2026-11-20-seed-proof-of-work
Published
26 claims
- Each block mined on top of a Bitcoin transaction adds work that anyone trying to reverse it must redo: the Bitcoin whitepaper says an attacker would have to redo the proof of work of that block and every later block, then overtake the honest chain.
- Bitcoin nodes do not let miners pick their own difficulty: Bitcoin Core's ContextualCheckBlockHeader rejects a header as "bad-diffbits" unless its nBits equals the value GetNextWorkRequired computes from the preceding chain.
- As of 20 November 2026, Bitcoin Core PR #35793, "Implement BIP 54 (Consensus Cleanup) without mainnet activation", is open and not merged; its description says the BIP 54 validation rules it implements are only enabled on regtest, with mainnet activation, if any, to be considered separately.
- BIP 54 proposes that the first block of each 2,016-block difficulty period (height divisible by 2,016) must have a timestamp no earlier than 7,200 seconds (two hours) before the timestamp of the block immediately preceding it, which the BIP gives as its fix for the timewarp attack.
- BIP 94 (Testnet 4) adds a timewarp rule: the first block of each difficulty period must not be timestamped more than 600 seconds before the previous block. Bitcoin Core enforces this rule only on testnet4 and regtest; its mainnet parameters set enforce_BIP94 to false.
- At the retarget at block 32,256, mined on 30 December 2009, Bitcoin's nBits changed from 0x1d00ffff (block 32,255) to 0x1d00d86a, raising mining difficulty from 1 to about 1.18.
- A 2018 statistical study of Bitcoin block arrival data (arXiv:1801.07447) reported that block arrivals do not follow the homogeneous Poisson process it says the original Bitcoin paper suggested, and proposed a refined model covering both periods of constant difficulty and how difficulty changes over time.
- Bitcoin Core counts each block's work as 2 to the power 256 divided by (target plus 1), the expected number of hashes needed to meet that target, and sums these values along a chain into nChainWork, the chainwork figure that decides which chain is best.
- Bitcoin limits each difficulty adjustment to a factor of four: Bitcoin Core clamps the measured timespan of a 2,016-block period to between one quarter and four times the two-week target before retargeting, so one retarget can at most quadruple or quarter the target.
- Bitcoin's mainnet difficulty can change only at every 2,016th block, so when total mining power rises or falls between adjustments, blocks arrive faster or slower than 10 minutes on average until the next adjustment.
- A Bitcoin block's hash is the double SHA-256 (SHA-256 applied twice) of its serialized header: Bitcoin Core's CBlockHeader::GetHash writes the header into a HashWriter, whose GetHash computes double SHA-256.
- Bitcoin mining costs real resources: the Bitcoin whitepaper compares issuing new coins through mining to gold miners spending resources, and names computer time and electricity as what Bitcoin miners spend.
- Bitcoin's genesis block (block 0, timestamp 1231006505, 3 January 2009) has nBits 0x1d00ffff, the compact form of mainnet's easiest permitted target, which corresponds to difficulty 1.
- Bitcoin Core's proof-of-work check in src/pow.cpp rejects a block header if the target decoded from its nBits is zero, negative, overflowing or above the network's proof-of-work limit, or if the header hash, read as a 256-bit number, is greater than that target.
- A Bitcoin block header is 80 bytes: Bitcoin Core serializes six fields, nVersion (4 bytes), hashPrevBlock (32 bytes), hashMerkleRoot (32 bytes), nTime (4 bytes), nBits (4 bytes) and nNonce (4 bytes), and proof of work is computed over these 80 bytes only.
- Bitcoin Core rejects a block whose timestamp is not later than the median time past, which is the median of the timestamps of the previous 11 blocks, ending with the new block's parent.
- At commit f498a5a6, Bitcoin Core's GetMinimumTime raises the minimum block template timestamp on all networks to respect both the BIP 94 600-second timewarp limit and the BIP 54 rule that a period's last block is not earlier than its first; the BIP 54 part came from Bitcoin Core PR #35949, merged 7 September 2026.
- Bitcoin Core's mainnet parameters set nMinimumChainWork, a minimum total chain work documented in the code with the comment that the best chain should have at least this much work; at commit f498a5a6 its mainnet value is 0x145ec036acc5ba740052af1a0.
- Bitcoin nodes follow the valid chain with the most accumulated proof of work, not the chain with the most blocks: Bitcoin Core ranks candidate chain tips first by total chain work, so a shorter chain of harder blocks can outrank a longer chain of easier ones.
- A Bitcoin header's 4-byte nBits field encodes the 256-bit target in a compact floating-point-like form: the top byte is an exponent in base 256 and the lower 23 bits are the mantissa, so the target equals mantissa times 256 to the power of (exponent minus 3).
- At each mainnet retarget, Bitcoin Core's CalculateNextWorkRequired multiplies the previous target by the period's measured timespan, first limited to between one quarter and four times two weeks, divides by the two-week target timespan, and caps the result at the proof-of-work limit, so faster periods make the target smaller and mining harder.
- Bitcoin mainnet recalculates its proof-of-work target every 2,016 blocks: Bitcoin Core derives this interval as the target timespan of two weeks (1,209,600 seconds) divided by the target block spacing of 600 seconds.
- Bitcoin Core measures each 2,016-block difficulty period from the timestamp of its own first block to that of its last block, which spans 2,015 block intervals rather than 2,016, so consecutive measurement windows do not overlap. This off-by-one is the property the timewarp attack exploits.
- Bitcoin aims for one block about every 10 minutes on average: the network adjusts how hard mining is so that, as more or less computing power joins, blocks keep arriving at roughly that rate.
- BIP 54 (Consensus Cleanup) states that the timewarp bug lets an attacker with a majority of Bitcoin's hashrate arbitrarily lower mining difficulty by manipulating block timestamps, and that in the worst case the attacker can bring difficulty to its minimum within 38 days.
- Bitcoin mining is a guessing search: miners repeatedly change a number in a candidate block and hash it with SHA-256 until the result is at or below a target, and only a block that meets the target can extend the chain.
Held for the owner
Their text stays private until the owner approves it.
- 5 claims on proof-of-work
names Murch, Zawy; names Antoine Poinsot, Matt Corallo; names Ittay Eyal, Emin Gun Sirer; names Ayelet Sapirshtein, Yonatan Sompolinsky, Aviv Zohar
Rejected
6 candidates
- Block 32,256 was the first mainnet difficulty change.
research: did not fetch every earlier retarget header, so the claim states only the observed change
- Inter-block times follow an exponential distribution with a 10-minute mean and median near 6.9 minutes.
research: only found in sources outside the registry (glossaries, blogs); no registered source fetched with an exact quote
- The bitcoin whitepaper says an attacker's chance of catching up falls exponentially with each block.
research: overlapped attacker-must-redo-work; kept the claim to one idea
- Satoshi's first release chose the chain by block count and was later changed to chainwork.
research: learnmeabitcoin returned 403 and no primary (commit) was fetched
- Optech coverage of 2026 BIP 54 follow-ups (2106 timestamp relaxation, chain-length bounds, testnet5 draft).
research: not read in detail in this run; candidates for a follow-up run
- Optimal selfish mining profitability threshold of about 23.21 percent.
research: the number is not in the arXiv abstract that was fetched
Could not settle
- The chain.header bits expect values are written as hex strings ("1d00ffff", "1d00d86a") matching bitcoin-cli getblockheader; Esplora returns integers (486604799, 486594666). The checker's expected format should be confirmed.
- Whether BIP 54 will be proposed for mainnet activation and with what parameters; settled by an activation PR or a BIP revision.
- The exact Delving post wording and the BIP 94 raw mediawiki line for the 600-second rule were read through a summarizing fetch; the quote check will confirm them.
- Line ranges were omitted for chainparams.cpp, validation.cpp and blockstorage.cpp because exact line numbers were not confirmed.
Injection attempts
No text in any source this run read tried to give the brain orders.
Read
32 sources
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/pow.cpp
retarget, clamp, off-by-one window, proof-of-work check
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/kernel/chainparams.cpp
mainnet powLimit, timespan, spacing, enforce_BIP94, nMinimumChainWork, genesis
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/consensus/params.h
DifficultyAdjustmentInterval, nMinimumChainWork comment
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/consensus/consensus.h
MAX_TIMEWARP
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/chain.h
nChainWork, median time past, MAX_FUTURE_BLOCK_TIME
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/chain.cpp
GetBitsProof work per block
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/validation.cpp
ContextualCheckBlockHeader (bad-diffbits, time-too-old, BIP94 check)
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/node/blockstorage.cpp
CBlockIndexWorkComparator
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/node/miner.cpp
GetMinimumTime timewarp and Murch-Zawy template rules
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/primitives/block.h
header fields and serialization
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/primitives/block.cpp
CBlockHeader::GetHash
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/hash.h
double SHA-256 in HashWriter
- https://raw.githubusercontent.com/bitcoin/bitcoin/f498a5a6718d9157b6afb85412c3462439117702/src/arith_uint256.h
compact nBits format
- https://raw.githubusercontent.com/bitcoin/bips/a7f91321f25b5be8d93e901775621c21519e80b7/bip-0054.md
BIP 54 preamble, timewarp motivation, rules and rationale
- https://raw.githubusercontent.com/bitcoin/bips/a7f91321f25b5be8d93e901775621c21519e80b7/bip-0094.mediawiki
BIP 94 timewarp rule and status
- https://github.com/bitcoin/bitcoin/pull/29775
testnet4 PR
- https://github.com/bitcoin/bitcoin/pull/35949
miner Murch-Zawy rule, merged 7 September 2026
- https://github.com/bitcoin/bitcoin/pull/35793
BIP 54 implementation PR, open
- https://delvingbitcoin.org/t/great-consensus-cleanup-revival/710/1
timewarp mechanism as described by the proposal author
- https://bitcoinops.org/en/topics/time-warp/
expert coverage of time warp
- https://bitcoinops.org/en/topics/consensus-cleanup-soft-fork/
expert coverage of consensus cleanup, newsletter index
- https://bitcoinops.org/en/topics/testnet/
testnet4 history (not cited)
- https://bitcoinops.org/en/newsletters/2024/04/24/
off-by-one Q&A pointer
- https://bitcoinops.org/en/newsletters/2026/09/11/
coverage of Bitcoin Core
- https://bitcoinops.org/en/newsletters/2026/07/17/
BIPs
- https://bitcoinops.org/en/newsletters/2025/04/04/
draft BIP 54 coverage (summary only, not cited)
- https://bitcoin.org/bitcoin.pdf
whitepaper sections 4, 6, 7
- https://arxiv.org/abs/1801.07447
block arrival statistics
- https://arxiv.org/abs/1311.0243
selfish mining (Eyal and Sirer)
- https://arxiv.org/abs/1507.06183
optimal selfish mining strategies
- https://arxiv.org/abs/2003.00001
checked for block-time distribution wording (not cited)
- https://mempool.space/api/block-height/0
header data for blocks 0, 32,255, 32,256 and 975,744 via /api/block