A fixed supply does not answer a separate question: who pays for proof-of-work after the subsidy fades?
Bitcoin's block subsidy declines every 210,000 blocks. The intended long-run replacement is transaction fees. A tail emission would give miners a recurring subsidy, but it would also change the monetary contract by creating new units indefinitely.
A recent Delving Bitcoin discussion about the diminishing subsidy surfaced another possibility: charge old UTXOs, recreate the remaining value under the same spending condition, and direct the charge to miners. Pieter Wuille argued that monetary-policy experiments of this kind belong in a separate currency where they can compete with Bitcoin—not in a contested change to Bitcoin's existing contract.
OBTC is such a separate experiment. It is not a proposal to modify Bitcoin, and it makes no claim that Bitcoin has chosen the wrong policy. It asks whether a recurring proof-of-work contribution can come from dormant value without increasing total supply.
REAP does not mint the tax. It consumes an expired output, recreates the refund under the original locking script, and allows the miner to claim only the difference.
362,880 blocks · 70% refund · 30% tax
One concrete dormant-value mechanism
Under the current candidate parameters, every OBTC UTXO has a lifecycle of 362,880 blocks—approximately 6.9 nominal block-years. Before expiry, its owner can spend it normally or renew it with an ordinary signed transaction. Renewal creates a new output and resets the lifecycle.
After expiry, the ordinary owner-signature spend path closes. A miner may include a consensus-constrained REAP system transaction. For a normal expired input, integer accounting assigns 30% to tax and recreates the remaining value under the input's original locking script. Very small expired inputs below the current 720-sat threshold are folded fully into tax instead of creating an uneconomic refund.
Miners cannot freely select the richest expired outputs. If a block includes k REAP inputs, they must match the first k entries of a globally ordered expired set. The current ordering is expiry height, then amount, then outpoint. Input caps and a separate REAP weight ceiling bound the work per block.
This is not ordinary storage rent
The charge is proportional to value, not to the bytes an output occupies. OBTC is therefore testing a dormant-capital contribution to security—not claiming that a high-value UTXO costs more to store.
Bitcoin block 953,697 · No-renewal stress case
The prototype was run against a full live UTXO snapshot
The offline rehearsal started from Bitcoin's live spendable UTXO set at block 953,697: 165,782,759 outputs holding 20,042,261.87498542 BTC. The snapshot was read from a fully validating node's UTXO database and checked for missing amounts, duplicate outpoints, and ordering violations.
The preview then applied the implemented REAP selector using a 30% tax, a 720-sat dust threshold, a maximum of 256 normal inputs, a separate maximum of 1,024 refundless dust inputs, and a 400,000-weight ceiling.
It deliberately assumed no wallet renewals and no voluntary consolidation. That does not model likely user behavior. It is a mechanical stress case: what happens if every live output in the snapshot is eventually allowed to cross the lifecycle boundary?
- Snapshot height
- 953,697
- Live spendable UTXOs
- 165,782,759
- Total snapshot value
- 20.04M units
- Missing amounts
- 0
- Normal input cap
- 256
- Dust input cap
- 1,024
- REAP weight ceiling
- 400,000
- Voluntary renewals
- 0
The mechanism ran · The economics were uneven
Four results that matter more than the headline tax total
Every source input was eventually selected. No simulated REAP transaction exceeded the 400,000-weight ceiling; the maximum estimate was 253,948.
The maximum pending queue was not a launch-day cleanup. Its first zero-backlog point arrived 173,871 blocks after the preview began—about 3.31 nominal years.
Median block tax was about 2.37 units, while the maximum was about 39,007 units. The mechanism produces a distribution-driven revenue stream, not a smooth subsidy.
Dust represented 44.40% of selected inputs but only 0.005855% of total simulated tax. It is mainly processing load, not security-budget value.
The full preview emitted 391,984 REAP blocks over a 392,077-height span, approximately 7.46 nominal block-years. That span includes outputs from across the snapshot aging into eligibility; it is not a prediction of a final launch schedule.
| Simulated block-tax statistic | Satoshis | OBTC units |
|---|---|---|
| Median | 237,018,183 | 2.37018183 |
| p90 | 1,564,887,706 | 15.64887706 |
| p99 | 18,566,430,000 | 185.66430000 |
| Maximum | 3,900,727,469,110 | 39,007.27469110 |
Implementation feasibility and economic desirability are different results.
The rehearsal supports a narrow engineering conclusion: under the observed snapshot and tested caps, the selector made progress without exceeding its modeled weight budget. It does not show that rational holders will remain dormant, that miners will always include low-value prefixes, or that the resulting revenue is sufficient to secure a network.
Related systems exist
Ergo makes the comparison unavoidable—and useful
UTXO lifecycle charges are not new. Ergo's storage-rent protocol allows miners to charge old boxes after four years. A miner can spend and recreate a box while preserving its state, minus a rent amount tied to storage. Ergo's miner documentation explicitly treats the mechanism as both garbage collection and a future miner-revenue source.
Freicoin explored demurrage more broadly, reducing monetary value over time rather than attaching a lifecycle specifically to inactive UTXOs. Tail emission takes another route: distribute the security cost across the entire monetary base through continuing issuance.
| Mechanism | Who bears the cost? | Primary rule | Supply effect |
|---|---|---|---|
| Transaction fees | Transactors | Block-space auction | No added issuance |
| Tail emission | All holders through dilution | Recurring subsidy | Continuing issuance |
| Ergo storage rent | Old boxes, mainly by state size | Per-byte rent after inactivity | Transfer to miners |
| OBTC REAP | Unrenewed dormant value | Percentage split with same-script refund | Transfer to miners |
OBTC's experiment is therefore not “the first expiring UTXO.” Its narrower distinction is an amount-proportional dormant-value contribution, implemented on a Bitcoin-derived chain with same-script refunds, global prefix ordering, bounded processing, wallet renewal, and a full historical rehearsal.
Five limits on interpretation
What the prototype does not prove
- Dormant does not mean lost. The protocol observes output age, not a person's intent, custody status, estate plan, or access to keys.
- REAP does not generally shrink the UTXO set one-for-one. A normal expired input is replaced by a refund output. Direct count reduction comes mainly from refundless dust processing or voluntary consolidation.
- The preview is not a revenue forecast. It assumes no renewal, no consolidation, continuous candidate processing, and value in the native OBTC unit rather than an external unit of account.
- The mechanism does not know the required security budget. A protocol can calculate tax, but it cannot observe its own market value or the cost of attacking it.
- A working selector does not create adoption. Network security still depends on users, miners, independent nodes, wallet reliability, and a monetary contract people voluntarily choose.
OBTC is a source-available technical experiment under review. It is not Bitcoin, not a production financial network, and not an instruction to move BTC or expose Bitcoin keys.
Questions for technical review
The open questions are economic and adversarial, not promotional
- Is an amount-proportional dormant-value charge a defensible way to fund proof-of-work, or is per-byte rent or broad dilution more coherent?
- Does recreating value under the original script preserve intended ownership semantics across P2SH, P2WSH, Taproot, timelocks, and covenant-like constructions?
- Does a mandatory global prefix prevent miner cherry-picking at the cost of a public-goods problem during low-tax stretches?
- Under realistic renewal behavior, does meaningful security-budget value remain—or do rational holders remove nearly all of it before expiry?
Criticism that changes the mechanism, narrows the claim, or shows that a simpler design dominates it is a useful result.
Code, data, and reproducibility
Inspect the implementation, not the story
The aggregate rehearsal data are public. Raw txid/vout-level snapshot rows are intentionally excluded; the published package includes per-block aggregates, checksums, summary JSON, derived tables, figures, and a verification script.
Reproduce the claims
Review the consensus implementation and download the aggregate 953,697-height rehearsal package.
Primary references
- OBTC candidate lifecycle and REAP parameters
- Consensus validation for REAP markers, accounting, and global prefix selection
- Machine-readable rehearsal summary
- Delving Bitcoin discussion: Addressing the Diminishing Block Subsidy
- Ergo protocol documentation: Storage Rent
- Freimarkets paper describing Freicoin and demurrage-capable assets