1kx's Thoughts and Feedback on the 2024 Tokenomics Rework

Dear RP community,

The team at 1kx appreciates the community’s enthusiasm for enhancing Rocket Pool’s tokenomics and recognizes the significant efforts behind the current tokenomics proposal. We would like to suggest a slightly different approach, viewing our proposal not as a replacement but as an extension of the existing one.

Given the timelines involved, we have attempted to strike a balance between getting the proposal in front of the community as soon as possible, while also providing sufficient detail to allow for an informed comparison.

In brief, our primary disagreement with the existing rework proposal is that it prematurely abandons RPL’s primary value driver—the RPL collateral requirement for validators. We recognize that the proposed token rework contains other benefits, but we believe Houston’s new features can be better leveraged to further accelerate our main goal of maximizing rETH’s sustainable yield to position Rocket Pool as the most attractive LST in the space – we call this the “maxETH” strategy.

Introduction: “maxETH” <3 rETH

One main reason Rocket Pool has become so popular is because of its alignment with Ethereum’s core values. However, this merited gravitation towards decentralization can only attract so much TVL, and the average user’s preferences are driven by three realities: high yields, deep liquidity, and DeFi utility. Since we believe that the former leads to the latter two, our primary goal is to maximize the rETH’s sustainable APY as it is the single most important attribute for any LST.

The “maxETH” strategy has 3 main components:

  • Align rETH growth with NO growth by adding a dynamic global commission structure based on deposit pool utilization
  • Introduce RPL delegation to provide additional utility to RPL and allow stakers to select highly performant NOs from a Node Operator Performance Dashboard
  • Safely accelerate protocol growth and reduce MEV theft risk by accelerating Bonding Curves adoption

Proposed Changes in Brief

The core of our proposal is a delegated staking system, described in more detail below.

We propose to initially set the surplus_share to zero, increasing rETH APY.

1. Dynamic Global Commission

Capital in the deposit pool is idle and does not contribute to rETH APY. When the deposit pool is full, 18k of that ETH is not earning rewards, dragging down the average APY. Thus, aiming for maximum capital utilization —an empty deposit pool—is doubly important for rETH APY.

The protocol previously operated a dynamic commission rate that was set at the time of creation and applied for the lifetime of the validator. However, this approach led to gamification, as operators would skip a few days of rewards and wait for the deposit queue to fill up to lock in a 20% lifetime commission rate.

This had a negative impact on rETH’s growth and APY. There was slower growth due to NO waiting games and lower overall rETH APY due to overcompensation to NOs.

While gameable, this mechanism elegantly tied NO incentives to the protocol’s growth requirements, increasing rewards when necessary to attract new NOs.

We propose a new implementation of the dynamic commission with one important change: the variable commission rate applies to all validators regardless of the prevailing rate at the time of initialization.

This helps maximize rETH APY in two ways:

  • The protocol pays a higher commission only for as long as is necessary to attract new NOs
  • An individual NO has less incentive to wait for the higher commission rate

The prevailing commission rate would be set by a contract function. The function would calculate the desired position on the commission curve based on the available capacity of the deposit pool. We envision this function being executed by the oDAO, although it would be executable by anyone so long as there was sufficient time between invocations.

The rate change function would be similar to that of a DeFi lending pool, where the commission would spike if the deposit pool hit a certain size. Some illustrative values are:

Deposit Pool Utilization Commission Rate NO’s share Staker’s share
10% 5.9% 1.18% 4.72%
50% 9.5% 1.9% 7.6%
80% 12.2% 2.44% 9.76%

For those who want a deeper explanation, the dynamic commission rate on ETH staking rewards grows linearly with the deposit pool size and can be parametrized as follows:

d is the current deposit pool size in ETH, d_min and d_max the minimum/maximum values of those we consider (proposal: d_min=0 and d_max=18k) and c(d) the commission rate which is capped at c_max (proposal: 14%) and c_min (proposal: 5%).

This overall commission rate is split between i) the node operators (NO commission), ii) the RPL stakers or more specifically delegators (stake commission) and iii) a protocol share (surplus share). The motivation is similar to the ones outlined in the initial community proposal, yet we suggest a different dynamic:

i) The node operator commission share is initially a flat share of the overall commission rate and increases at a given kink point (k) of the deposit pool size, similar to a DeFi lending pool:

The node operator hence receives s_NO(d)*c(d) as commission from the ETH staking rewards. We propose a minimum share of 20%, a maximum of 70% and the kink at 80% if the maximum deposit pool size is reached (i.e. 14.4k ETH).

ii) The stake commission share is the share of ETH rewards that gets swapped to RPL and paid out to stakers that are delegated to node operators. We propose that the initial stake commission rate equals the total commission minus the node operator commission (see i). See next for our rationale.

iii) The protocol surplus share: we propose to leave this share at 0 for now for two main reasons: first, we want to maximize rETH yield by lowering the overall commission rate. Second, as mentioned in our previous forum reply, the protocol’s current TVL is not high enough to justify any sort of value capture. If Rocket Pool were to use the surplus share, we have a recommendation that we reference in point #6.

With that, the proposed commission share and its distribution is as follows:

The node operator commission share we propose (20-70%, blue area) is, on average, 3.2% when considering all deposit pool sizes, which is only slightly lower than the initially suggested 3.5% figure, as shown by the gray line. If the deposit pool is empty, the total commission will equate to ~5%, which boosts rETH’s APY by a notable 10% compared to the current product. A theoretically empty deposit pool would allow rETH to surpass stETH’s yield by a full 15 bps, which massively incentivizes more TVL for Rocket Pool. If the deposit pool is full, we’re temporarily over-incentivizing NOs to spin up new validators with higher commissions, ensuring our overall protocol efficiency stays high. This mechanism is a very powerful feedback loop that can lead to sustainable TVL growth.

Combined with some assumption of gains for performance in general based on the research by ArtDemocraft, we could see rETH’s yield pass stETH over the long run, or at the very least a significant reduction in the current yield gap between stETH and rETH. This would make rETH one of the best-yielding LSTs:

For RPL stakers, the actual APY depends largely on the total amount of RPL staked and the price of RPL in terms of ETH. We fund the RPL staking rewards by purchasing RPL with earned ETH staking reward commissions. Thus, if the RPL price increases vs. ETH, then the RPL staking yield decreases and vice versa. This is true for the original proposal as well. Hence, the following assumes the current price relation of RPL and ETH. If the proposed minimum of 150 RPL per validator is staked, we end up with about 25% of the total RPL staked and consequently the highest yield. The peak income for stakers is at the kink of our commission split between node operators and stakers at 14.4k ETH in the deposit pool (see chart above). At that level and minimum stake ratio, stakers would earn over 10% APR:

This is almost double the yield for stakers compared to the original proposal, represented by the orange line. Please click here for more simulation data and associated assumptions above.

This change would introduce the following protocol parameters:

min_commission Minimum commission rate
max_commission Maximum commission rate
commission_change_min_interval Minimum time that must pass between changes to commission rate, in seconds.
commision_relative_step Maximum amount that commission can be changed per update, expressed as a percentage of current rate

The introduction of a dynamic global commission will contribute towards the maxETH initiative by minimizing protocol value capture, optimizing for node operator and staker payouts at varying deposit pool sizes, and prioritizing rETH yield payouts at all times.

2. RPL Delegation

Key to our proposal is the introduction of delegated staking.

Houston introduced the concept of “whale marriages”, whereby one party can provide the RPL while another provides the ETH. This has previously been possible through the use of splitter contracts but technical complexity and poor UX led to limited uptake.

Like many in the ecosystem, we believe this is of paramount importance to unlocking NO growth as it allows the protocol to onboard NOs who do not want to take on RPL exposure.

We propose to introduce delegated staking, effectively a matchmaker for “whale marriages”. This provides a Schelling point where NOs can compete to attract delegators, and RPL delegators can make informed decisions about their delegations.

The goal is to eliminate the friction of pairing these counterparties, unlocking the full potential of Houston.

Should our proposal pass, we would follow up with a future proposal to integrate the existing penalty research into the delegation system, further increasing the incentive for NOs to perform and for delegators to choose performant NOs. This involves making NOs (and their delegators) liable for the cost of poor performance, ensuring the rETH APY approaches its theoretical maximum.

To mitigate centralization risk a per-NO delegation cap would be applied, preventing any single NO from accumulating too much voting power (e.g. a cap of 1% would ensure that no single NO could have more than 1% of RPL supply delegated to them).

In the future, we could also introduce the idea of separating reward delegation from voting delegation. An RPL staker could receive rewards from the NO that provides the highest yield, while still delegating their voting power to a community member who supports their ideals.

Improving Performance

Among notable ETH staking protocols, Rocket Pool currently has the lowest performance on Rated Network, falling behind players such as Swell, EtherFi, Ankr, Lido, Stader, and Stakewise. While each missed attestation is relatively small, the aggregate effect of the missed attestations drags down rETH’s APY.

We believe this can be corrected with a market design that incentivizes delegation to operators with strong performance, which in turn can reduce missed rewards and help increase rETH APY. Based on the assumptions in ArtDemocraft’s research, we could optimistically expect to see a potential APY increase of up to 0.07% simply due to performance improvements amongst the largest struggling NOs.

To do this, we need to introduce a forcing function that encourages NOs to improve performance. The best way to do this is to create a direct link between performance and rewards.

As part of our proposal, a portion of the NO’s rewards will be shared with their delegators. The higher the NOs rewards (i.e. the fewer missed attestations), the more rewards received by their delegators. The more delegators an NO has, the more collateral they have for launching additional validators. Thus we introduce a positive feedback loop where operators with the best performance will likely attract more rewards. This introduces competition between NOs to miss the fewest attestations, which will increase the total amount of rewards, which will increase rETH’s APY.

To allow RPL holders to make informed decisions when choosing delegates, we will introduce a Node Operator Performance Dashboard (NOPD). This will be built on publicly available data so we might see multiple NOPDs deployed by different parties.

The variable nature of MEV makes it challenging to compare validator earnings side by side. We, therefore, propose making participation in the smoothing pool mandatory for those opting into delegation, reducing the inherent variability in rewards.

Democratizing Whale Marriages

Under the current system, Whale Marriages have some centralization risk, benefiting large RPL holders who can strike deals with node operators at scale. Our delegation system democratizes the process of providing RPL for delegation and gives smaller RPL holders just as much chance to take part as larger holders.

It would be a lot of work for a polygamous NO to individually coordinate with 10,000 delegators providing 10 RPL each. This gives an advantage to the whale with 100k RPL, who would find it easier to match with a “marriage partner.” We believe that larger holders should not have an unfair advantage merely because they have more capital, and leveling the playing field provides more support for the Charter goal of prioritizing decentralization.

By removing the friction of matching RPL and ETH counterparties through a simple dashboard, we can rapidly accelerate the rate at which ETH can be deployed, further accelerating rETH’s APY growth.

Reducing Risk of MEV-theft

In addition to the above benefits, implementing delegation allows us to further reduce both the probability and impact of MEV theft. Delegated RPL provides additional capital which can be slashed to a) punish NOs, and b) make the protocol whole after MEV theft.

As noted in the Bond Curves RPIP, the current proposal does not fully protect against MEV theft. While the nature of MEV makes it challenging to achieve full protection, we believe we can improve upon the current proposal.

In the current proposal, if a NO with a single 8 ETH minipool steals 20 ETH of MEV, rETH APY is lower than it should be.

As part of the delegation system, delegators will stake their RPL against a NO’s megapool contract. This provides additional capital which can be used to slash NOs who steal MEV. This has three benefits:

  • The cost of MEV theft is increased, making it less likely to occur
  • If MEV theft does occur, the protocol has a higher chance of recovering the stolen funds
  • We do not need to force-exit the NO in order to recover the capital - we can immediately slash the staked RPL (and still force-exit the NO if more collateral is required, and might wish to do so as punishment)

RPIP-42 notes that:

> “We willingly take on somewhat more MEV-theft risk for the smallest NOs”

Our proposal allows the smallest NOs to attract delegated RPL, strengthening the anti-MEV-theft provisions. Thus, we provide further support for this goal without taking on additional MEV theft risk.

With additional RPL collateral available for slashing misbehaving or poorly performing NOs, and sufficient buy-side demand for RPL, we believe the Bond Curves (RPIP-42) work could potentially be safely accelerated.

Delegation Parameters

Delegation Parameters

The incentive to delegate RPL to the top-performing operators maintains RPL’s core utility and enhances the protocol’s overall validator rewards. This, in turn, leads to higher rETH yields, advancing the maxETH initiative.

3. Updates to RPIP-42: Bond Curve Proposal

Currently, node operators need to put up a set amount of ETH per validator (e.g. 8 ETH collateral, 24 ETH borrowed) when launching validators. Also as suggested within RPIP-42, there is general alignment within the community that from a risk/reward perspective, lower amounts of collateral for incrementally borrowed ETH should be experimented with. We believe that the amount of reasonable collateral needed to slash an operator for misbehavior does not scale linearly with the amount of ETH they borrow from the deposit pool. For example, if they operate one validator, an example amount of collateral could be 4 ETH, but if they operate 8, there might only need to be 2 ETH per validator. The exact sublinear formula can be debated and explored at a later date.

We propose that depending on the total number of validators for a given node operator v, the total collateral requirement in ETH C(v) is given by: C(v) = C(1)*v^p

Using C(1)=4 ETH and p=⅔:

Currently, about 35% of nodes run one validator, so for them there would be no change in collateral if they run a minipool. However, for 20% of node operators who operate 8+ validators (in aggregate that is 80% of the validators), the reduction is 50% and potentially more:

The left axis/blue line shows 1 minus the ratio of proposed collateral needed per operator (operating v validators on the x-axis) divided by the current collateral needed when assuming 4ETH per validator. The right axis/gray bars show the current distribution, e.g. about 1200 operators running one validator. The conclusion is that compared to the world of vanilla LEB4, relatively less collateral is needed as the number of validators per node increases.

With slight updates to RPIP, we can efficiently lower the collateral requirement for larger node operators while reducing MEV-theft risk. This allows the protocol to support more demand at scale and ensure that the deposit pool remains empty, supporting maxETH’s main goal of maximizing rETH yields.

4. Optimize Deposit Queue Access for rETH APY

We expect that the Houston upgrade, accelerated by our delegation proposal, will lead to significant increases in deposit pool utilization. This raises the question of how to handle an empty deposit queue.

To address this, we propose prioritizing access to the queue based on the total RPL staked by a node operator (including delegated RPL). This approach ensures that those most committed to the protocol’s long-term health are given the first opportunity to serve the community. Additionally, it enhances protocol security by directing ETH towards operators who can provide the strongest security guarantees, as indicated by their slashable collateral.

To prevent centralizing forces in the deposit queue, we propose a mechanism where queue slots are reserved for new NOs. Additional protection avenues, such as limiting the number of slots per NO within a given timeframe, can be explored as part of further community discussion.

5. RPL Collateral for Node Operators

Given the newly proposed RPL delegation system, we also propose some tweaks to the current RPL collateral design. A fixed amount of 150 RPL will be required per validator, and this value will be static until the community or pDAO votes for a change.

Whilst the specific value is open to debate, our rationale for 150 RPL as the initial value is twofold:

  • At current price levels, this adds approximately 1 ETH in collateral value per validator. This significantly increases collateralization for node operators running multiple validators and counter-balances the reduction in ETH collateral per incremental validator proposed in Bond Curves (RPIP-42). The result is maintained overall network security.
  • Taking the current validator count, the theoretical required RPL stake of 150 equates to around 25% of circulating supply. This minimum threshold is a 50% reduction versus the current stake ratio of 50% coming from the current RPL staking requirement, but we believe our new RPL yield mechanism attracts staking beyond this new minimum of 150, and hence the drop in RPL staked under our new proposal in practice will likely be much less than 50%. Please click this here if you’d like to see some simulations we made around % RPL staked, RPL price, and more.

Static RPL values make it easier for both delegators and operators to track. The amount of RPL required per validator will not change regardless of how much a single node operator earns. If a node operator becomes undercollateralized, for example, when it does not have enough delegated RPL for the number of validators, there will not be a forced exit.

However, while a NO is undercollateralized, the full delegate_share is used to purchase RPL on the open market and stake it under the ownership of the pDAO. Once the node is sufficiently collateralized, the delegate_share reverts to the NO’s delegates. The pDAO continues to own the purchased RPL and benefit from its yield. A NO can not launch new validators while undercollateralized. One downside of an RPL delegation system is the elevated pressure on delegators to monitor their positions as they may temporarily be forfeiting their share to the pDAO.

The primary motivation for choosing a static value is to simplify life for NOs and delegators and alleviate the burden of constantly checking collateralization levels. An alternative approach would be to retain the mechanism where the required RPL is based on the value of ETH borrowed from the protocol. We are open to discussing this, as it has advantages and disadvantages, but believe the simpler approach is more suitable for this entire proposal.

A fixed amount of required RPL collateral minimizes the risk of prolonged periods where node operators lack sufficient delegations. This ensures a consistently optimal delegation market, allowing delegators to earn yields with the highest-performing node operators, thereby supporting maxETH’s goal of increasing rETH yields.

6. Introduce protocol-owned yield as part of the surplus share

We believe protocol commissions should, at present, only go to node operators and RPL stakers. However, in the future, one alternative value capture mechanism that is to be explored is the introduction of a protocol-owned staking pool where pDAO stakes the respective surplus share to earn yield on behalf of the protocol’s newly owned ETH assets. The subsequent yield is then distributed to rETH holders. This staking pool would be equivalent to the “surplus_share” recipient when referencing current proposals. The pDAO, guided by RPL voters, could select operators with less market share or more decentralized infrastructure and therefore counteract any potential centralizing forces that might result from crowding around the most performant NOs. This new scheme potentially strengthens rETH’s long-term APY, the primary goal of MaxETH, while simultaneously building up protocol-owned assets.

RPIPs - new and modified

Implementing this proposal would require two additional RPIPs, and modifications to multiple existing draft RPIPs. 1kx would provide the initial version of the new RPIPs and submit them to the community for discussion. We would also work with the authors of the existing proposals to integrate our changes in a way that minimizes the impact and allows us to continue along the previously accepted timelines.

New RPIPs

Below is a high-level overview of the modifications we propose to the existing RPIPs:


RPIP Changes (Pt 2)

Personas

Personas (Pt 1)
Personas (Pt 2)

Concluding Thoughts

The maxETH initiative intends to increase rETH’s yield, making it the most desirable liquid staking product on the market. Meanwhile, we also aim to preserve the RPL collateral requirement, address the deposit queue buildup, and accelerate TVL growth, all of which are key to Rocket Pool’s future success. We believe that by modifying the existing rework to take full advantage of the Houston upgrade, one of the most exciting in the protocol’s history, we can create the most exciting liquid staking product on the market.

We appreciate your engagement and consideration of our post. We look forward to the community’s feedback and collaboration on refining and enhancing the current RPIPs. Your input is invaluable in shaping our collective future.

Disclaimers

Disclaimer: Rocket Pool is a 1kx portfolio investment.

Disclaimer: This article is for general information purposes only and should not be construed as or relied upon in any manner as investment, financial, legal, regulatory, tax, accounting, or similar advice. Under no circumstances should any material at the site be used or be construed as an offer soliciting the purchase or sale of any security, future, or other financial product or instrument. Views expressed in posts are those of the individual 1kx personnel quoted therein and are not the views of 1kx and are subject to change. The posts are not directed to any investors or potential investors, and do not constitute an offer to sell or a solicitation of an offer to buy any securities, and may not be used or relied upon in evaluating the merits of any investment. All information contained herein should be independently verified and confirmed. 1kx does not accept any liability for any loss or damage whatsoever caused in reliance upon such information. Certain information has been obtained from third-party sources. While taken from sources believed to be reliable, 1kx has not independently verified such information and makes no representations about the enduring accuracy or completeness of any information provided or its appropriateness for a given situation. 1kx may hold positions in certain projects or assets discussed in this article.

8 Likes

Thanks for the detailed proposal! It’s going to take a while to understand this in aggregate, but I’ll share the first things that stood out to me

No ETH-only

The main part of the current tokenomics proposals is to grow through eth-only minipools, but this proposal retains RPL collateral requirements through whale marriages. Gut reaction is, if I’m a potential NO, an eth-only node is a significantly lower hurdle than entering an “RPL delegator market” to attract RPL. I.e. I think the the current proposal would grow the protocol faster via eth-only NOs than this proposal’s fixed RPL collateral + marriages.

Obvious centralizing force

Introduce RPL delegation to provide additional utility to RPL and allow stakers to select highly performant NOs from a Node Operator Performance Dashboard

This would obviously favor larger entities and hurt home stakers. It doesn’t feel very “Rocket Pool”. The potential benefit is stated later as “up to .07%”. I think 0.07% is a reasonable cost to NOT shift rewards from home stakers to professional staking operations.

It’s also complicated. Later we deal with the flows from this 0.07%

…this staking pool would be equivalent to the “surplus_share” recipient when referencing current proposals. The pDAO, guided by RPL voters, could select operators with less market share or more decentralized infrastructure and therefore counteract any potential centralizing forces that might result from crowding around the most performant NOs

The burning question

I appreciate the disclosure you’re an investor in the project. To me, this proposal feels like an effort to limit any downside risk to RPL (through retaining fixed collateral instead of eth-only), rather than an attempt to maximize the growth of the protocol. Not sure how you can best address that concern / impression, but wanted to share it.

The dynamic stuff

I liked the introduction of existing functions in DeFi to set some of the system parameters. The current proposals punt that to a future iteration. I wonder if it’s worth considering them based on this proposal and the 1kx team’s experience.

Thanks!

2 Likes

I don’t think this idea is close to functional at this time.
I’m going to address what I think is the biggest issue. Then I’ll have a section hitting some secondary ones. This is far from complete - I’m trying to focus on things that I consider untenable conceptually (ie, not just tuning numbers).

The dynamic global commission proposed forces linkages that do not exist in the market. Said otherwise, they are trying to tell the market what to think. As we’ve already found out once, telling the market what to do doesn’t work; sooner or later it’ll get the last word.

  • As deposit pool is fuller, commission increases
    • This part is directionally sensible, but misses the timescales relevant to each party.
    • The slope here may be steep enough to cause perverse incentives. Ie, if I’m a large rETH holder I may choose to change some rETH to stETH in order to decrease the deposit pool size and get better commission. This is desirable at the extreme of “very full”, but not at the low end; however, the slope is constant which means this behavior is equally incentivized at 500ETH in the DP as 18k. Also – this incentive actually goes away at the very top end for some reason? Ie, the case where a rETH holder is least incentivized to burn some rETH is… when we have the most overdemand.
  • As deposit pool gets fuller, RPL gets more revenue. Why??? A full DP means high rETH demand, which requires supply. RPL doesn’t provide that.
    • When the deposit pool gets very full, NO share rapidly increases. This acknowledges the need for supply. But it also introduces perverse incentives.
    • Large NOs may be able to earn more by exiting validators when we’re near a supply crunch. Imagine we’re at 14k ETH, with NO commission around 2.5%. We’ll assume a large NO with many 2-ETH-bond validators. They can exit ~130 validators to spike their commission to ~9.5%. An entity or group of entities running about 2k validators would net-gain by exiting validators in this situation. This would also, ofc, damage the protocol by creating a supply crunch.
    • Similar tricks can also be played if any NOs hold other ETHlike position. For example a wstETH/ETH LP might be a position they like for some funds due to liquidity – it may make sense to simply sell that position to buy rETH if they’d take advantage of the large kink to ~4x their commission.

This is the biggest issue. There is an attempt to create incentives in a way that warps the market. I have spent not-even-10 minutes thinking of issues and came up with 3 separate perverse incentives. I think anything that enmeshes incentives will suffer this way.

Secondary thoughts
  • RPL delegation. We don’t have a way to identify entities on-chain, so “caps” are just “incentives to use sock puppet nodes”. Take a look at SSV Network Explorer to see transparent sock puppeting in action (note that their centralization isn’t special - Cosmos validators end up similar, eg). And yes, you can crack down on that… which incentivizes covert sock puppets. Each step gets worse.
  • RPL delegation in general. Essentially, this means folks with a known name or the ability to market have access to an ETH-only product, but random solo homestakers (some of the most ethereum-aligned folks possible) will not be granted access to an ETH-only product. That seems quite wrong.
  • Improving performance X delegation. I love artdemocrat’s general theme of providing for rETH restitution or similar. Note that this type of fix can be achieved without delegation. I have always been particularly strong in the belief that we must not simply scale with performance, instead we should crack down on outlier-bad cases, but not reward “home staker” differently than “high-availability professional staker”. This creates a direct incentive to use the highest performer and ignore everything else (ie, an allnodes-run node should expect delegation before a home staker).
  • Reducing Risk of MEV-theft. This section is deeply confused. Three benefits are listed, and I agree with the second one (funds would be more recoverable with an additional forced collateral source). That said, the other 2 benefits are backwards. The cost of MEV theft is not increased by having delegates slashed. It is decreased by protecting the thieving NO’s ETH in favor of the delegators RPL. If we don’t force exit, the NO can continue to steal or otherwise act with impunity.
  • RPIP-42 already proposes a sublinear bond curve, going down to 1.5 ETH per validator.
  • Flat RPL collateral per validator. There’s currently ~20.5M RPL. At 150 per validator, that would be a hard maximum of 136.7k validators, which is 13% of the current marketshare (assuming 100% is actively used). Quite an odd stance.
  • Flat RPL collateral per validator. If 1 ETH per validator is sensible (current 150 RPL), why would it also be sensible to require 4.5 ETH per validator (ATH ratio 150 RPL)? Why would it be sensible to tolerate 0.2 ETH per validator if price crashes? This number seems unrelated to anything. This is because it’s no longer the source of value capture (as it tries to be in current tokenomics), and it’s also not a primary bond.
  • Undercollateralization. This hasn’t been thought through. Can delegates undelegate? If not, that’s a huge red flag cuz people are stuck indefinitely. If so, why would they opt to receive a zero share? I expect they would instead completely leave that NO and delegate with someone that isn’t undercollateralized. Ie, if the price of RPL falls and an operator becomes slightly undercollateralized, they can expect to immediately lose all delegation.
1 Like

General

Thanks for sharing the ideas! I think much of the delegation you described can work as projects on top of Rocket Pool (to earn “Voter Share” rewards) without requiring integration into the core tokenomics rework. For example, someone building a “Node Operator Performance Dashboard” with publicly available data could apply for a grant from the GMC. Projects like NodeSet or RocketLend can provide healthy markets, and someone could also work on a project to specifically assist in democratizing marriages through focusing on delegation of RPL.

Some other comments/feedback on the rest of the proposal below:

1. UARS vs Dynamic Global Commission

  • Two key points about UARS is that the variables are universal and adjustable, so they already are designed to function in the same way as what you call global and dynamic. With UARS we are not “locked in” to any values, and we can adjust to market conditions as needed. There is an active thread https://discord.com/channels/405159462932971535/1228755310794113036 where we are discussing what UARS values we expect to need at equilibrium. Even among 3-5 people there is a wide variety of what we expect to be required to find an equilibrium. Instead of trying to predetermine a function to handle all possible variable amounts, I think we should just listen to the market to determine what values will create a healthy equilibrium once we have all the relevant information in the future to make that decision. In that sense it can be simple: do we have a consistently high deposit pool? We can vote to increase the NO share (and decrease surplus share) to incentivize supply to meet that demand.
  • Related to “listening to the market”, Rocket Pool will need to provide a competitive product for Node Operators. If we are only offering Node Operators 1% commission at an empty deposit pool, how do we expect to compete with something like Lido CSM that offers 7.5% commission? It seems with the function you described you will end up consistently struggling to incentivize Node Operators to join until/unless the deposit pool is almost full and the yields shoot up closer to the competitive landscape at 7.5%. But this seems to be the scenario you were trying to avoid (commission for rETH also shoots up, and now we have a less attractive product again)?
  • I think the wrong variables seem forced to be “static”, and the variables that should ideally be more static are made to be more “dynamic”.
    • Picking a static 150 RPL seems confusing, the RPL/ETH ratio can be highly volatile so it’s possible in the future this will be a massive barrier to growth, or it could be a negligible requirement (and what’s the point in keeping it then?)
    • Picking a dynamic total rETH fee seems like it would make marketing/communication difficult. Sometimes rETH commission is only 5%, but then other times it can swing up to 14%. An investor may choose to buy rETH after reading the commission is 5%, and then be surprised to learn later that it jumped up to 14% the majority of the time they held it (unless they had studied the intricacies of Rocket Pool and how the Deposit Pool impacts commission/yields)

2. Delegation and staker’s share:

  • You proposed a “Staker’s Share” and I noticed one difference from the existing proposal’s “Voter Share” is you suggested paying in RPL instead of ETH (using the ETH to buy RPL first and then distributing the RPL)? I was curious what do you see as the benefit of first converting to RPL before distribution (vs just distributing ETH)?
    • The existing proposal could consider paying out “voter share” in RPL if there is reasonable justification for that (for example, if buy+burn option is chosen then all “RPL value capture” could first convert ETH revenue to RPL, and then “surplus share” would get burnt and “voter share” would go to staked RPL). This seems like a relatively minor change in terms of impact though, and I’m not sure what the benefit would be.
  • Another difference seems to be rewards are not socialized? (They are paid out per validator?). This, combined with the RPL staking requirement seems like it could lead to weird outcomes:
    • All validators require some minimum RPL to start, so any RPL “delegated” to a node after initialization effectively dilutes all existing RPL staked to that node. This incentivizes max RPL yield at the minimum RPL staked amount.
    • A Node Operator may have RPL delegated to them which allowed them to create more validators (at close to the minimum RPL staked amount). If one of the RPL stakers decides they are ready to unstake and remove their delegation, this could “undercollateralize” all the rest of the staked RPL on that node for all of the validators. This could lead to a “musical chairs” environment where everyone is constantly being forced to readjust their delegations and positions to maintain their RPL yield, having little to do with the actual Node Operators underneath.
    • You mentioned this could happen, and that the pDAO would earn surplus share yield in the meantime. You didn’t mention what happens to the NO commission, it continues to earn like normal?… I don’t think this is a viable solution. A node operator would continue to earn their “NO commission” share but any “undercollateralized RPL” on the node becomes unproductive. The rational thing to do then would be to delegate any remaining RPL to a different node that isn’t undercollateralized where it can become productive again.
    • This leads to the most broken outcome… Take the same staked RPL: delegate to a node, create new validators, unstake it (now undercollateralized), continue running ETH-only validators, rinse/repeat, (and now strangely most of the revenue ends up going to the pDAO…)
    • I think your proposal also suggests this pDAO revenue should be given back to rETH holders? That seems like an unnecessary extra step (vs just charging a lower fee on rETH). This is a worse option for rETH holders since the protocol first used the ETH revenue to buy RPL instead of the entire ETH value being directed to rETH holders.

We propose that the initial stake commission rate equals the total commission minus the node operator commission… surplus share: we propose to leave this share at 0”.

  • I just wanted to reiterate that “surplus revenue to voter share” – was one of the potential options suggested in the existing proposal, and could very well be how the surplus revenue decision lands. The “vibe check” vote showed a pretty even distribution of preference for the surplus revenue decision, so it would seem aligned with your goals to lobby for the “voter share” option.

3. Bond Curves:

The exact sublinear formula can be debated and explored at a later date.

  • There is already a sublinear bond curve in the existing proposal, do you have any specific research or justification for changing it to the formula you described?
  • Generally we discussed “rETH protection” as specifically punishing outlier-bad performing node operators (instead of trying to reward the best performance). Rocket Pool is a stalwart of decentralization for Ethereum, and I think our diverse community of home stakers is our greatest strength. Home stakers often do not have the best attestation performance, but they can certainly avoid being “outlier bad” which tend to be the largest source of drag on the protocol. We decided to postpone “rETH protection” for a future upgrade instead of including it in the core tokenomics rework since there is already a high volume of changes for the rework, but maybe this could be prioritized for the next upgrade after Saturn.
  • Generally there is a principal agent problem if we want to ensure the Node Operator performs well but we focus on punishing/rewarding them through “delegated” RPL. Under the existing proposal we allow ETH-only node operators, so we have to ensure sufficient collateral with ETH-only. Maybe in the future RPL could still be used as a limited form of collateral for bond reductions/rETH protection though, since it does still have value that the protocol could use to recover lost yield from underperformance/theft.
2 Likes

It seems like our main and most important disagreement is the RPL collateral requirement; this should take discussion priority above else. We have provided some thoughts below:

Tl;dr: We should not remove the RPL collateral requirement unless it is necessary, and it is too close to the Houston release to know if it is necessary yet

Our issues with the current proposal relate to a small number of important issues. While we share many of the views expressed in the “Tokenomics Explainers” and supporting materials, we do not agree with the final conclusions.

The Why Rework document outlines three core ideals:

Ideal 1 - RPL stake serves as an entry ticket to ETH commission via the minimum staking requirement to start a minipool (10% of borrowed ETH)

Ideal 2 - RPL rewards are attractive enough that node operators want to maintain this minimum staking level, so we expect them to stay at or above the minimum by adding RPL as needed

Ideal 3 - In the long run, the fundamental value of RPL trends to what’s needed for all minipools to hit this minimum staking level (ie, RPL market cap should be ~10% of rETH TVL)

The document concludes that because ideals 2 and 3 did not play out in reality, we should abandon the ideal of “RPL as an entry ticket to ETH commissions”. In our view this is an incorrect - or at least premature - conclusion.

The rework research has identified a number of potential reasons for this. One reason is that as RPL/ETH performance continued to decline, it became more expensive for NOs to maintain the minimum collateral level. Without RPL delegation, the only option was for NOs to buy or borrow the RPL. There was no trustless way for a NO to allow third parties to contribute their RPL towards that NO’s collateral maintenance (without using splitter contracts and trusting the other party).

Post-Houston, this is no longer the case. Whale marriages make it possible for a NO to maintain this collateral requirement without shouldering the entire financial burden of doing so. This is an improvement on the previous situation but we need to go further still: enshrine the concept of delegation into the protocol and eliminate the friction of collateral maintenance.

So far, the only options for an undercollateralized NO to launch new validators were to purchase RPL or exit minipools to reduce borrowed ETH and reduce their collateralization level.

Our proposal introduces a third option for attracting RPL delegates. This avoids the need to exit validators, which is good for rETH TVL and supports our Charter goal of prioritizing the health of the Ethereum network.

From its inception, RPL has been required as collateral. Changing this would significantly change the social contract between RPL holders and other ecosystem participants. The Charter specifically states that our goal is to serve the entire Rocket Pool community, which includes RPL holders (many of whom are also NOs).

The Houston upgrade already facilitates the primary goal of ETH-only pools (allowing NOs to onboard without RPL exposure). However, this change was only recently applied. As such, all of the analysis on whether we need ETH-only pools was performed on pre-Houston data.

When changing a complex system it is important to limit the number of changes made simultaneously and to allow time for the impact of each change to be sufficiently analyzed.

Introducing a second method for NOs to onboard without RPL exposure before we have had a chance to adequately assess its impact is premature.

RPL as an Entry Ticket

The main difference between our proposal and the existing proposal is the retention of the RPL collateral requirement, so we feel it is important to share our thinking on this point.

The Rework Intro document provides insight into the benefits of being a NO versus solo staking:

8 ETH bond: 10.5% boost compared to solo staking rewards.

4 ETH bond: 24.5% boost compared to solo staking rewards.

1.5 ETH bond: 71.2% boost compared to solo staking rewards.

Given those numbers, it is clear that NOs have a significant financial advantage over solo stakers. We believe it is equally clear that potential NOs would be willing to pay for an entry ticket that allows them to receive those boosted yields.

Consider this thought experiment: imagine there was a Validator Boost NFT for sale. If a solo staker buys and burns this NFT they get a lifetime 10.5% increase in their validator rewards. This scales linearly - if they have two validators they can buy two NFTs and earn boosted rewards on both. There is a limited supply of these NFTs and no more will be minted.

How much would this NFT be worth? A validator who expects to earn 1 ETH over the lifetime of their validator would stand to gain 0.105 ETH by using it.

Now consider a slightly different scenario: instead of burning the Validator Boost NFT the staker can resell it after they have exited, and another validator can use it. It seems likely that the resellable NFT would be valued more highly than the burnable NFT. How much more would depend on how the buyer perceives their ability to sell the NFT in the future.

What can we say about the value of these hypothetical NFTs?

  • The true value of the NFTs is somewhere between 0 and the amount of extra ETH received by the validator
  • Because the boost applies for the lifetime of the validator the NFT is worth more to a long-term staker than a short-term staker
  • Short-term stakers - who would sell the NFT before recouping its cost - would be subjecting themselves to more market risk than long-term stakers
  • The higher the perceived future value of the NFT, the lower the perceived Total Cost of Ownership
  • The number of people willing to buy the NFT with >0 resale value will be larger than the number willing to buy the one with 0 resale value
  • The true value of the NFT can only be expressed in ETH because the conferred benefit is measured in ETH
  • An NFT with the 71.2% boost trait would be more valuable than a 10.5% boost trait

What conclusions can be drawn from this thought experiment about RPL?

  • RPL has measurable value as an entry ticket, and giving this away for free is not economically rational
  • A staker with longer time horizons will perceive RPL to have a lower cost. Those most willing to pay for the entry ticket are those who will most benefit the health of the network (reduced validator churn)
  • A potential NO is more likely to buy RPL as an entry ticket if they believe they can resell it in the future, so providing buy-side demand could actively accelerate network growth
  • Collateral requirements should be denominated in ETH, as that is the ROI metric potential buyers will use

If RPL has value, it can serve as collateral to improve protocol safety (e.g., by increasing the cost of MEV theft). However, this mechanism only functions if RPL has sufficient buy-side demand. By abandoning the historic primary value driver of RPL as an entry ticket, we limit the ability to rely on RPL as collateral.

For these reasons, we believe that introducing ETH-only validators would be tantamount to giving away one of the protocol’s most valuable resources—access to protocol ETH—for free.

The ability to borrow protocol ETH is a privilege, and we believe it is a privilege worth paying for.

Disclosure: Rocket Pool is a 1kx portfolio investment. 1kx also operates validators on the Rocket Pool network.

3 Likes

Hi Mikey,

I wanted to show some support for your / 1kx contributions here. There are some novel ideas here, that I think deserve deep consideration. I want to specifically +1 your contention that Rocket Pool should not entirely remove the requirement to stake RPL at this time. I suggested the same when I wrote about the tokenomics changes back in March. Note: the proposals have changed somewhat since then so my writeup isn’t up to date, but the broad strokes remain.

I’d need a bit more time to go over in more detail whether I agree with some particular items. I think Samus mentioned this, but a flat number of RPL staked seems a bit of an odd choice. Might an entry ticket of a particular percentage of protocol ETH make more sense? Seems like you could more easily reason about its relation to future rewards.

I agree that we should let Houston play out before so fundamentally changing the Rocket Pool protocol in pursuit of TVL, and some of your proposals may serve to expand what’s possible now that Houston is here.

I think too many believe we have no hope to scale without acquiescing to those that have not supported the protocol to this point, whereas I think we should enable and incentivize the operators that we have, and that will only serve to attract more Node Operators. I think there’s also a real centralization vector with ETH-only minipools, as I wrote about in my doc. Some think there’s a power law when it comes to ETH staking and centralization is inevitable. That may or may not be true, but we certainly don’t need to lean into it, we should endeavor to lean away when practical.

I think it’s also critically important to keep things simple. The community should have their eyes open on how long it took to get Houston fully complete, and how long “Rocket Pool 2.0” would take to implement in Saturn. I think the scope in front of us is not realistic in any timeframe that the community is thinking. At the same time, we should be developing alongside the core protocol changes. I think Rocket Pool’s next release should be as closely following Pectra as possible, just as Atlas was with Shanghai.

It’s the start of 2024 H2 already. I don’t believe a single RPIP in the proposed canonical set is realistic for 2024, which means we’re likely bumping up with Pectra. As part of our commitment to Ethereum we should be evolving with Ethereum, and leveraging the features being implemented in Pectra.

I’d like to remind the community these items from the pDAO charter:
6. The pDAO SHOULD prioritize the health of the Ethereum network
7. The pDAO SHOULD prioritize decentralization

Building Megapools atop MAXEB (EIP-7521) should be seen as a must IMO. Allowing Rocket Pool to help collapse the validator set as we scale will be an important narrative, and it will be a negative if other LSPs are collapsing their validator set, while we expand it.

Collapsing the validator set will tend to reduce network computational load. This opens the door to work on Single Slot Finality, and is also (for a different reason), in the critical path for Enshrined Proposer Building Separation. It could also improve validator performance.

My points on MAXEB may seem out of place in a discussion on tokenomics, but this should be considered for Saturn. We only get so many Rocket Pool releases per year (i.e., one). We should not lose the Ethereum forest for the tokenomics trees.

Lastly, I’d like to thank you again for the input from one of our largest node operators. To get any input from a large node operator and investor perspective is important, and it’s one I think should not be minimized, just because it arrived a bit late and a lot of work has been laid on a particular set of tracks.

Thanks,
Chris Rock

2 Likes

I added some more thoughts in the discord thread given that’s the more active discussion.