GIP-0002: Rewards destination for indexer's indexing rewards

Yeah I do get that, but the indexers of today are not the indexers of tomorrow and keeping incentives aligned is a very, very difficult task.

Course correcting this early in the development of the protocol seems, IMHO unwise.

If this proposal were to pass I think the immediate proposal that comes next is why shouldn’t delegators be able to un delegate without the freezing/thaw period? Delegators experience a financial opportunity cost that is equivalent to the experience of indexers.

In my opinion, the financial opportunity cost, and the “delay” which prevents participant from reacting immediately to price fluctuations ensures that participants stay aligned on long term vision and goals.

If indexers are allowed to withdraw rewards immediately you will see them optimizing for price action rather than protocol health. By removing the ability to react to price action, you dampen the effect that price will have on decision making.

Additionally I don’t think the loan alternative is a bad one, access to capital is a normal business expense and in the case of indexer might actually provide a healthy proxy.

1 Like

This proposal is NOT “course correcting”, it’s attempting to fix something that was not intended to begin with.

Again, this is not about undelegating, it’s about rewards only. If indexers want to take out part of their stake, they WILL need to thaw for 28 days.

Again, this is not about “opportunity to dump”, it’s about taking out significant operational costs, by the actors who keep the network alive and take the risk

A normal business environment allows you to use your income for paying your expenses, which currently is not possible without harming delegators and the protocol.

2 Likes

@Koen Would you support Delegators getting their rewards immediately?

Delegators don’t have to pay thousands of dollars per month for allocations, server infrastructure and personnel.

3 Likes

But delegators DO have opportunity cost on their capital. Servers are expenses, it is true, but capital opportunity cost is enormous in a bull market, so delegators are taking a very similar cost/tradeoff ratio.

I personally would be more supportive of simply paying indexers more rather than unlocking that money early.

Also, I want to be clear, I’m not arguing to deprive indexers of rewards to cover their capital expenditures. I’m saying that the wider ecosystem will be very sensitive to monetary adjustments, and the success of The Graph network depends on a stable environment. I happen to believe that while today’s indexers are the best of the community- long term running an indexer will probably be abstracted to 3rd party staking organizations that optimize for-profit and cut deals with indexers and dapps to control query flow. Minimizing potentially net negative actors ability to arbitrage or find clever exploitative economic strategies should be the priority for long term success.

1 Like

Indexers have also the capital opportunity cost, seeing they need at least 100k stake.

If the question is, give delegators a similar option to forgo automatic redelegation and opt for immediate payout of rewards, depending on specifics, potentially yes.

Anything that’s actively open to receiving rewards, including for indexers, should go through 28 days thawing.

The thing with “paying indexers more” is we cannot currently withdraw those. Our only option would be undelegate our entire stake, wait 28 days and withdraw the surplus. Then optional spin the indexer up again and restake. THAT is what this proposal is addressing.

The “success and stability of the graph network” will equally rely on many indexers being able to stay in business. We’ve been paying out of pocket since July 2020 (start of testnet) and since December out of pocket for ETH gas. The reason for this proposal is to avoid that only big staking organisations can afford to keep their indexers afloat, to keep net positive actor able to be part of the network.

4 Likes

Hmm, I see.

If as you explain it, you are required to undelegate EVERYTHING to access your rewards, then I would be in favor of indexers being able to receive their rewards at a second address, as long as that separate address also had a lockup period. That makes sense to me. I also think the question of receiving rewards at a different address is a different proposal than reducing the thawing period to receive rewards.

EDIT: For staking facilities, 100K GRT is easy to acquire, especially now and I’m sure many have acquired that much it to potentially move into indexing in the future when it will be expensive to do.

1 Like

The entire thawing period should not be thrown out the window for indexer rewards, but rather a portion of indexing rewards let’s say 1/3 can be made available without having to go through the thawing period. Some may say this will help with offsetting operational cost but there will be indexers that will attempt to game the system and use the removal of the thawing process to cash out when the time seems right which in my opinion is not good

1 Like

I assume you mean thawing, the 28 days required for undelegating ?
(As “locked” is a “reserved word” with different meaning.)

1 Like

I disagree with any thawing period for rewards for Indexers in the method proposed. Having this thawing period would be a serious problem for tax reasons (at least in my country) where you are liable for this at the point at which the tokens are minted. Being able to set aside this at the time of closing allocations is important as if the tokens are available for transfer at a later date they may have lost value and we would be out of pocket without an opportunity to set aside a portion of their value for tax at generation.

Also can I remind contributors to the discussion here that this proposal is to solve a specific problem for Indexers with a token lock contract, as raised by Brandon in the original thread here.

If you have ideas around unlocking delegators rewards or anything unrelated to the proposal specified by @ari (which is an attempt to address the tokenlock issue) then it would be best to raise a suggestion in the feedback forum.

2 Likes

That would be even better! But I assume that ariel came up with the idea of a separate withdrawRewards() call to save gas, or because it simplifies the implementation, hence would make this change faster to deploy and less risky.

1 Like

I understand very well the desire of the indexers, delegators with vested contracts not to get access to their reward without the 28 thawing period but keep in mind like you spend your personal money to keep the network alive, the same way they trusted in the project and they’ve put significant amount of money too and they deserve to be able to get rewards in my humble opinion. Other thing is with some of the vested contracts when the time comes and depend on the contract there will be a problem when you get your first unlock you won’t be able to keep delegating if you want to move your unlock, so thats going to be another problem. In my opinion the changes should be made for all vested contracts not only for the indexers, they keep the network alive, but a lot of delegators with vested contracts believed in the Graph before it went on mainnnet and they also deserve their rewards.

P.S. The idea of thawing period is not to prevent delegators with vested contracts not to get their rewards, in the end the delgators should have the same rights with vested contract or not, at the current state there is difference between delegators and delegators with vested contracts, it makes sense to stay like that as far as all vested contracts has the same rights, but not if some of them don’t.

No changes are proposed to any token lock contracts.

If you wanted access to your GRT as it is unlocked by your vesting contract you should not have delegated the entirety of it (I assume this is what you have done).

1 Like

Bare with me, i am not tech guy, i just want to say if indexers have access to their rewards, like the delgators have, should be the same for delegators with vested contracts.

It sounds like you are wishing for a change to be made to all delegation behaviour, as the 28 day thaw period after undelegation to withdraw rewards from the network applies to all delegators (token lock contract or otherwise).

Delegators giving their 2 cents should bear in mind that while they share the same opportunity costs as Indexers on their collateral, they don’t bear the weight of ongoing expenditure opex for infra and gas. This proposal is an attempt to rebalance an already imbalanced situation that was not intended for mainnet in the first place.

That said, i wouldn’t mind if we also had to thaw our rewards, but that means additional gas and also the risk of price changes at the start/end point of thawing. Not to mention the tax-related issues already mentioned.

Delegators should consider, when giving their opinion, whether they want the costs of the above risks passing down and eating in to their own returns, because that’s what will happen. And at (currently) 21.5% of the network stake, the effect on the market from Indexers having this benefit would also be lessened, compared to allowing it for all. Not to mention already likely quite steady from some opex costs being naturally returned on an ongoing basis.

Let’s not forget that this is what the aim is here - to allow Indexers, especially smaller ones, to continue to function. And are people really opposed to the additional risks and efforts taken upon by Indexers having some kind of benefit returned anyway? If so, i’d remind you that anyone with enough stake could join team Indexer if they feel the odds are tipped in their favour at any point in time.

Incidentally, this would be a big plus for the network, so by all means, please consider it. :ok_hand:t2:

5 Likes

@ari and @davekaj,

In an attempt to bring this conversation back on topic, what are the next steps here? I suppose you will want to give some more time for those affected (Indexers) to chime in with their thoughts but assuming what you’ve suggested is generally accepted will it then be deployed on testnet?

1 Like

The thawing period for indexer’s stake is a mechanism not mean to manage liquidity in the protocol but a way to provide security to avoid an indexer easily avoid being slashed by removing the stake immediately before a dispute is accepted.

Koen, thank you for the feedback, I see your point.

These are the reason for the proposed route:

  1. When we added the operator key feature we thought it to be used to perform operations that do not take tokens out or in any of the protocol contracts. All the functions that can be called by operator just move tokens already deposited to different allocations, or frees them, while the main indexer key is used for token-external operations.

  2. We considered that pulling tokens out by calling a specific function like withdrawRewards() is what we see in most UX where you claim accumulated tokens from a protocol, it is easy to understand and provides a way to explicitly check and set the beneficiary before you send the transaction.

  3. I’ll do some calculations of gas efficiency, if you withdraw a number of accumulated rewards from multiple allocations it would mean just the cost of one more SSTORE.