Rewarding Active Delegates — April 2026 Results: Correction & Post-Mortem

DAO Programs & Initiatives Rewarding Active Delegates Program (RAD)
1

Executive summary

The April 2026 RAD report was built from an incomplete automated pull of Snapshot votes for the month. A bug in the monthly automation left out one proposal that closed in April 2026: Updating the Code of Conduct & DAO Procedures to Become Living Documents (off-chain decision making, non-constitutional).

That gap affected participation columns, rationale checks, incentive budgets, and total rewards. The first post listed 14 delegates and $14,500 on three Snapshot votes. The corrected workbook uses four votes, the same enrollment (34 delegates), and $20,000 total rewards (+$5,500 vs what was reported).

Automation is fixed for future months. This post is for transparency and to reconcile any delta with the 5 May payout. Delegates don’t need to take any further action.

April participants

34 delegates enrolled in April (unchanged from the original post).

Voting breakdown (corrected)

During April 2026, there were four Snapshot votes on arbitrumfoundation.eth with end within the month (UTC):

Off-chain decision making (non-constitutional): $7,000 budget, $500 cap/delegate

  1. Updating the Code of Conduct & DAO Procedures to Become Living Documents (missing from the first report; added in correction)

  2. Improvements to the Arbitrum Audit Program

Off-chain temperature check (non-binding): $5,000 budget, $300 cap/delegate

  1. Transfer 6,000 ETH and Idle Stablecoins from the Treasury to the Treasury Management Portfolio

  2. [Constitutional] AIP: ArbOS 60 Elara

Total RAD budget envelope for April (corrected): $24,000 ($7k + $7k + $5k + $5k).

The first report implied $17,000 ($7k + $5k + $5k) by omitting the Code of Conduct vote.

Incentive tiers remain as in Table 1 of the original post.

What went wrong

1. Missing Snapshot vote (root cause)

The pipeline uses an automated monthly job that should list every Snapshot proposal whose vote end falls in the reporting month. Because of a bug in the automation, the Code of Conduct vote never made it into the dataset. The other three April votes were processed. The omission was not policy; the proposal fell off the radar of that run.

2. Participation metrics: what changed and what did not

The original post reports 64.71% program participation across enrolled delegates. That aggregate figure is the same in the corrected build:

Metric Calculation (corrected)
Program participation 88 “Yes” votes ÷ 136 slots (34 delegates × 4 proposals) = 64.71%

What changes is which votes count, not that headline ratio:

First report Corrected
Proposals in scope 3 4
Per-delegate Participation formula COUNTIF(3 cols) / 3 COUNTIF(4 cols) / 4
Program-wide participation 64.71% 64.71%

Row-level participation rates in the corrected workbook do not include anyone at exactly 75%. With four votes, observed values are only 0%, 50% (2/4), or 100% (4/4):

Participation (row) Delegates
100% (4/4) 19
50% (2/4) 6
0% (0/4) 9
75% (3/4) 0

The ≥75% monthly rule still applies (inclusive at exactly 0.75 in the formula). In April’s data, delegates who clear the threshold on four votes are at 100%, not 75%. With only three votes in the first report, ≥75% effectively required 3/3 (100%), because 2/3 ≈ 67% fails.

The correction is therefore not “delegates were secretly at 75%.” It is: one vote was missing, so participation and rationales were wrong for anyone who voted on Code of Conduct, and rewards used the wrong proposal set.

3. Incentive caps (decision making)

Per program rules, off-chain decision-making uses a $ 500-per-delegate payout cap (and a $7,000 budget per proposal). The corrected workbook applies $500 on both decision-making sheets (Code of Conduct and Audit Program).

4. Downstream effects

Adding the fourth vote increases rational instances (88 vs 66) and changes reward math across four proposal sheets. Footer metrics per sheet (Eligible VP, Total spent, Dollar per vote, ARB per USD) should be read from each tab in the corrected workbook after recalculation.

Corrected metrics

Participation

  • Enrolled: 34

  • Program participation (aggregate): 64.71%

  • Voted on Code of Conduct: 22 / 34

  • Voted 4/4 Snapshot votes: 19

  • Below 75% row threshold: 15 (same count as first report; meaning of each row updated)

Rationale requirement

Among delegate-vote instances where the delegate voted:

  • 88 instances (22 voters × 4 proposals)

  • 60 with on-time rationale ≈ 68.18% (same rate class as the original post; larger denominator)

Voting power & incentivised VP (per proposal sheet)

Proposal Total VP deployed (approx.) Incentivized VP Incentivized %
Updating the Code of Conduct 108,693,295 ARB 60,638,764 ARB 55.79%
Improvements to the Arbitrum Audit Program 107,643,549 ARB 62,746,244 ARB 58.29%
Transfer 6,000 ETH and Idle Stablecoins… 107,637,009 ARB 62,746,244 ARB 58.29%
[Constitutional] AIP: ArbOS 60 Elara 108,693,295 ARB 60,638,764 ARB 55.79%

Incentivized VP = VP that counts after ≥75% monthly participation, 200k ARB minimum on the row, and rationale Yes on that proposal sheet.

Economics (footer of each sheet)

As in the original post, each proposal tab reports Dollar per vote and ARB per USD at the bottom. Those values are a redistribution construct when caps bind; they may not match naive total VP ÷ total USD.

Original post (three votes only):

  • Decision making (Audit only): $0.0001115604625 / vote · 8,963.75 ARB per USD

  • Temperature check (pooled): $0.00006081728275 / vote · 16,451.33 ARB per USD

Corrected: Four footers (one per sheet). Decision-making pools are split across two $7k budgets; temperature-check pools across two $5k budgets. See the corrected workbook for recalculated footer lines.

Corrected rewards

Source: RAD Program - April 2026 Results, sheet Comparison

Summary

Published (May 2026) Corrected
Delegates paid 14 14
Proposals counted per delegate 3 4
Total rewards $14,500.00 $20,000.00
Net adjustment n/a +$5,500.00

Delegate table

Delegate Proposals voted Corrected rewards Previously reported Net
Arana Digital 4 $1,600.00 $1,300.00 +$300.00
Areta 4 $1,600.00 $1,300.00 +$300.00
DZack23 4 $1,600.00 $1,300.00 +$300.00
Griff 4 $1,600.00 $1,300.00 +$300.00
Jojo 4 $1,600.00 $1,225.98 +$374.02
Layer3 4 $1,600.00 $927.58 +$672.42
MaxLomu 4 $1,600.00 $1,300.00 +$300.00
Mihal 4 $1,600.00 $1,300.00 +$300.00
Reverie 4 $1,600.00 $1,300.00 +$300.00
Uniswap Delegation 4 $1,600.00 $924.11 +$675.89
Zeptimus 4 $1,600.00 $671.84 +$928.16
cp0x 4 $800.00 $350.49 +$449.51
GMX 4 $800.00 $300.00 +$500.00
L2BEAT 4 $800.00 $1,000.00 −$200.00
TOTAL $20,000.00 $14,500.00 +$5,500.00

Reading the pattern

  • $1,600: ten delegates at the combined outcome across two decision-making votes ($500 cap each) and two temperature-check votes ($300 cap each), with full budget redistribution on eligible VP.

  • $800: cp0x, GMX, and L2BEAT (lower corrected total; caps, eligible VP, and rationale weighting differ from the three-vote build).

  • L2BEAT is the only row with a negative Net (−$200): the first report overpaid relative to the corrected model.

  • GMX and cp0x receive more in the corrected total than in the first report, not less.

What we are not changing

  • 34 enrolled participants for April

  • 200k ARB minimum VP per proposal row

  • ≥ 75% monthly participation rule (now applied over four votes)

  • Rationale, timing, and detection methodology

  • Budget tiers from Table 1

Conclusions

April 2026 showed that reliable month-boundary automation is as important as delegate activity. The first report omitted a live Snapshot vote, understated total rewards by $5,500, and misstated individual totals for several delegates.

The corrected workbook restores four votes, keeps 64.71% aggregate program participation, and sets $20,000 in April rewards across 14 delegates.

We regret the confusion for participants and thank delegates who voted and published rationales across all four April proposals, including the Code of Conduct, which was missing from the first report.

Attachments / links