[Validator Contribution Encouragement]

Hub Proposals Proposal Ideas
1

Every dPoS network and its coin investors do not want to see the domination of operating-only-commoditized validator set. Active participation and contribution from validators is what all of us hope for.

From Martin/Ethan post(The curious case of Vitamin B, bootstrapping and the concentration of power. | by Martin Worner | TgradeFinance | Medium), I came up with an idea about a on/off chain procedure to completely curve the mindset of delegators and validators upon delegation choice.

I would call this, “prediction/oracle market on future validator contribution by delegators”. Detailed pseudo process is explained below.

An idea about contribution encouragement

  1. we can have a contribution reward for each quarter. Contributors advertise their contributions, and community discuss it offchain

  2. we will have a gov voting for winner. We might can have gold/silver/bronze winner as an example

  3. those winners will get a prize of atoms. But the prize will be not distributed to the validator, but to its delegators! The distribution of the prize will be related to 2 major components, length of delegation and amount of delegation. Delegators will get more prize if he/she has delegated longer time and more amount.

  4. it will naturally cause delegators to delegate more tokens to “contributing validators” than purely operating validator because it has better chance to get prizes in the future. The dramatic change of such mindset of delegators and validators on choosing validators are so natural in economic incentivization and also align with moral intuitiveness on community collaboration.

  5. the effect of suggested prize is relatively low for high power validators so it is naturally decentralization friendly! It encourages more extensively to small validators because it can more efficiently gather more delegation by contribution.

  6. Prize amount can be correlated with commission rate so that low rate validators gain less prize so that the ecosystem encourage adaquate level of commission rate in protocol level, but not enforcing it.

  7. for better accuracy of such contribution Oracle, we can budget little bit from prize to distribute it to the winner pickers(incentives for voters to choose the right contributor). It will be a small prediction game for all delegators :slight_smile:

  8. currently there is a strong trend that validators becoming selfish and put more efforts on providing services only to their delegators. It is definitely not a efficient activities to grow the network. This kind of protocol can change validator’s mindset from concentrating on exclusive service to providing more inclusive service, which will become the new evolutionized standard of competitiveness definition on blockchain ecosystem.

2

I agree we need to change the mindset of delegators. Your suggestions have inspired me to think of some other aspects.

  1. I also noticed that some validators have become more selfish than before. It may because they are really focused on their business. At the end of the day, they need to report to their stakeholders (eg. investors), whose goals may not necessarily align to that of the community.

  2. I have some concerns on incentivising delegators in the way according to your suggestion:

  • Validators are already struggling. Contributing validators are struggling harder. Given they have contributed a lot, and the new incentives have gone to their delegators but not themselves. I think this is not fair.
  • In contrast, the delegators are already enjoying the staking service of that validators. They may enjoy other contributions by those validators. And now they can even enjoy an extra monetary incentive. Sounds not quite right to me.
  • The benefits to the validators under this mechanism are too indirect and too little as I can guess (but need some more concrete numbers)
3

I will show you an example of the suggested algorithm

Say a small validator wants to grow, by contributing to the network. When it is payed by onetime direct to validator, it is not continuous growth. Also he/she even does not need to be a validator at all.

Also, delegators do not have any economic incentives to delegate to contributing validators. Hence, contribution activity does not impact much on delegation growth for the small validator.

Let’s say the validator won 10k atoms and assumes that it has 100k atoms in delegations. All delegators to the validator will win 10% gain at once, which is a huge gain for delegators. Now the validator shows intention to further continuous contribution. It is obvious that delegation will flow in, to gain exposure to future prize.

Now the validator got 1million delegation. Now it generates 10k atoms per year assuming 10% commission. It is continuous growth and also it gives the validator responsibility to keep validating healthy and contribute more. Still if he/she win again 10k atoms, it is 1% gain for delegators, which covers all the 10% commission for a year. It is still quite an incentives for delegators to stay in the validator if they think winning chance is high enough for near future.

Now it has 5million atoms in delegation, becoming one of the top validators. Now delegators will gain only 0.2% if he/she wins. The incentives become less when the validator becomes top rank. So delegators will redelegate to another “rising tiger” who has less delegation but having good chance to win the contribution prize. It is naturally decentralization friendly.

We can design to have budget for this activity and pull taxes from rewards. Then, in total, delegators do not earn more. It is just a more efficient redistribution of rewards. Therefore delegators are not taking their free lunch. It is just changing the way how delegators choose their validators.

Now, most top validators are composed of highly contributing entities. Validators keep competing on better contribution for entire network. It is the most efficient way to spend delegators’ commissions. At the same time, because of the nature of decreasing incentives for delegation to higher power validators, those top validators will be quite well decentralized.

Also we can easily expect those contributing validators are mostly relatively community friendly, transparent, well communicative, and honest. Network becomes stronger and grows faster and gains more trust.

Many validators are building new services for their own delegators. What a waste?? Isnt it better for them to serve entire users and get more delegation from those contribution? Isnt it better for entire network?

Generally, business competitive entities are good for entire economy. But in blockchain ecosystem, it is little bit different. If too many validators are business profit oriented entities, it has better chance to cause inefficient equilibrium of the entire network. Ethan Buchman also explain this kind of situation with more elegant words. I hope more entities in Cosmos Network are less business profit oriented and I also hope those collaborative validators are supported in this network, supported not only by one time money, but also by delegators’ delegations naturally by designed incentives.

1 Like
4

there is tremendous merit to the proposal of rewarding contributors to the cosmos hub. perhaps this should not be limited to validator or even delegator contributions but even to other stakeholders?

on the other hand, are there ways to incentivize delegation to smaller validators solely based on, say the two slashing criteria, availability and double-signing, that are very easily quantifiable?

take uptime, for example. would it make sense to:

(a) set aside a certain percent of the funds that flow into the community pool during a certain period as additional rewards for the 20% smallest validators for reaching an aggressive xx.xx% availability for the same period. a lesser percentage of the funds could be rewarded to the next 10% smallest validators for attaining a slightly higher availability than the xx.xx%.

the rewards would redound to the group.

(b) inversely, the largest 10% validators would have their normal contributions to the community pool doubled (or increased by z%) if they fail to maintain an even higher yy.yy% availability.

i can definitely see how some may feel this wealth redistribution or progressive taxation may be no better than arbitrarily setting a minimum floor commission rate.

however, this is directly tied to network security without abstraction - availability. there are delegator incentives (towards smaller validators) and disincentives (for delegating to larger ones). and larger validators charging 0% commission may feel a heavier burden to stay at that level.

finally, the criteria can actually be algorithmically adjusted upwards as conditions warrant, as our entire set of validators gain experience and polish, further increasing security.

5

The thing is that in normal circumstances, 99% validators have +99.9% uptime. Validator service is becoming commoditized in normal circumstances and it is difficult to tell what happens when crisis comes.

So, double signing and uptime will not be a proper metric.

1 Like
6

you are correct, thank you.

perhaps straight forward availability may only have applicability on the incentives part for delegators to smaller validators?

7

This is a great concept. Well done, guys. We should definitely explore it in more detail.

  1. It really does solve the problem of how to incentivize community engagement. Building network effects is the most important thing we can all do to to ensure the success of the Hub.
  2. Competitions are always exciting. Having small number of teams who win every quarter is going to bring the best out of everyone. It will create a great buzz as everyone pitches their contribution and the votes come in.
  3. I like that this is token holder-centric. That they get the rewards but they have to put in some thought about how to earn them.
  4. It gives validators a way to compete that is more measurable than the current mechanism. Not all delegators can tell who is running the most secure infrastructure with the lowest risk of slashing. But they will be very good judges of the teams doing the most to build Cosmos network effects.
  5. I like the way it seeks to counter-balance the dangerous trend to 0% commission fees.
  6. And most of all is: this looks like an easily implementable variation of futarchy!

Looking forward to discussing this more. Great work Hyung and Dongsam!

1 Like
8

Thanks bharvest, this seems very promising.

I just find somenthing that may lead to the same problem we actually have. It is your point number 2. If we based it in a normal gov voting we will be facing the same stake centralization problem we are trying to solve.
I think, this kind of solutions would work better with a voting system not attached to the stake.

I also can not see how this great idea could help to solve the “network debt” for past contributions. It would not be fair to just declare new winners for new contribution and forget all our history until here, would it?

Anyway, congratulation!

Best regards

9

Thanks for this proposal.

I agree with the overall sentiment. Today, delegators primarily choose validators based on commission rate. Validators that contribute to the network do not get compensated for that, which will cause them to contribute less so as to generate the revenues elsewhere. I think this is a real trend and has a negative impact on Cosmos’ chance of success.

But I don’t think the proposed mechanism would work. The revenues would happen at unpredictable times. It will be difficult to convince delegators to stake with a higher commission validator, because they contribute and they might get some reward. It will have little affect on the behavior of delegators and just be a welcome bonus if it comes through.

It’s also perverse for the funds to go to delegators instead of validators. In a way, that exacerbates the current problem.

In my view, a more promising direction would be to pay out for network contribution on a quarterly basis to the top contributors. Who this is could be decided by governance. But it should go to those who actually contributed (and that can also be non-validators), not to the delegators of contributing validators.

Of course, this is not a simple system to build if it is to be fair and resistant to corruption. And if it is not to burden governance overly by having to make too many decisions. But it’s a direction that seems more promising to me than the one proposed.

10

One fact check.

It will have little affect on the behavior of delegators and just be a welcome bonus if it comes through.

The affect of the mechanism is decided by the size of the prize and the power of the validator. So, it is a quantitative. We can make it significantly effective by making the number big enough. It is always affectable. Especially, it is relatively more affectable for smaller validators.

From my example above, the prize can be up to 10% reward at once. Is that not enough? Even 1% is the 1 year commission which is big enough to affect.

Purpose of this mechanism is not only financing contributing validators, but more importantly, reorganizing the power distribution to more active participating validators, which will strengthen the security of the network significantly and also at the same time, achieving decentralization.

The revenues would happen at unpredictable times.

This is correct and intentional affect. The prize is decided by future event. It makes delegators to think about long term contribution of each validator.(Because prize distribution is ALSO related to the length of delegation : the royalty of the delegator to the validator. It is basically reputation based decision.)

Reward of contribution should not be immediate to validator, because it can make validator opportunistic. Instead, the system should convince validators to become long-term supportor of Cosmos Network. Indirect reward by bigger delegation is what I think the most efficient way to make top rank validators composed of most contributing validators.

For non-validator contributor, we have ICF, community fund to support. Support for contributing validator is a different category. Both should exist in my opinion. Amount of each part can be organized for best efficiency.

11

I agree that this is a pain point of any incentivized governance process. What I imagined to overcome this issue is the fractional distribution of prize to the delegators who correctly predict the winner. This incentive will incentivize delegators to vote, and also to vote to the validator who has the best chance to win the prize.(the prime advantage of the prediction market)

I also can not see how this great idea could help to solve the “network debt” for past contributions.

I also agree that the incentive is only focusing on future contribution. Maybe we can design to distribute the prize in longer period, for example like 6 months.(aka : prize vesting. If the delegation leaves the validator, all the rest prize will be burnt or return to the reserve) Then the winning delegators should be delegated to the validator for 6 months to fully appreciate the prize. Then, the delegation will remain in the winning validator longer.

12

Your response is hardly a “fact check”.

My point was a prediction about the behavior of delegators, which I do not believe would be substantially changed by your proposal.

Since it hasn’t been tried and we don’t have any data, we don’t know. Maybe you’d be right and it would be. But we’re both making guesses about future behavior.

13

As an extreme example, we can make universal rewards as zero, and whole rewards coming from prize. Then we cannot argue that it has no effect.

What I wanted to explain is that it is just a matter of how “much” would be enough to change the behavior. It is not about mechanism itself, but about how much quantitative the incentive is.

Also it is related to the power of validator because each prize is distributed to the delegators of a validator.

And my expectation was 10k atom per quarter and this number is arguable as you said. I expect 10k atom prize per quarter is enough to change the behavior, but it might not enough.

Although I think it is better than pay out validator directly, i think we can use hybrid approach, 50% prize 50% directly to validator.