It seems to me the main point is this mechanism will introduce economic security, where if an oracle misprint happens, the protocol can refund the affected users. This is a feature that no other oracle has afaik.
It’s hard to me to be of the opinion of rewarding stakers for passively staking without any skin in the game. While it’s true that stakers could be rewarded for voting, that also seems hard to implement soundly with the right incentives because the strategy of voting everytime without even checking the proposal might optimize short term returns for the staker.
Delegated Staking: Non-publishing $PYTH stakers can delegate their stake to one or many publishers.
I have to warn about about this. There is currently a really huge risk on the Zksync network. I think Syncswap got like 25% of all delegations, so if this route is to be going, then there should be a cap on how many tokens which can be delegated to each publisher.
The reward cap does not protect against this issue, as most people will just delegate and forget. And reap the benefits directly to their wallet
That’s a very fair and good reason to think of the situation which I can truly understand however I still seem to disagree about it.
I would 100% agree that the deposited funds should be slashed and used as a refund in case of a misprint but I believe that there’s no reason for the stakers to take part in the refund as they essentially have no impact on the publishied data. Therefore, the responsibility of refunding users should be the publisher’s who is the only one responsible for the prices that it publishes.
In addition to that, I believe that unlike traditional blockchains that use delegated PoS models, where the delegates actually use the delegated tokens to actually produce more blocks, the data publishers don’t need any delegations to produce the exact same data and price feeds as they would’ve published without those additional tokens.
Thus, publishers should be the only responsible for misprints and refunds since the only ones “messing” with the data.
As for the voting and staking, I think delegating votes to someone who apply as a delegate and who you trust might help for the problem you brought up in addition to slashing of staked tokens that didn’t participate in over a certain percentage of the votings over a given time. I also think that in a decentralized organizations, such as the pyth DAO, there’s never a way to actually make everyone go through the details and vote wisely, with or without rewarding them. you simply cannot force anyone to actually read everything before voting and the best possible solution might be to allow them to at least delegate votes to someone who they trust to vote for them.
Just like you can’t force a civilian of a country to actually go through the whole candidates list before voting you can’t do that here as well
I love this ambitious proposal … this is one of the reason this forum exists . Let’s bring it up !
We need more like this as the next chapter for staking will be to generate value for the protocol and for the stakers.
Great proposal, very thought out & well stated. I think the incentives are def. aligned - I would love to see what a proposal looks like where the “stakers” also have some skin in the game & potential rewards !!
This sounds a lot like Pyth Stakers getting a sustainable and productive source of yield for staking $PYTH!
Additionally, this also builds on @oraclemaxi forum which encourages good behaviour for data producers and adds additional security to Pyth as an oracle. My ideas on that forum mentioned still apply:
Rewards data producers for data integrity while punishing bad behaviour
Improves chance of higher quality/accurate data for consumers
Increased security for Pyth as an Oracle
First of all great proposal @CMS, I think this idea is doing something innovational that’s going to take $PYTH to the next level!
If stakers aren’t involved it wouldn’t be community lead, by having the ability for publishers to have a delegation fee they are rewarded for allowing stakers to have a share of their reward pool! As they add more symbols they also don’t need to fill that $PYTH cap themselves and can rely on the community. This is a great win-win situation that every PYTH staker should be excited for!
I think this proposal might be a key in making Pyth the biggest and best Oracle. By rewarding data publishers higher caps per symbol contributed they are encouraged to continue providing data to Pyth. Additionally, this will increase the # of data providers per symbol reducing reliance on a signal point of reference! (INCREASED SECURITY to DEFI). By having slashing, data providers and their delegates will be impacted by the negative behaviours, which will help the community identify bad actors and encourage more accurate/quality data in general!
I think we need to merge the community and these data publisher rewards/punishments together! This proposal does that and pushes for true governance and participation
Simply put every PYTH Staker should want this proposal to pass as this is great for PYTH + DEFI and Stakers (REWARDS/YIELD)
Honestly I’m all for this system of incentivizing & slashing, based off results & voting (with your staked Pyth). Love to see governance in a true & valid form coming to life
Lets keep these productive discussions going & discover something special !!
I actually disagree here slashing will impact people delegating to publishers therefore this will actually increase good behaviour!
If a publisher decides to publish a ton of symbols to increase their reward cap and people delegate to them! If the publisher screws up anything ever it will result in slashing which impact them + everyone staking! There is reputational risk here as people who get slashed will probably never delegate to the same publish again. The more Pyth staked in your pool the more your on the line to keep publishing the highest quality data
Additionally, to your point of discouraging new data publishers I don’t think that will be the case. If a data publisher believes they have the best/most accurate data they will be encouraged to compete for the Pyth rewards. With new tokens being released everyday I doubt we will fall into a situation where there is a single one “best” publisher! Looking at how delegation works for staking on other protocols things usually get distributed across many leaders. If these leaders are providing excellent data I think thats desired!
I think ignoring stakers would actually cause more issues! If data publishers are the only ones getting rewarded and punished here and the community has no involvement or say…how is that community lead or great for our governance?
It also would be weird to expect these data publishers to have even a 10th of the total staked $Pyth normal stakers have! I wouldn’t be surprised if its more like 1/100th or 1/1000th.
I think the community standing side by side with the data publishers and getting rewarded and slashed together is the way forward!
Normal Pyth stakers will help these publishers cover their caps in return for a delegation fee. The Publishers and Stakers will get rewarded together and if something bad happens they will be slashed together!
This will allow normal users to get skin in the game while encouraging data publishers who don’t want reputational damage and backlash from the Pythain brotherhood to act in the best interest of DEFI!
This person sees the vision of what this proposal is taping into!
Being reborn as a Data Layer is super wild !!
I think this is the grand vision of what’s possible when we have a proposal like this pass! In the digital age Data is the most valuable resource and in the DeFi space that secures billions of dollars of TVL/transactions - we need a leader that will show the way to the highest quality + accurate data. I’m convinced that LEADER WILL BE PYTH!!
First of all, with or without delegating to publishers the community still has 0 impact on the data thats being published and the fact a certain delegate has got more delegated tokens wouldnt just magically change the data they are publishing and make it more accurate all of a sudden.
As for the point about slashing and reputation, if a publisher is found publishing “bad” data and gets slashed his reputation will be much lower regardless of how many stakers it has. In order to publish the most accurate data out of pyth we must rely on statistics and facts and not on peoples opinions.
I understand the “we are in this together” sentiment and that delegating tokens is familiar to many from other networks but unlike other networks where the staked tokens are used for the actual building of the network’s blocks etc, here in pyth the publishers dont need any tokens to simply publish their data to the network so besides putting their own stake at risk (because they are the ones putting that data out to the public) there is no reason for additional tokens for them to publish the same data. If they know they get slashed/rewarded for their own stake they will simply do their best to publish the highest quality of data, therefore further collateral would not change a thing about the way they publish the data.
And besides, if the actual tokens of the stakers (that could definetly be rewarded in a model of their own) are delegated to the publishers who’s gonna take part in the DAO voting? Also the publishers? Where will be the community “say” if all their voting power is in the hands of the publishers?
I think your underestimating the power the masses/community have on organizations/businesses. The stakers won’t have impact on the data published BUT they will have an influence on publisher through delegation fees. If the publisher doesn’t have integrity and acts badly they will get slashed but also face backlash from every person who staked in their pool! That is a powerful deterrent for data publishers cause now they are in the spot light more so than ever before. The Pyth community is a powerful entity that can bring attention and love to these data publishers while also keep them in line from bad behaviour in fear of backlash.
While your right that unlike other protocols the staking/delegating in this situation is different, for what Pyth is trying to do I would argue it’s just as important. Pyth is an oracle which powers Defi with the best data. Without having the Pyth community involved in this mission of securing DEFI it falls short and doesn’t make Pyth any different than a centralized organization. Pyth has always been community first therefore I think its important that the community gets rewarded and punished with these publishers! This will encourage Pyth stakers to make informed decisions and trust in data publishers that are good for our ecosystem. I think this is a great Win-Win situation.
Pyth publishers getting tokens for publishing good data allows Pyth to grow with more data feeds and rewards them for a valuable resource they are providing. I don’t understand your point on “there is no reason for additional tokens for them to publish the same data”, as the cap only increases for an additional symbols a publisher adds not for the same one. If a different publisher adds a symbol that already exist that’s not a bad thing as it gives another source for data aka counters single point of reference risk.
I think there is misunderstanding, when you stake your tokens your able to participate in governance automatically cause your a staker. Delegating to publishers shouldn’t effect our DAO or governance in any negative way. If anything it will actually encourage more people holding Pyth to stake (to get yield/rewards) which will bring more people to our forums and in rich our governance!
I think the delegation of staked Pyth doesn’t relinquish the user’s voting power. I’m making this assumption as no where in this proposal does it mention that the publisher gets all the voting rights of Pyth Stakers @CMS would you like to comment on this point as I think @TheDetective has misunderstood this
Sir with all the respect I’m feeling like you repeating the same things just in other words so I’ll try to sharpen my points here.
First, publishers can use their own stake to get rewarded and slahsed and dont need any delegations in order to put out the exact same data. Their own tokens being slashed or getting rewarded on top of them is already enough for them as an incentive to publish the best data possible. Therefore when I said “no need for additional tokens” I was referring to the stakers tokens that arent needed to make this process happen.
Second, as I already said multiple times before, pyth stakers CAN and SHOULD get rewarded but they deserve a different model for that where they actually and truly impact something, like they already do in the DAO. This model of slashing and rewarding publishers proportionally to their stake on each price feed is great but its for them to be involved in. This model will be no different with or without the stakers in it and thats the main issue here as they dont change a singe thing about the accuracy and quality of the data.
Also, people choosing their favorite publisher will not make any publisher release better data in any sort of way and once again, there’s no reasonfor stakers to get rewarded/slashed based on things they have 0 impact on, such as the data and they should be rewarded/slashed based on things they can do, like DAO size and voting participation
1 ) Sure publisher can use their own stake but they would need to own a ton of $Pyth depend on the caps. If the caps are designed to be inclusive to the community it would factor this in so that they wouldn’t be allowed to just fill their own cap themselves. I think there would be a need for additional tokens if the caps are designed to be large. Additionally, due to delegation fees publishers would be more than happy if stakers are contributing to the pool and they don’t need to keep filling the caps with their own $Pyth. Another point here is the potential for Data Publishers to gain supporters and attention through their data publishing efforts. By having stakers join their pool they are in a way also gaining supporters + credibility + reputation. Being a top data publisher on Pyth could open new doors and be more valuable in this new more public model than it is currently.
2 ) “where they actually and truly impact something, like they already do in the DAO.”…In this proposal they would actually truly impact the way data publishers, take their role more seriously. I think you totally missed my point on the importance communities have in crypto. Additionally, stakers will also impact the data publisher’s Delegation fees which is another important portion of rewards they can get. Stakers are absolutely critical to the proposal by CMS but for some reason you think it would work the same without them.
I disagree that the model would be the same with or without stakers. In traditional law there is the Judge and Jury. While the Judge has the knowledge of law, our system uses a Jury (group of normal citizens) to ensure a fair and justice outcome + to reduce single person bias. Bringing it back to this proposal, though the publishers are the ones issuing the data the stakers (community) are working hand in hand to ensure publishers are get more rewards (delegation fees) and have pressure to perform at high standard (reputational risk + improved attention/marketing). Without the stakers this system would:
Require data publishers to have a ton of Pyth (Pay to win)
Lack direct community involvement and education on the various data publishers (Not decentralized, not leveling up community)
No additional encouragement from community pressure/support
On point 3 there is something called the Social Identity Theory in psychology which states: "Individuals derive a sense of identity and self-esteem from the groups they belong to, including their organizational or leadership roles "
If Data Publishers are able to get the support of the stakers and establish themselves as leaders in providing accurate and quality data, they will per this theory identify with this trait. The stakers will be the fuel that pushes them into maintaining this high quality of standard
3 ) This was addressed in point 3 regarding the Social Identity Theory. I believe this proposal will encourage a new level of standard for quality and accuracy of data. Involving stakers is absolutely critical for this system to work. Stakers should be rewarded and punished with their publishers to encourage more education and involvement of the community. This is the decentralized way of giving POWER to the people! They have a ton of impact in terms of influence and sadly I don’t think you see this point.
Great to see so much interest in this idea! Two questions for now: 1) what’s the utility (destination) of any slashed assets (i.e. $PYTH), and 2) “Pyth stakers can delegate their unlocked tokens” - does this mean holder of unvested tokens can participate in rewards? If so, would it be worth exploring this for those that only have vested and staked tokens?
This is a great proposal that will benefit stakers and make publishers more accountable of the data they publish.
The proposal aligns the interests of all stakeholders
If publishers rely on the staked tokens of the community in order to maintain certain price feeds it could ver yeasily blow up for them in a bear market conditions for example where many stakers are likely gonna take their tokens out of these pool and sell it and then these pool won’t have the required minimum to function and work. Thus, publishers simply cannot rely on the community for this and must have their own minimal stake at any of those pool in order to make sure they will flawlessly function.
besides, the stake per price feed doesn’t have to be insanely high anyways so that smaller publishers will be able to pretty easily take part in it as well.
The data publishers role is to release the best data possible to the pyth network. If they have tokens at risk it will be an extra reason of why they should perform at the higehest level. I understand how strong community helps project grow but this isn’t the case here. The pyth community is surely great and strong but a data publisher who has HIS OWN tokens at risk will try publishing the best data possible regardless of whether people gave them extra tokens or not. And again whether community stakes by its own or through data publishers will not change the prices that are being published by publishers and will not change the decisions that are being taken by the PYTH DAO.
Psychology shouldn’t have anything to do with this discussion if being honest here. The fact it might feel “safer” or “familiar” to perform a some sort of delegated staking model doesn’t mean the ecosystem will benefit from it rather than 2 seperate models where publishers get their allocated reward (aka “publisher reward”) and stakers get their own rewards (“ecosystem growth”)
Its been a good discussion here but I feel like we should end it here as we start going in circles. plessure talking to you!
