Name of Grant
Saturn 2 Exit and Bond Impact Model
Applicant: DAOplomats
Technical lead: @winverseDP
Request: $3,600 USD, paid in USDC in two milestones after acceptance
Delivery: Six weeks from approval
What is the work being proposed?
DAOplomats proposes to extend our published minipool exit simulation into a reproducible model of Saturn 2’s exit, bond and withdrawal trade-offs. The aim is to give the pDAO numbers it can use when choosing or revisiting parameters: how much liquidity a rule provides, what it costs remaining rETH holders, and how exits are distributed across node addresses.
Our preliminary work already shows a measurable trade-off. At a 50k rETH net shortfall, commission-first rules wipe 206–251 eligible minipool books, versus about 118 under uniform random selection. They also leave a lower capital-weighted commission on the remaining eligible minipools. We have not yet measured whether alternatives can keep that efficiency while reducing concentrated losses.
The grant funds four extensions:
- Exit-rule comparison. Extend the data to validator-level megapool state and compare the six candidate megapool criteria discussed in RPIP-71: random, lower RPL stake, FIFO, bond shortfall, larger-node bias, and avoiding repeated hits to the same node. This includes the tournament mechanism, with explicit scoring definitions and sensitivity to tournament size. For minipools, add a node-balanced tie-break within each exact commission tier and a broader node-balanced ordering. Compare all of these with our existing commission and random rules at 10k, 50k and 200k rETH net shortfalls.
- Bond and reentry analysis. Compare 1.5, 4 and 6 ETH marginal bond scenarios using the full bond curves, including initial requirements and reward-funded top-ups under RPIP-42 and RPIP-83. Report capital requirements, queue clearance under stated inflow paths, reentry effects, and the incremental return from topping up compared with solo staking.
- Exit-penalty economics. Model the RPIP-80 penalty and re-request schedule together. Compare the cost of non-compliance with additional operating rewards and exit/reentry opportunity costs across bond and commission levels. Report break-even conditions under stated assumptions; we will not infer how operators will actually behave.
- Mint-and-withdraw incentives. Test the concern that repeated large withdrawals could create costly validator churn. Account for the mint fee, withdrawal buffer, reward treatment of queued rETH, capital lockup, liquidity inflows and exit/entry delays. Compare the baseline with illustrative exit fees and hysteresis settings, reporting actor costs and protocol reward losses separately.
Exit comparisons will report validators selected, credited ETH, nodes touched, eligible books wiped, each node’s fraction of pools removed, top-10 exit share, and associated RPL exposure. Where execution and consensus data establish it, we will also count addresses losing all active Rocket Pool validators across minipools and megapools. Incomplete cases will stay marked unknown, and addresses will not be presented as verified independent operators.
Efficiency comparisons will separate rETH liquidity from returned operator bonds and total validator principal removed, and will distinguish remaining commission from modeled reward losses during queue and reinvestment delays. We will publish the measured trade-off between concentration and efficiency, including results that do not support a proposed alternative.
Is there any related work this builds off of?
Yes. We published preliminary findings in the RPIP-86 discussion, post 35. The snapshot is block 26,127,238 (5 October 2026), with 11,571 eligible minipools across 1,323 node addresses. Already-exiting and withdrawn validators are excluded. The post documents the method, seeds, a second-seed rerun, cap-crossing sensitivity, and 10k and 200k runs.
| Preliminary finding | Commission ordering / observed | Random comparison |
|---|---|---|
| Commission-homogeneous multi-minipool nodes | 72.28% | 18.97% shuffled mean |
| Eligible minipool books wiped, 50k rETH | 206.0–250.8 across four tie-breaks | 117.6 mean |
| Top-10 node share of exits, 50k rETH | 39.5–43.0% | 22.8% mean |
| Remaining capital-weighted commission | Approximately 10.64% | Approximately 11.68% |
The commission difference corresponds to about 3.1 basis points of annual yield at an illustrative 3% gross reward rate. This is a static fee proxy, not a protocol APR forecast. Wiped books are not verified operator departures, and associated RPL is not a forecast of unstaking or selling. At 200k rETH, some commission rules have a lower top-10 share than random, so the result depends on the metric and request size.
This grant pays only for the extensions above. The published baseline is completed work and is not a paid deliverable. The project builds on the Saturn 2 proposals and community discussion; no core-team or GMC endorsement is claimed.
Will the results of this project be entirely open source (MIT), GPL, Apache, CC BY? If not, which parts will not be, why, and under what license will they be published?
Yes, entirely. Python code and tests will be published under MIT, reports and original charts under CC BY 4.0, and original data exports under CC0. Third-party sources keep their original licensing and attribution. License files will be included in the first milestone release. Inspecting the funded analysis will not require any proprietary model or paid subscription.
Benefit
| Group | Benefits |
|---|---|
| Potential rETH holders | Clearer evidence about withdrawal liquidity, queue delays and the costs of alternative designs when evaluating rETH. Adoption growth is not assumed. |
| rETH holders | Quantifies the reward and capital-efficiency consequences of liquidity exits, including idle-capital costs and withdrawal incentives. |
| Potential NOs | Shows how bond choices and reentry priority affect required capital and modeled queue access. |
| NOs | Measures the distribution and severity of exits, surviving megapool validators, top-up economics and penalty incentives. |
| Community | Reproducible comparisons for Saturn 2 deliberations and later parameter reviews, with assumptions and uncertainty exposed. |
| RPL holders | Evidence about operator participation and associated RPL exposure. It does not claim to predict RPL sales or price. |
This work serves the GMC goal of improving quality of life for the protocol and its community. It informs parameters that affect the experience of node operators and rETH holders, which bears on whether operating a node or holding rETH is attractive.
Which other non-RPL protocols, DAOs, projects, or individuals, would stand to benefit from this grant?
Ethereum staking researchers and other staking protocols can reuse the public methods. DAOplomats gains a reusable research contribution and a public record of delivery. The funded questions and outputs are specific to Rocket Pool. This is not a cross-DAO advocacy campaign.
Work
Who is doing the work?
DAOplomats. @winverseDP is the technical lead and is accountable for data collection, model implementation, validation and publication. Other members of the DAOplomats team contribute to the work. Public contact: Telegram @Win_Verse.
We will invite protocol authors and node operators to challenge assumptions and inspect results on the public forum. Delivery and acceptance do not depend on any particular reviewer volunteering, and no independent reviewer appointment is represented as confirmed.
What is the background of the person(s) doing the work? What experience do they have with such projects in the past?
Winverse’s background includes DAO governance, protocol research, grant-program design and technical communication, including prior work on the RARI Foundation grants framework and playbook, and the Cryptex’s Treasury Strategy Optimization. DAOplomats works on governance across multiple ecosystems.
The most relevant evidence is the Rocket Pool work already published in post 35: a pinned block, randomized comparisons against a stated baseline, a second-seed rerun, cap-crossing sensitivity and multiple request sizes. The GMC can assess this directly.
What is the breakdown of the proposed work, in terms of milestones and/or deadlines?
| Milestone | Due from approval | Deliverables | Payment after acceptance |
|---|---|---|---|
| M1: Exit rules and bonds | End of week 3 | Extended data and coverage report; minipool alternatives and six defined megapool criteria; 10k/50k/200k comparisons; bond, top-up and queue/reentry scenarios; tests, CSVs, charts and a forum summary. | $1,800 |
| M2: Penalties, withdrawal incentives and release | End of week 6 | Penalty break-even analysis; mint-and-withdraw, fee and hysteresis sensitivities; a final data refresh and comparison; integrated decision report; documented refresh commands, release and handoff. | $1,800 |
We will freeze and record RPIP versions at kickoff, log any model-relevant changes during delivery, and reflect them in the final comparison. The work stays useful for parameter review after the initial votes, and delivery does not depend on any vote still being open.
How is the work being tested? Is testing included in the schedule?
Yes, throughout both milestones:
- Reconcile registries and execution/consensus joins; report missing state, duplicates and exclusions.
- Check exact capital accounting, hard-cap boundaries, and the distinction between gross requests and net shortfalls.
- Use synthetic node layouts to test selection scores, tournament sampling, reentry priority, and full-node versus eligible-book counts.
- Run at least 1,000 draws for each stochastic comparison at each request size, repeat with a second seed, and report empirical 5th–95th percentiles with sensitivity to cap handling and eligibility assumptions.
- Reconstruct a representative extended exit scenario independently in a spreadsheet, and check bond, reward, fee and penalty equations against small hand-calculated cases, including zero-inflow and zero-fee boundaries.
The release will include an offline replay command and automated checks. Queue and incentive results will be conditional scenarios, with parameter ranges and data gaps documented.
How will the work be maintained after delivery?
All deliverables and saved datasets will stay public. The final handoff will document how to collect a new pinned snapshot and regenerate the reports on GitHub. We will correct reproducible defects affecting the delivered scope for 90 days after final acceptance, within this grant. Later feature development, or support for materially changed protocol mechanics, would need a separate scope.
Costs
What is the acceptance criteria?
M1 is accepted when the public release contains the extended data and coverage report, both new minipool alternatives, six explicitly defined megapool criteria using tournament selection, all three withdrawal sizes, and the bond, top-up and reentry comparisons. Each rule must state its scoring, eligibility and capital-accounting assumptions. Tests, per-draw results and reproduction instructions must accompany the report.
M2 is accepted when the penalty and withdrawal-incentive modules publish their equations, parameter grids and break-even results; the refreshed data and integrated report show the efficiency and concentration comparisons; and the release passes its documented checks and offline replay.
For both milestones, every headline number must trace to a saved input and generated output. RPIP versions, blocks, seeds and limitations must be recorded. Unknown validator state, proxy metrics and illustrative alternatives must be labeled. Code and reports must carry the stated licenses. A forum summary and public issue log must accompany each delivery. Acceptance concerns completeness and reproducibility, not a preferred policy conclusion.
What is the proposed payment schedule for the grant? How much USD $ and over what period of time is the applicant requesting?
$3,600 USD, paid in USDC, over six weeks. The two $1,800 payments are requested only after the GMC accepts the corresponding milestone. No upfront payment is requested.
The planning estimate is 100 hours, an effective rate of $36/hour. This is a fixed-price request; the hours explain the estimate and do not create hourly billing.
| Work | Estimated hours | Budget |
|---|---|---|
| Extended data and exit-rule comparisons | 30 | $1,080 |
| Bond, top-up and queue/reentry analysis | 20 | $720 |
| Penalty economics | 15 | $540 |
| Withdrawal incentives, fees and hysteresis | 20 | $720 |
| Release validation, reports and scoped support | 15 | $540 |
| Total | 100 | $3,600 |
Routine data-access and compute costs are included. No recurring infrastructure funding is requested.
Who will directly receive the payment? (Required — the GMC can only send grants directly to a recipient address and cannot accept invoices.)
DAOplomats, in USDC on Ethereum mainnet to:
0xEe5A54F83Ef0F983e9173BcCAc7615DBE12a78A0
No invoice is required.
How will the GMC verify that the work delivered matches the proposed cadence?
Each milestone will have a tagged public release, a completion report mapping artifacts to the acceptance criteria, and a forum update. The GMC can replay the saved inputs, inspect the tests and the spreadsheet check, and reconcile report figures with CSV outputs before payment. Weekly forum updates will show progress and flag any material scope or data issues.
What alternatives or options have been considered in order to save costs for the proposed project?
We will reuse our existing collection pipeline and simulation rather than charge to rebuild them. Batch scripts, cached snapshots and static reports avoid ongoing hosting. The work concentrates on numerical policy comparisons and excludes production exit automation, Smartnode features, a security audit and a hosted dashboard.
A prose-only review would cost less but would not measure the node-level distribution or identify break-even thresholds. The existing commission-versus-random study is already available at no extra cost. This request pays for the missing comparisons and incentive analysis.
Have you already been compensated by the RP protocol in any way for this work?
No. Neither winverse nor DAOplomats has received compensation of any kind from Rocket Pool or the GMC, and no other grant, client or DAO is funding this work. The milestones exclude the already-published baseline, and the same work will not be charged to another grant or retrospective award.
Conflict of Interest
Does the person or persons proposing the grant have any conflicts of interest to disclose? (Please disclose here if you are a member of the GMC or if any member of the GMC would benefit directly financially from the grant).
None to disclose. No member of the GMC is involved in this work or would benefit financially from it. Neither winverse nor DAOplomats runs a Rocket Pool node or holds RPL or rETH, and neither has a relationship with any GMC member or the Rocket Pool team. DAOplomats takes part in governance in other DAOs; none of those paid roles touches Rocket Pool or rETH, and this grant carries no mandate to promote another protocol.
Will the recipient of the grant, or any protocol or project in which the recipient has a vested interest (other than Rocket Pool), benefit financially if the grant is successful?
DAOplomats will receive the grant payments and may benefit reputationally from delivering public research. No other protocol or project in which we have a vested interest will benefit financially. The outputs are free to use, no token or fee is attached, and results will be reported whichever policy comparisons they support.