to the losers if they are 49%? Is that going to generare a split in the project?

Based on my understanding of the arrow impossibility theory, it would seem the best choice for voting would be simple yes/no voting.

Some things and a sort of suggestion that I think are worth considering:

Are those with the biggest bags given the most weight in voting thus ensuring unequal decision making by those with the most money? Whales would control the decisions under this scenario. This is positive and negative. I think it should be discussed though. Maybe the below thoughts could counterbalance the negatives which would arise from this issue.

As we are aware decisions on proposals will affect everyone’s economic prosperity, but I question whether I have the financial/defi expertise to make these calls with my votes? If I feel incapable of making an informed decision, surely others feel the same, can we somehow set up a system to delegate our voting power to others we believe to know more about these things? We would also need the ability to take our voting power back to use ourselves or redelegate to another.

If we do set up something like this, can we track each proposal how those delegates voted, what the economic affects are of that proposal across time, and then identify or rank those delegates over time based on their record of success in making decisions that grow the bottom line for everyone? We would need to be able to objectively assess a delegate by this mechanism

Yeah, competence and weight are important.

About weight, you may want to consider quadratic voting, in use by ETH communities, that gives more power the more people are voting something. So the math behind it would give more importance to whales but overrun them if one hundred smaller caps would reach the same weight/cap. Like a whale with 1B would have less voting power than 1000 users getting to 800M.

But then, what if the 800M are made of another dev whale who’s got 1000 anonymous accounts?

Competence is also tricky. You may have one user making 5 good decisions and then making a very bad one that screws everyone.

Yes/no questions are good, but you are imagining a question like:

Should we add this new token, $ABRA?
Yes / No

then you have different kind of voting to Yes / No. One is weighted, another one is allowing blank votes, another one is not counting people who didn’t vote…

even the simplest vote is different.

And finally: what do we do if 49.99 people say YES and 50.01 say NO?

Good thoughts on every point. This just highlights the complexities we have to deal with.

As to the 4 good decisions one bad, yes, of course, this is always a possibility. There is always the possibility of downside. You cannot eliminate that. However, in a complex society we are required to rely on those with different skills and specializations to do things that are in our collective best interests. We want this, because without this, we will not be able to maintain the complex systems that make our lives so relatively easy.

This is all just a long way of saying, you’re right, but regardless of you being right, we should still give competence and weight and allow delegation to others. The best we can do is design systems to retract delegation or affect reputation, etc., when that delegate makes a bad decision that harms us.

Regarding your comment on yes/no questions, I agree with you here too, but again, maybe this should, at the beginning be the limit of it, and proposals must be simplistic enough to fit into this rubric until we can ensure this is effective and working. This yes/no idea also avoids the serious problems raised by the arrow impossibility theory raised by you. Again, not sure there is any correct answer here. This just seems like one way to simplify things (at least at first) and reduce risk.

There will be no perfect system because human beings are involved. The best we can do is mitigate risk on both sides without devolving into fantasies about any side of the equation.

I just want reiterate these are just some early thoughts on this, that I felt worth putting out there for discussion. Please feel free to tear these apart.

Your answer is wise. It is a great conversation, thank you.

“in a complex society we are required to rely on those with different skills and specializations to do things that are in our collective best interests. We want this, because without this, we will not be able to maintain the complex systems that make our lives so relatively easy.”

agree, well said

“but regardless of you being right, we should still give competence and weight and allow delegation to others”

I don’t care, obviously, to be right, I am here to look for a good conversation with you and anyone else interested, in order to find a good possible option. And, yes, there’s always a downside. That’s actually the trick, not to have the perfect system, but a learning system. To ALLOW mistakes, not to forbid them. It’s like in trading, you need retests, you need dips, you need some mistakes to learn and grow.

“There will be no perfect system because human beings are involved.”

oh that’s not because of humans, it’s because good and bad are not part of nature, in physics there’s no good or bad, so you don’t have “perfection” but “adaptation” :slight_smile: I like to say we need an adaptive and learning system. Our community should fall, but learn from falling. No need to tear anything apart.

My interest is let’s say we use a yes/no question, without doubting of the yes/no, once we have 51% yes and 49% no, are we really sure we should do the yes? That’s my main concern.
Maybe in that case it’s better to discuss things more, because what could stop in that case to tear the community in two, or better, to tear the market cap in two?

Quadratic Voting (QV), first put forward by Glen Weyl years ago and really the only way to adapt democracies and voting to the modern needs. In short, it allows for people not only to express the direction of their vote on certain polcies (yes or no) but also simultaneously express the intensity of their preference…

To put it simply: Individuals buy as many votes as they wish by paying the square of the
votes they buy using some currency …now this sounds backwards and vulnerable to exploitation but has received support and endorsement as the most efficent method of voting from guys like Buterin.

It greatly mitigates a few issues:

  • tyranny-of-the-majority

  • fractional control issues

And it has seen real world use, an example is the Colorado senate which went as such:

In 2018, they used Quadratic Voting to decide which appropriations bills to fund first. Since legislators were likely to sponsor their own bills and vote for them, the Democratic caucus sought a method to gauge which bills had everyone’s support.

Initially, the Colorado Democrats assigned 15 tokens for each legislator to use on their preferred 15 bills. After this didn’t work well, they talked to Microsoft economist Glen Weyl, who explained how Quadratic Voting could provide a solution.

Weyl saw Quadratic Voting as a solution to the ‘tyranny of the majority’ issue. Regular voting assumes that everyone cares for an issue equally, which is rarely the case. The reality is that some legislators do not care about certain issues, care moderately about others and care deeply about a few.

So, instead, each legislator was given 100 tokens. If a legislator cast one vote each for several issues, it would cost them one token each. However, a legislator could cast more than one vote for an issue, at the following cost in tokens:


The only real vulnerability is large scale sybil attacks but there many ways around this especially with zero knowledge proofs becoming more and more available to developers.

1 Like

Yes, we were considering that too. The problem is always the result, not just the voting process.
What happens if the output of a quadratic voting says 51% yes and 49% no?

Is it too simple to change the passing threshold for votes depending on their implications?

Perhaps significant changes could have a higher threshold of votes to pass?

it seems reasonable that more impactful changes should require a larger proportion of votes to pass. and smaller changes should require fewer votes as to not impede proposals that are narrow in scope and can be implemented quickly.

Of course this brings the issue of deciding the passing thresholds for proposals as well as the total vote threshold for a proposal to be validated. We definitely don’t want a small amount of votes on a proposal that vote together to pass a proposal that mayb ewasnt visible enough or hadn’t received the proper planning or due diligence from the community.

Some examples:

Partnerships or proposals that involve revenue sharing may be required to have certain threshold of votes in favour → maybe 75% in favour with 65% of total available votes cast.

Capital expenditures of certain amounts may require different passing thresholds.

  • Level 1 expenditure - $X → 65% in favour with 65% of total available votes cast.
  • Level 1 expenditure - $X X-> 75% in favour with 65% of total available votes cast.
  • Level 1 expenditure - $X X X-> 85% in favour with 75% of total available votes cast.

Of course these are just simple examples but I believe this method will be able to ensure the votes are a good proxy for the community if implemented correctly, the biggest thing is to ensure these thresholds make sense statistically (this would also have to be studied) in respect to the potential impact of any proposal.

We do want to avoid rehashing proposals or reversing course if possible since this is a decentralized decision making process. It’s going to take a bit longer than centralized decision making processes but if the proposals are well thought out, and the community is diligent, we can be as efficient as some dudes sitting around the board room in uncomfortable clothes shooting the shit for half their existence.

Our biggest asset as a decentralized decision making body is that many hands make light work and being around enough large cap-ex projects that I have come to understand the ones that fail are those that aren’t backed by the requisite planning, the ones that get rushed because people are too invested in the short term. In these situations the short term sacrifice (usually time and efforts) makes for long term gains. This should always be the focus of the decentralized decision making process imo.

It’s a good idea, similar to quorum, but yes then you have to define who chooses when to have a higher threshold.

Every community ability to decision is based on common knowledge. Understanding. When a voting is 51/49, similar to UK Brexit, it is probable that the community does not understand what is going on.

Other voting instead can be addressed with proportionality. For example, if some money has to be destined (yes/no) to a project, having 51/49 may say “let’s give 51% of what they asked” for that project.

Some other voting can be “integrated”. If 50% want more security and 50% more development, we could invest in the development of a new form of security that may be disruptive for the market.

Is this something that we could talk about?

Great conversation and discussion here, all posts have some valuable ideas.

Just a note to @amdp - please when you are opening these topics in the future restrain from using examples like you just did with Brexit, that imply that majority of people in the UK was wrong to vote yes. I can tell you right away, that is not easy to argument.

It doesn’t matter what is mine or your subjective opinion on the matter, but it was a democratic vote and it needs to be respected. Because not respecting it will open up a pandora box of different opinion of many other different elections.

That being said, I’m always in favor of democratic majority, the issue with our voting system here will not be 51-49, but how many people will actually vote.

1 Like

Hi Federeico,

DeFi is maybe one of the places where freedom is defended, so why should I restrain to talk about Brexit or to express my ideas, regardless if they are not what you think? Is there some sort of rule about it?
I am not in opposition, I just want to understand you.
Another question is: why technically you think we won’t have a 51-49 situation here?

Thank you for your participation in this topic, again, I’d like to understand you better.

Hi amdp, no problem to explain more what I thought, its a healthy discussion.

My train of thought is the following - in your opening post you say “Should we have an old democratic voting system that may lead to brexit-like outcomes?” implying that the brexit outcome is not favorable by your opinion, and i respect your opinion.

But if my opinion is that it was a good choice - then your whole argument looses its strength. I would just nod my head and say - yes, why not as you can see it resulted in a good decision.

That would lead us to discussing whether or not Brexit was a good decision, because your main argument why old democratic voting system is a failure is based on the outcome of it. And outcome is highly subjective, and I think if we are to design a better system for voting here, it needs to as neutral as possible.

I think that’s why those old democratic voting system still prevail - they value opinion of a sheep farmer and phd in math the same - which is what I like about them and that’s where they remind me of DeFi. Everyone scales the same.

That was my whole point, as these topics need to be approached from a neutral philosophic way, to say it like that.

As why do I think we will not have a 51-49 situation here, that’s not based on any math or data, just my experience that communities are either highly excited about something, or not at all. Could be I’m wrong!

Hi Federico,

thank you for your warm explaination. Yes I agree, it is an healthy discussion.
59-41: okay, good for the DeFi, I see there is not that much need for cases like the one I brought, but you know, to make a system work, you need to see the risks in a stressful situation, non in a plain one.
However I respect your opinion and maybe it is not that frequent.
We could discuss even a 71-29 case, in that situation there’s 30% of the people against it, they’re not a few people!

I will make then a brief round on the Brexit topic and then conclude my answer. I don’t care at all about the Brexit outcome. You are generally right, we shouldn’t discuss the outcome of a voting like that. I wrote stuff in a weird way and generated confusion: I am not saying Brexit outcome was good or bad, I am just saying "look, in cases like that, is voting and “winner takes all” outcome the best one? What I see is that is almost HALF of the country was against ti. That is enough. It is a strong base for a conflict, it seems like a soccer match, a competition, not an emergent profound thought coming from the people involved. During that voting a politician was killed, many interests were conflicting, a lot of lies have been told in order to “win”.

“Old” democratic systems value the vote as executive, I think we are the future and can proceed with a system that treats the vote as “informative” instead.

If you have a 90% majority, it’s clear what people want, and you can just memorize the opponent position and keep it for further reference. But when 50%, 40%, 30% are against it, a lot of people will be disappointed and feel unrepresented if you approve it without a decent discussion or similar.

To be clear, I see votes like a representation of the people, an “indication pointing out where to go” not a dynamic mechanissm that “scores” the ideas like the points in a sport match, and who wins, wins. I mean, it’s not a match, it’s people reflecting on a subject and trying to find the right answer :smiley:

1 Like

@amdp thank you for reply, nothing to add, I understand your point of view much more , and I support the future search for a voting system (if needed).

+1, those close calls especially need quite a lot of discussion and shouldn’t be green flagged with same effort as 90-10 situations.

Best of luck in 2022 !