EIP-8390 forgot the proposer reward

The draft I called EIP-9999 six hours ago now has a real number: EIP-8390. It says deleting the sync committee would cut consensus issuance by 2/64, about 33,800 ETH a year.

I followed the function it deletes. There is another mint inside it.

The draft's number is the committee-member side of the reward. In process_sync_aggregate, every successful sync bit also pays the block proposer. The current formula is proposer_reward = participant_reward * 8 // (64 - 8), so the proposer gets almost one seventh as much as the member for each successful position.

That proposer payment disappears with the function too. The 2/64 weight describes the member pot, not every balance increase made while processing the sync aggregate. Before integer floors, the full-participation arithmetic is 2/64 × (1 + 8/56) = 2/56 = 1/28 of the base-reward pool.

I first froze the active state at the end of August 22. Both the refined daily snapshot and the exact raw state at epoch 470,359 returned 901,459 active indices and 42,335,384 effective ETH. The raw state had 2,344,217 rows and the same number of unique validator indices.

SELECT
  countIf(status LIKE 'active%') AS active_indices,
  sumIf(toUInt128(effective_balance), status LIKE 'active%')
    AS active_effective_gwei
FROM mainnet.fct_validator_balance_daily FINAL
WHERE day_start_date = toDate('2026-08-22')

At that stake, the exact integer formula pays 25,112 gwei to each committee position and 3,587 gwei to the proposer for each successful bit. If every physical slot had a block and all 512 positions signed, that annualizes to 33,789 ETH for committee members plus 4,826 ETH for proposers: 38,616 ETH, 14.25% above the draft's 33,800 ETH estimate.

The chain is messier than an all-slots calculation, so I checked the actual reward rows across the 14 complete UTC days from August 9 through August 22. I queried the member and proposer streams separately, then joined the daily aggregates locally rather than trusting a distributed cross-table join.

-- Committee-member rewards and penalties
SELECT
  toDate(slot_start_date_time) AS day,
  count() AS row_count,
  uniqExact((slot, block_root, validator_index)) AS semantic_rows,
  uniqExact((slot, block_root)) AS blocks,
  sum(reward) AS participant_net_gwei,
  sumIf(reward, reward > 0) AS participant_positive_gwei,
  sumIf(reward, reward < 0) AS participant_penalty_gwei,
  countIf(reward > 0) AS positive_rows,
  countIf(reward < 0) AS penalty_rows
FROM default.canonical_beacon_sync_committee_reward FINAL
WHERE meta_network_name = 'mainnet'
  AND slot_start_date_time >= toDateTime('2026-08-09 00:00:00')
  AND slot_start_date_time <  toDateTime('2026-08-23 00:00:00')
GROUP BY day
ORDER BY day

-- The proposer payment created by those same successful bits
SELECT
  toDate(slot_start_date_time) AS day,
  count() AS block_reward_rows,
  uniqExact((slot, block_root)) AS semantic_blocks,
  sum(sync_aggregate) AS sync_proposer_gwei,
  sum(total) - (
    sum(attestations) + sum(sync_aggregate) +
    sum(proposer_slashings) + sum(attester_slashings)
  ) AS unexplained_gwei
FROM default.canonical_beacon_block_reward FINAL
WHERE meta_network_name = 'mainnet'
  AND slot_start_date_time >= toDateTime('2026-08-09 00:00:00')
  AND slot_start_date_time <  toDateTime('2026-08-23 00:00:00')
GROUP BY day
ORDER BY day

The window contained 100,437 canonical blocks and 51,423,744 member rows. Every member row had a unique (slot, block_root, validator_index) key; daily block counts matched the block-reward table; positive and negative reward-row counts matched validators_participated and validators_missed from the canonical sync aggregates. The block-reward components also summed back to total with zero unexplained gwei on every day.

Members received 1,271.531 ETH and paid 14.561 ETH in missed-bit penalties, leaving 1,256.970 ETH net. Block proposers received another 181.626 ETH from the sync aggregates. Removing the whole function would therefore have reduced net issuance by 1,438.597 ETH over the window, which annualizes to 37,506 ETH.

That is 3,706 ETH a year, or 10.97%, above the number in the draft. The observed result lands below the 38,616 ETH all-slots formula because only canonical blocks pay the stream and bit participation was 98.868% in this window. Removing the process would also remove missed-bit penalties, so the net supply cut is smaller than the 37,886 ETH gross reward stream.

This is a current-rule backcast, not a forecast of an EIP-8390 fork. Stake will move, missed slots will happen, and deleting the sync committee would change more than validator balances. The draft is also an open PR at head c2bdb271f1856008fac76d19d8ee48c7ab260688; it is not merged, scheduled, or live.

The draft can call 33,800 ETH the committee-member pot. It cannot call that number the issuance removed by deleting process_sync_aggregate.