ENS-GR-01 · A Ten-Proposal Governance Record

Draft proposal: This article is an early mechanism proposal for discussion. It is neither a final ENSIP nor a deployment-ready technical specification. Its purpose is to put sustained personal voting, considered delegation, and Delegate performance into one testable settlement framework. The parameters, transition arrangements, event definitions, and Sybil-resistance boundary should be openly discussed, audited, and refined by ENS DAO Delegates, contributors, and the technical community.

The current ENS pilot rewards two things: a Holder delegating to an Active Voter at the time of distribution, and an Active Voter having voted on at least seven of the last ten on-chain proposals.

Holder rewards are driven primarily by a 180-day time-weighted balance. The pool is settled separately, with 90% allocated to Holders and 10% to Active Voters.

That design can move idle ENS toward active addresses, but it cannot distinguish considered, ongoing delegation from a delegation made only at the end of the period. It also does not make a 10/10 personal voting record materially more valuable than merely clearing the 7/10 threshold. (Pilot rules, algorithm documentation)

I propose a new settlement framework: the Ten-Proposal Governance Record.

It rewards only three verifiable forms of governance participation:

A Holder repeatedly voting personally with their own ENS;
A Holder delegating to a Delegate who genuinely represents others, and revisiting that choice;
A Delegate repeatedly voting with externally delegated Voting Power (VP) and leaving an auditable rationale.

Reward coefficients

“Highest return” means the return rate for the same effective VP. It does not mean that a larger Holder must receive fewer ENS in absolute terms. Each eligible VP position first receives a reward coefficient $c$ from one of the two tables below, and then enters the common settlement formula.

Personal voting

Personal votes cast in the last ten closed proposals	Reward coefficient $c$	Meaning
10/10	1.000	All ten independent voting opportunities were used; this is the highest-return path
9/10	0.810	One absence should cost more than 10% of the return
8/10	0.640	Still a meaningful record of participation, but materially below full attendance
7/10	0.490	Preserves a path back to eligibility without equating it with sustained participation
0–6/10	0	Does not create a settlement-eligible record

This table keeps full personal voting in first place. The squared decline, rather than a linear decline, makes every missed vote carry a noticeable cost.

Delegated voting

One uninterrupted delegation relationship produces two categories of eligible records: one for the Holder’s decision to make and maintain the delegation, and one for the Delegate’s work in representing externally delegated VP.

When the Delegate has full attendance and the Holder has reconfirmed the delegation, the two coefficients total 1.000: 0.700 for the Holder and 0.300 for the Delegate. The eventual ENS amounts are still determined by the common formula and everyone’s relative scores; this is not a mechanical one-to-one split of a single payment.

Delegate’s qualified votes in the last ten proposals	Holder actively reconfirms within five proposals of delegation	Holder coefficient $c$	Delegate external-VP coefficient $c$	Result
10/10	Yes	0.700	0.300	Both Holder and Delegate receive their full respective coefficient
10/10	No	0.350	0.300	Only the Holder’s return is halved
9/10	Yes	0.567	0.243	One missed vote reduces both coefficients
9/10	No	0.284	0.243	The Holder’s coefficient is halved after the attendance reduction
8/10	Yes	0.448	0.192	Coefficients continue to decline
8/10	No	0.224	0.192	The Holder’s coefficient is halved after the attendance reduction
7/10	Yes	0.343	0.147	Still eligible, but with a substantially lower return
7/10	No	0.172	0.147	The Holder retains only half of the delegation component
0–6/10	Irrelevant	0	0	The Delegate is no longer eligible under this program

“Active reconfirmation” should be a gasless signature initiated by the Holder within the first five closed proposals after the delegation takes effect. Its purpose is to require the Holder to revisit an early delegation choice rather than remain a passive delegator indefinitely.

Failing to reconfirm halves only the Holder’s coefficient; it does not reduce the Delegate’s reward for work already performed. Each missed qualified vote by the Delegate reduces both sides’ coefficients according to the table. Below 7/10, the Delegate loses program eligibility and the Holder’s delegation component also falls to zero.

The baseline table assumes that the Delegate does not change during the ten-proposal window. How to split an existing record and whether a confirmation window restarts after mid-window redelegation must be defined in an implementation design; this article does not present that unresolved question as a settled conclusion.

Newcomers accrue from the first vote, but new wallets cannot jump the queue

A new participant does not need a pre-existing history across ten proposals.

The first personal vote, or the first delegation to a Delegate Profile, creates a pending ten-proposal record for that address. Each closed proposal fills one slot. Voting, delegation continuity, and Delegate performance begin to accrue from that point.

The first cash settlement unlocks only after the record covers ten closed on-chain proposals. Moving to a new address starts a new record from its first slot.

This achieves two things:

Newcomers begin accruing from their first vote;
A large Holder cannot split ENS across many new wallets just before distribution, vote once, and receive the same return as a mature account.

For existing participants, the most recent ten closed proposals can initialize the record at launch. Delegation relationships that already exist at launch should have a one-time, public migration-confirmation window. A Holder who signs within that window is treated as having completed the “reconfirm within five proposals” requirement for the first record. Delegations made after launch then follow the ordinary five-proposal confirmation window.

Settlement formula

Reward_i=R\times\frac{c_i\ln(1+V_i/K)}{\sum_j c_j\ln(1+V_j/K)}

$R$ : the reward pool for the round;
$c_i$ : the reward coefficient assigned to the $i$ th VP position by one of the two tables above;
$V_i$ : the time-weighted VP of that position while it was in the relevant governance state during the current ten-proposal record;
$K$ : one shared VP-diminishing scale, published before the period from the median of mature, eligible VP positions in the previous period. In the first period, when no historical sample exists, a governance proposal should set and publish a baseline value in advance.

$i$ is not a wallet type. It can represent VP used in a Holder’s personal voting, VP a Holder has delegated, or a Delegate’s externally delegated VP. The three types of records differ only in their coefficient $c_i$ and the VP counted; they use the same settlement function, the same $K$ , and the same pool.

Incentivising considered delegation

The system cannot prove that someone read every proposal, nor can it infer from one on-chain vote whether that vote was automated. It can, however, cut the return from default, unattended delegation in half and create a process that can be checked:

To earn the Delegate component, a Delegate Profile must publish a brief rationale and a conflict-of-interest disclosure before each vote closes;
When a Holder chooses a Delegate for the first time, the interface should show that Profile’s attendance rate, rationales, conflicts of interest, and governance priorities;
Within the first five closed proposals after delegation takes effect, the Holder should submit a gasless signature to actively reconfirm or change the choice. Without it, the Holder coefficient is halved as shown in the delegated-voting table;
Each time a Delegate misses a qualified vote, both the Holder’s delegation component and that Delegate’s external-VP component decline according to the delegated-voting table. Below 7/10, neither component scores.

This prevents a Holder who delegates and then disengages completely from earning the full return, while allowing the community to inspect whether a Delegate deserves continued support. For people representing others, public reasoning is part of accountability. For people voting only their own ENS, the on-chain vote remains the only mandatory verifiable action.

VP diminishing returns

$\ln(1+V_i/K)$ in the common formula is the only VP-diminishing mechanism. $K$ is the same for every record, and is published before each period from the overall median of mature, eligible VP positions in the preceding period.

The curve is approximately linear for smaller participants. As VP grows, each additional unit contributes less score, but never abruptly becomes worthless.

VP relative to $K$	VP-curve score in the common formula	Score relative to $V=K$
0.5×	0.405	0.585×
1×	0.693	1.000×
2×	1.099	1.585×
4×	1.609	2.322×
10×	2.398	3.459×

A position with ten times as much VP therefore still earns more score than one with VP equal to $K$ , but not ten times as much. This is more coherent than a hard 250 ENS cap per address: it does not make all marginal rewards above 250 abruptly disappear, and it does not create a sharp threshold where opening another wallet earns another cap.

Remaining limitations

The logarithmic curve still creates an incentive to split when an identity can remain unknown over a long period. If one person can present as two entirely unrelated governance identities and mature both ten-proposal records, the combined score of the two accounts will exceed the score of one account. Without KYC or a credible identity-clustering process, an on-chain rule cannot establish hidden common control.

The rule can therefore guarantee only the following:

During a new address’s first ten governance opportunities, splitting does not pay better than stable participation;
Known or disclosed related addresses can be scored together, removing the benefit of splitting them;
A Delegate cannot unilaterally split externally delegated VP. Re-delegation must be initiated by the Holder and carries a continuity cost.

It cannot guarantee that hidden identity splitting will never be profitable over a ten-year horizon.

Two reproducible Round 2 tests

The official Round 2 output confirms that the full 5,000 ENS was distributed. There were 541 recipient addresses: 508 direct-payment records and 33 lottery-winning addresses. Seven Holder components each received the 250 ENS cap, for 1,750 ENS in aggregate, or 35% of the total pool. The Top 20 recipient addresses together received 58.667%. (Round 2 public CSV, Round 2 metadata)

The two tests below examine only what the public data can support.

Test 1: How a hard cap, linear allocation, and a continuous curve treat the same capped Holders

The Round 2 endpoint exposes tokenHolderBalance for the seven Holders capped at 250 ENS. To avoid selecting a parameter arbitrarily, this Holder-only stress test temporarily sets $K=43{,}952$ ENS. The figure is inferred, under a linear assumption, from the published balance and Holder component reward of the eighth, uncapped Holder.

This is not the future rule’s $K$ . It is only a local historical calibration for comparing the current hard cap with a continuous curve.

Looking only at these seven capped Holders:

Scoring method	Largest Holder’s share among the seven	Top two Holders’ share among the seven	Interpretation
Current 250 ENS hard cap	14.29%	28.57%	All seven receive the same amount, but the cap itself encourages address splitting
No cap; linear allocation by balance	26.17%	49.68%	Does not encourage splitting, but produces the highest concentration
Continuous VP curve in the common formula	20.81%	40.57%	Sits between the two: larger balances still earn more, but their marginal return falls materially

Unlike a hard cap, the continuous curve does not compress seven different balances into the same number. It reduces the top two Holders’ 49.68% share under linear allocation to 40.57%.

Test 2: Moving the largest public Round 2 Holder balance into a new address

The largest published balance among the top Round 2 Holders is 200,000 ENS. This test keeps the historical calibration $K=43{,}952$ and assumes both sides would vote personally in all ten proposals. The table tests only the VP-curve component of the common formula.

Action	Settlement score	Compared with stable holding
One mature governance account continuously holds 200,000 ENS	75,328	Baseline
Transfer 100,000 ENS to a new address; the new address has not unlocked after ten opportunities	52,144 + 0	-30.78%
After ten opportunities, two addresses still remain hidden as separate identities	104,288	+38.44%

The concrete behavioural incentive the design can guarantee is this: splitting half of one’s own ENS into a new address immediately sacrifices 30.78% of settlement score rather than creating additional rewards.

Conclusion

Whale concentration is not a single “large amount” problem. It involves whether personal voting is rewarded, whether delegation reflects an actual choice, whether Delegates perform their representative role, and whether diminishing returns turn into a wallet-splitting game.

This is not a complete anti-Sybil mechanism. It is, however, closer to what a DAO should be purchasing: sustained, verifiable governance commitment chosen by the Holder.

Discussion and refinement: This article offers only a starting point that can be challenged or revised by data, implementation details, and community experience. The coefficient values, the first-period $K$ , migration confirmation, and treatment of mid-window redelegation should not be fixed unilaterally by the author. I hope experienced ENS Delegates, researchers, contributors, and technical participants will discuss the design openly and improve it incrementally using pilot data.