Last Updated: February 2026

Most blockchain analysis starts and ends with the confirmed chain – the single linear sequence of blocks that every node agrees on. But the confirmed chain is not the only chain that existed. Every time two miners solve a block at nearly the same moment, the network briefly maintains competing versions of history. One version wins. The other becomes a stale block – discarded by nodes, invisible on block explorers, and ignored by most investigators. That discarded history is where double-spend evidence, attack signatures, and operational intelligence hide in plain sight.

Chain reorganizations – the process by which the network switches from one version of the chain to another – have enabled over $30 million in documented double-spend theft across Ethereum Classic, Bitcoin Gold, and Bitcoin SV between 2018 and 2021. In each case, the attacker exploited the gap between when an exchange credited a deposit and when the blockchain reorganized to erase it. Investigators who monitor stale blocks and reorg events can detect these attacks as they unfold, trace the double-spent funds to their destinations, and build the forensic timeline that connects the attack to its beneficiary.

At Crypto Trace Labs, our forensic team incorporates chain reorganization analysis into investigations involving exchange fraud, 51% attack recovery, and disputed transaction finality. This guide explains what orphaned blocks reveal, how investigators use reorg data, and what has changed since Ethereum moved to proof-of-stake.

What Are Orphaned Blocks and How Do They Differ from Stale Blocks?

The terminology around discarded blocks is widely confused in the cryptocurrency space, and precision matters for forensic work. A stale block is a valid block that lost the race to be included in the longest chain. It has a known parent on the main chain – it simply was not accepted first by the majority of the network. This is the technically correct term for what most people call “orphan blocks.”

A true orphan block – one whose parent is unknown to the receiving node – has been impossible on Bitcoin since Core v0.10 introduced headers-first synchronization in early 2015. Nodes now always know a block’s parent before downloading the full block data. Despite this, the term “orphan” persists because Bitcoin Core labels discarded block rewards as “orphaned” in its interface.

On Ethereum’s pre-Merge proof-of-work chain, the equivalent was called an uncle or ommer block. Unlike Bitcoin, Ethereum’s protocol rewarded uncle block miners with 1.75 ETH (compared to the full 2 ETH block reward), a design choice that encouraged mining participation from less well-connected miners. Since the September 2022 Merge to proof-of-stake, ommer blocks no longer exist.

Bitcoin’s stale block rate has declined significantly over the past decade. Approximately 60 stale blocks were recorded in 2017, dropping to 5 in 2018 and just 2 in 2019. This decline reflects improvements in mining pool communication infrastructure – particularly the FIBRE network and compact block relay – that reduce propagation delays. For investigators, this means natural stale blocks are now rare enough that a sudden spike in stale block frequency for any chain is itself a forensic signal worth investigating.

Anatomy of a 51% Attack: How a Reorg Unfolds

Understanding how chain reorganizations enable theft requires walking through the attack chronologically. The following timeline represents the structure documented across the Ethereum Classic, Bitcoin Gold, and Bitcoin SV attacks – the specific durations vary, but the sequence is consistent.

Hour 0:00 – Private mining begins. The attacker, controlling majority hashrate (rented or owned), begins mining blocks on a private chain that only they can see. They do not broadcast these blocks to the network. Meanwhile, the public chain continues producing blocks normally, unaware that a competing chain is being built in parallel.

Hour 0:05 – Deposit transaction broadcast. The attacker sends a large deposit to an exchange on the public chain. This transaction is included in a public block and begins accumulating confirmations. In the Bitcoin Gold attacks, these deposits ranged from 1,900 to 5,267 BTG per attack. On their private chain, the attacker creates a conflicting transaction sending the same coins back to their own wallet.

Hour 0:15 to 1:00 – Exchange credits the deposit. Once the deposit reaches the exchange’s confirmation threshold (6 confirmations for Binance’s BTG deposits in 2018, approximately 60 minutes), the exchange credits the attacker’s trading account. The attacker immediately converts the deposited cryptocurrency to a different asset – typically Bitcoin – and withdraws it to an external wallet they control.

Hour 1:00 to 2:00 – Private chain overtakes. The attacker continues mining their private chain until it contains more accumulated work than the public chain. Because they control majority hashrate, their private chain grows faster than the public chain despite starting later.

Hour 2:00+ – The reorg. The attacker broadcasts their private chain to the network. Nodes compare the two chains and, following the most-accumulated-work rule, switch to the attacker’s longer chain. Every block on the old public chain since the fork point becomes a stale block. The deposit transaction is erased from history and replaced by the conflicting transaction returning funds to the attacker. The attacker now holds both the withdrawn Bitcoin and the original deposited cryptocurrency.

The exchange is left holding a credit on their books for a deposit that no longer exists on the blockchain. Tracing where the double-spent funds went after the reorg – both the withdrawn Bitcoin and the returned deposit coins – is where on-chain forensic analysis begins.

Which 51% Attacks Have Investigators Documented?

Three cases demonstrate different scales and investigative responses to chain reorganization attacks.

Ethereum Classic – $10 million across four attacks

Ethereum Classic suffered its first 51% attack in January 2019 when an attacker executed 15 reorganizations, 12 of which contained double-spends totaling 219,500 ETC (approximately $1.1 million). Coinbase detected the attack and halted trading. Then in August 2020, three separate attacks within a single month caused combined losses exceeding $9 million. The deepest reorganization replaced 4,000 blocks – roughly two days of chain history. Coinbase responded by increasing ETC confirmation requirements to 5,676 blocks, and during the August attacks raised the threshold to approximately two weeks of confirmations. The cost to execute each attack was estimated at $200,000 in rented hashrate – a fraction of the funds stolen.

Bitcoin Gold – $18 million in exchange fraud

In May 2018, attackers stole 388,201 BTG (approximately $18 million) by targeting exchanges that credited deposits after just 6 confirmations. The attackers used 14 to 15 block deep reorganizations to reverse deposits after withdrawing the proceeds as Bitcoin. The estimated cost per reorganization was approximately 0.2 BTC ($1,700 at the time) – meaning each attack cost less than $2,000 to execute but yielded millions in stolen funds. A second round of attacks in January 2020 produced smaller but still significant losses of approximately $72,000.

Bitcoin SV – 100-block reorg and simultaneous triple chains

Between June and August 2021, Bitcoin SV suffered five separate 51% attacks. The most severe in August produced three simultaneously competing versions of the chain for approximately 12 hours. The initial reorganization was 100 blocks deep, wiping 570,000 transactions from the canonical chain. The attack cost was estimated at $5,200 per hour in rented hashrate. Exchange Bitmart sought a restraining order after 43 users were defrauded by double-spent BSV deposits, creating one of the few cases where 51% attack evidence was formally presented in legal proceedings.

How Do Investigators Detect Double-Spend Attempts Through Stale Blocks?

The forensic value of stale blocks lies in what they contain that the winning chain does not. When a block is orphaned, its transactions return to the mempool and most are re-included in subsequent canonical blocks. The critical exception is when a transaction appears in the stale block but a conflicting transaction – spending the same inputs to different outputs – appears in the winning chain. That conflict is direct evidence of a double-spend, whether successful or attempted.

Investigators detect these conflicts through several methods. Real-time reorg monitoring tools compare the transaction sets between competing chain tips. When a reorganization occurs, automated systems diff the transaction lists to identify any inputs spent differently between the two chains. Each conflicting pair produces an attribution lead: the address that received funds in the stale block version (typically an exchange deposit address) versus the address that received funds in the winning version (typically the attacker’s own wallet). Address clustering then connects the attacker’s receiving address to their broader wallet infrastructure.

Stale block analysis also supports selfish mining detection. A seminal 2014 paper by Eyal and Sirer demonstrated that selfish mining – withholding blocks to gain disproportionate rewards – is profitable at approximately 33% hashrate. A 2024 study in Nature Scientific Reports provided the first empirical evidence that selfish mining occurs in practice, identifying abnormal mining patterns on Monacoin and Bitcoin Cash through statistical analysis of orphan rates by mining pool. If a single pool produces a disproportionate share of stale blocks relative to its hashrate, it suggests deliberate block withholding.

Crypto Trace Labs applies reorg analysis in exchange fraud investigations where the victim suspects a deposit was reversed after confirmation. The forensic report documents the exact block heights involved, the conflicting transactions, the timing of the attacker’s withdrawal relative to the reorg, and the transaction graph connecting the double-spent funds to identifiable endpoints.

How Does Ethereum’s Proof-of-Stake Finality Change the Landscape?

Ethereum’s September 2022 Merge to proof-of-stake fundamentally altered the orphaned block equation. Under proof-of-work, mining was a race – two miners could solve a block simultaneously, creating competing chains. Under proof-of-stake, one validator is selected per 12-second slot to propose a block. There is no race, and therefore no natural mechanism for producing competing valid blocks at the same height.

More significantly, Ethereum now achieves deterministic finality after two epochs (approximately 12.8 minutes). Once two-thirds of all staked ETH attests to two consecutive checkpoint blocks, those blocks are finalized and cannot be reversed without destroying at least one-third of all staked ETH – currently worth billions of dollars. This makes 51% attack-style reorgs economically impossible at Ethereum’s scale, a stark contrast to proof-of-work chains where attack costs are measured in thousands of dollars per hour of rented hashrate.

The closest equivalent to orphaned blocks on post-Merge Ethereum is the missed slot – when the assigned validator fails to propose a block. The overall missed slot rate sits at approximately 0.9% of total slots, with validator participation consistently above 99%. Missed slots do not create competing chains or double-spend opportunities. They simply mean one 12-second interval passed without a block.

Ethereum has experienced brief reorg events under proof-of-stake – including a 7-block reorganization in May 2022 caused by different client implementations rolling out updates at different times, and two finality delays in May 2023 lasting up to an hour. Neither was a security attack, and both were resolved through client software updates. For investigators tracking assets across multiple chains, the key takeaway is that Ethereum post-Merge is essentially immune to the reorg-based double-spend attacks that continue to threaten lower-hashrate proof-of-work chains.

What Tools Monitor Chain Reorganizations?

Investigators and exchanges use several monitoring systems to detect reorgs as they occur. ForkMonitor, sponsored by BitMEX Research and launched in 2018, connects to multiple Bitcoin nodes running different software versions and detects consensus divergences between them. Coin Metrics’ Reorg and Fork Tracker monitors conditions at the blockchain tip across multiple chains, providing real-time alerts when reorganization events occur. For Ethereum specifically, open-source tools like the Ethereum Reorg Tracker detect chain reorganizations through parent hash mismatches between consecutive blocks.

Bitcoin Core’s getchaintips RPC command returns information about all known chain tips, including stale branches – but only if the node was running during the reorg event. Historical stale block data is maintained by the community at bitcoin-data/stale-blocks on GitHub. For forensic investigation firms, the critical infrastructure requirement is maintaining archival nodes that capture stale block data in real time, because standard block explorers typically discard this information.

Exchange confirmation requirements reflect the practical risk calculus. Coinbase requires 2 confirmations for Bitcoin deposits and 14 for Ethereum. Binance requires 1 confirmation for Bitcoin deposits and 12 for Ethereum. For chains that have suffered 51% attacks, requirements increase dramatically – Coinbase raised Ethereum Classic confirmations to over 5,000 blocks during the 2020 attacks. These thresholds are set so that the cost of a reorg attack exceeding the confirmation count is higher than the maximum depositable value, making the attack unprofitable.

Frequently Asked Questions

What happens to my transaction if it was in an orphaned block?

In the vast majority of cases, nothing. When a block is orphaned, its transactions return to the mempool and are re-included in subsequent canonical blocks within minutes. The only scenario where a transaction is permanently lost is if a conflicting transaction (spending the same UTXOs to different addresses) appears in the winning chain – which indicates a deliberate double-spend, not a natural stale block event.

Can Bitcoin be 51% attacked?

Theoretically yes, but the cost is prohibitive. Bitcoin’s hashrate requires billions of dollars in mining hardware and electricity to achieve majority control. No confirmed 51% attack has ever occurred on Bitcoin. The 24-block reorganization in March 2013 was caused by an accidental software bug (BIP-50), not a deliberate attack. Lower-hashrate chains like Ethereum Classic, Bitcoin Gold, and Bitcoin SV have been successfully attacked because renting sufficient hashrate costs orders of magnitude less.

How many confirmations protect against a reorg?

It depends on the chain’s hashrate and the attacker’s resources. For Bitcoin, 6 confirmations (approximately 60 minutes) is the long-standing standard, providing high confidence against all but state-level attackers. For lower-hashrate chains, exchanges may require hundreds or thousands of confirmations. The calculation is economic: confirmations must make the attack cost exceed the potential profit from a double-spend.

Do investigators have access to orphaned block data?

Only if they were capturing it in real time. Standard block explorers and most blockchain analysis tools only index the canonical chain. Stale block data must be captured from full nodes at the time of the reorg event – once the network resolves to a single chain, the stale blocks are discarded from most nodes’ active storage. This is why Crypto Trace Labs maintains archival node infrastructure that preserves stale block data for forensic purposes.

What is selfish mining and has it been proven?

Selfish mining is a strategy where a miner withholds discovered blocks, secretly building a private chain to orphan competitors’ work and claim disproportionate rewards. The 2014 Eyal-Sirer paper proved it is theoretically profitable at approximately 33% hashrate. A 2024 study in Nature Scientific Reports found empirical evidence of selfish mining on Monacoin and Bitcoin Cash, but no peer-reviewed study has confirmed selfish mining on Bitcoin mainnet.

Does Ethereum still have orphaned blocks after the Merge?

No. Under proof-of-stake, one validator is selected per slot to propose a block – there is no mining race that could produce competing blocks. The closest equivalent is a missed slot (approximately 0.9% of all slots), where no block is produced at all. Ethereum achieves deterministic finality after approximately 12.8 minutes, making reorg-based double-spend attacks economically impossible at its current staking scale.

Need Help Tracing Funds Through Chain Reorganization Events?

If your exchange, payment platform, or custodial service has experienced a suspected double-spend through a chain reorganization, or if you need to investigate transaction finality disputes on any blockchain, professional forensic analysis can reconstruct the reorg timeline, identify the conflicting transactions, and trace the double-spent funds to their destination.

Crypto Trace Labs conducts chain reorganization forensics across Bitcoin, Ethereum, and all major proof-of-work chains. Our forensic team – including analysts like D. Hargreaves – holds ACAMS certifications, MLRO qualifications across UK, US, and European jurisdictions, and Chartered status at Fellow Grade. Our founders held VP and Director positions at Blockchain.com, Kraken, and Coinbase.

Contact Crypto Trace Labs to discuss your case with our forensic investigation team.

About the Author

This guide was prepared by the blockchain forensics team at Crypto Trace Labs. Our founding members held VP and Director-level positions at Blockchain.com, Kraken, and Coinbase, bringing over 10 years of combined experience in cryptocurrency operations, on-chain analysis, and forensic investigation. Our team holds ACAMS certifications, MLRO qualifications across UK, US, and European jurisdictions, and Chartered status at Fellow Grade. We have analyzed vanity address exploitation patterns in hundreds of investigations and provided expert witness testimony on blockchain attribution methodologies in court proceedings.

This content is for informational purposes only and does not constitute legal, financial, or compliance advice. Crypto asset recovery outcomes depend on specific circumstances, regulatory cooperation, and technical factors. Consult qualified professionals regarding your situation.