DAO improvement proposal of Gnosis - shift towards futarchy and proxy voting

[koeppelmann]

Hey - Martin Köppelmann here - Co-founder of Gnosis. We recently did a code review of "The DAO" and discovered some problems that we want to fix.

In short we want to:

• automatically create prediction markets for every proposal
• give the option to fork/sell out of the DAO under the condition that a proposal you don't like passes
  (currently you have to fork before you know whether a proposal you don't like will win the vote)
• allow putting your tokens in a proxy voting contract where
  1) others can vote for you (liquid democracy)
  2) you automatically vote according to the market (Futarchy)

Here is a video explaining why:

We will provide more details soon but would love to get some initial feedback. Exciting times.

[cryptowanderer]

In short, Gnosis for the win

Look forward to a full proposal, but these are good points in general.

[cryptoanarchy]

Was just in the middle of writing an article about the in-built prediction market, delegative democracy combo. Definitely the most important and exciting idea in creating a DAO 2.0 - bring it on!

[DRW]

I fully support the proposal, and think it indeed addresses a major problem. Great work. You say the system would work through prediction markets.. Does that mean that the proposed system will be add-on rather than a re-write of the DAO code?

One thing you don't mention, and that I think we need to have clarity about soon is whether or not the voting status (having voted or not) or the vote itself (having voted yes or no) is secret or public during the voting period. (Of course prediction markets will be public, I'm only talking about the real vote here).

I think that in order for the 'wisdom of the crowd' to emerge rather than herding behaviour and gaming/strategic timing of votes, it is imperative that we get a system where votes and voting status, as well as their summary is kept secret during the voting period (like in regular elections). Does your proposal incorporate this aspect in any way?

[jbaylina]

I created a DTH pool that can act as a proxy voting and cold storage voting. It's in the prototype stage. You can find it here:

jbaylina/DAO/blob/dthPool/dthpool.sol

There are some tests implemented.

[profitofthegods]

Sounds very interesting and the improvement for splitting is very useful.

I would like to see more information about how you stop people from buying a particular outcome that they want when people are voting automatically according to the market.

[bustya]

I dont see how this adds any advantage over simply adding a 'no, else split' option hard coded into the DAO contract.

[davids]

Martin, this makes sense. The one thing I would mention is that there will be other mechanisms proposed for proposing and voting on projects. Specifically, Backfeed, which could play a useful role here (disclaimer - I'm an advisor to Backfeed as well as Consensys). I wonder if perhaps these things would be better to split up, so they don't come as a package? Or perhaps that's how you envision it anyway? I think a modular approach driven by tests and approvals seems like a good way to go.

[ranford]

Very interested, and really like the idea of having an option for voters to assign proxies/hoefully full on liquid democracy or instead choose futarchy vote. I find both very interstuing and feel theres a need to see them in action to tell how they compare, but if you can have both options available when voting this could be fantastic.

also as per the post above, how can we maximize the overall picture that gnosis, backfeed and others can bring as a whole, Is there a way to combine the best attributes the gnosis and backfeed

If nothing else I think TheDao is already bringing great and important value by stimulating and pushing such things like these forward. For me this far outweighs the potential loss of funds we (Daoists) may lose if the experiment goes south

[Daofanista]

Is it an idea to built in small but smart incentives to reward the more active token holders that actualia do the voting, 'paid' by token holders that don't feel confident enough or don't spend time for that proposal to have a clear opinion? I think in spirit of a community it must be voluntary. For example made easy with a thank you button that will sent small amounts of ether to the voter that was representing you (preset like 0,01 eth 0,05 eth, 0,1 eth)? Can think of other possibilities.

[koeppelmann]

We published now a more formal proposal: find the discussion on reddit and the link here.

More will come soon.

[profitofthegods]

In this example from your proposal:

One market asks, “what will The DAO token value be at a time in the future if we implement this proposal” and the other asks ““what will The DAO token value be at a time in the future if we do not implement this proposal.”

What happens to the dead end vote? What I means is if you put in a vote for the value of the token if we do not implement the proposal, then the proposal is implemented, then accuracy cannot be gauged - so is the money wagered simply returned?

If the money is returned then isn't that very easy to game: If I want to get my own proposal implemented then instead of voting that it will increase the dao price if implemented, I just put lots of money into betting that the dao price will be very low if the proposal isn't implemented; this makes accepting the proposal relatively more appealing so the proposal is implemented, but I haven't risked any of my funds to manipulate the market because I just get a refund.

[profitofthegods]

If the money from a dead end vote isn't returned, then it is effectively a yes or no vote with money wagered as a degree of confidence, but this has its own problems.

1. People voting will not be selected whether which side to vote on based on their own opinion, but based on which side they think will win. I may think a proposal will go through but we would make more money elsewhere, but I am disincentivized to express this opinion

2. It gives a one-dimensional answer: dao price at a certain time. It does not consider dao price over different time-scales, legal or ethical considerations, possible dividends etc.

3. When the specific date for gauging the success of futarchy market bets comes along, there will be heavy incentive to manipulate the dao price.

[koeppelmann]

but I haven't risked any of my funds to manipulate the market because I just get a refund.

You do need to risk your DAO tokens - because if the proposal is not implemented then you will have to sell your DAO tokens undervalued. However - more important is that we think that a manipulated price will just attract more people to participate and bring the price back to the normal level because a manipulated price means basically "free money".

We are currently setting up experiments that will try to give empirical answers to the questions you raised: see here and here

[profitofthegods]

Then see my next post: if you do risk your token then you are asking people about price at a certain date, and making decisions about what to do based on that information, but what people are actually voting on is whether the proposal will be implemented or not - so you are misconstruing the result.

[koeppelmann]

It gives a one-dimensional answer: dao price at a certain time. It does not consider dao price over different time-scales, legal or ethical considerations, possible dividends etc.
When the specific date for gauging the success of futarchy market bets comes along, there will be heavy incentive to manipulate the dao price.

So in the version we have in mind we will not trade a prediction on theDAO price in the future but theDAO price at the moment when the proposal is accepted or rejected. We basically create two markets for the regular DAO tokens. However - on one market all trades will be executed IF the proposal is accepted and on the other IF NOT. Finally every one who decides to use this mechanism would vote for the version where theDAO tokens are values higher. Those to markets allow the option to make conditional buys/sells (I want to buy/sell IF or IF NOT a proposal is accepted)

Again - you are right - there is for sure incentive for price manipulation. However - on efficient markets where everyone has access to this should not be possible. But we will (try to) prove this theory by creating a incentive structure in a separate experiment setup.

[thecryptofiend]

give the option to fork/sell out of the DAO under the condition that a proposal you don't like passes(currently you have to fork before you know whether a proposal you don't like will win the vote)

I agree with it all except the above. This is the same as allowing people to split if they don't agree with the vote and is a big mistake which goes against the whole principal of the DAO. It will cause massive fragmentation of the DAO to the point where it will become useless as an entity. As has been stated by myself and others multiple times splitting is a last resort to prevent 51% attacks etc and is not a voting option.

Also there doesn't need to be an option to sell. People can do that any time they like anyway.

[Gnosis]

Edit: Replied to wrong comment in this thread.

[Gnosis]

Greetings decentralized autonomous denizens,

We have some updates!

Over the past few weeks we have made two posts to reddit which have received overwhelmingly positive feedback.

The first is a video which introduces the problems and solutions that we are proposing:
reddit post: https://www.reddit.com/r/TheDao/comments/4jmnhr/a_video_introducing_a_futarchy_oriented_proposal/

Post on dao.consider.it:
https://dao.consider.it/gnosis-futarchy-and-token-selling-if-bad-proposal-passed

The second includes a written proposal to the DAO:
reddit: https://www.reddit.com/r/TheDao/comments/4kaj3l/a_gnosis_written_proposal_to_the_dao_futarchy_and/

Our proposal was also mentioned in this feature Coindesk article!
http://www.coindesk.com/futarchy-dao-governance/

If you would like to join our Slack to discuss, please write to matt.liston@consensys.net!

We greatly appreciate any feedback that you can provide and are excited to formally submit this proposal once a path forward has been decided for the moratorium concerns!

Long live The DAO, Your friends at Gnosis