Chapter 27

Historical Hard Forks

A Post-Mortem Analysis

Bitcoin's history is marked by several contentious hard forks, each representing a fundamental disagreement about the protocol's direction. This chapter provides a rigorous analysis of major chain splits—their technical motivations, implementation details, economic outcomes, and lessons learned. We examine these forks not to judge their merits but to understand the dynamics of decentralized protocol evolution.

27.1 The Block Size Debate: Background

To understand Bitcoin's hard forks, we must first understand the conflict that produced them: the block size debate.

Definition 27.1 (Block Size Limit)

Bitcoin's consensus rules include a maximum block size, originally bounded only by the 32 MB network message limit, then set to 1 MB in 2010 (commit 172f006):

if (block.GetSerializeSize() > MAX_BLOCK_SIZE)
    return error("CheckBlock(): size limits failed");

This limit was later redefined as 4 million weight units (SegWit, 2017), allowing blocks up to ~4 MB with witness data.

Remark 27.1 (The Scaling Trilemma Applied)

The block size debate is a manifestation of the scaling trilemma:

  • Throughput: Larger blocks → more transactions per second
  • Decentralization: Larger blocks → higher node costs → fewer nodes
  • Security: Fewer nodes → increased centralization risk

The debate centered on where to position this tradeoff.

2010 1MB limit added 2015 Bitcoin XT 2016 Classic/ Unlimited Aug 2017 Bitcoin Cash Aug 2017 SegWit activates Nov 2018 BCH/BSV split Soft fork Proposal Hard fork
Figure 27.1: Timeline of the block size debate and resulting forks.

Definition 27.2 (The Two Scaling Philosophies)

On-chain scaling: Increase block size to accommodate more transactions directly on the base layer. Advocates argued this was Satoshi's original vision and necessary for mainstream adoption.

Layer-2 scaling: Keep base layer small and secure; scale via payment channels (Lightning), sidechains, and other off-chain solutions. Advocates argued this preserves decentralization and enables novel use cases.

27.2 Failed Fork Attempts (2015–2017)

Before Bitcoin Cash, several attempts to increase the block size failed to gain sufficient support.

27.2.1 Bitcoin XT (August 2015)

Example 27.1 (Bitcoin XT)

Bitcoin XT was a fork of Bitcoin Core by Mike Hearn and Gavin Andresen implementing BIP-101:

  • Initial 8 MB block size
  • Doubling every 2 years until 8 GB
  • Activation at 75% miner support

Remark 27.2 (XT Failure Analysis)

Bitcoin XT failed due to:

  1. Insufficient miner signaling: Never approached 75% threshold
  2. Community opposition: Perceived as hostile takeover
  3. Technical concerns: Exponential growth criticized as reckless
  4. DDoS attacks: XT nodes were targeted; the attackers were never identified

Peak XT node share: ~12% (late 2015), then rapid decline.

27.2.2 Bitcoin Classic (February 2016)

Example 27.2 (Bitcoin Classic)

Bitcoin Classic proposed a more conservative increase:

  • 2 MB block size (one-time increase)
  • Activation at 75% miner support over 28 days
  • Positioned as compromise solution

27.2.3 Bitcoin Unlimited (announced December 2015, released January 2016)

Example 27.3 (Bitcoin Unlimited)

Bitcoin Unlimited took a more sweeping approach: no hard-coded limit at all. Instead:

  • Miners set their own limits ("Emergent Consensus")
  • Market forces determine equilibrium block size
  • Users configure acceptance depth for larger blocks

Remark 27.3 (Emergent Consensus Critique)

Critics of "Emergent Consensus" identified a game-theoretic objection: without a Schelling point (hard limit), miners have incentive to produce maximum-size blocks to:

  1. Collect more fees (more transactions)
  2. Disadvantage smaller competitors (propagation delay)

This could lead to unbounded growth and centralization.

27.3 Bitcoin Cash (BCH)

Bitcoin Cash is the largest persistent Bitcoin hard fork by market value.

Example 27.4 (Bitcoin Cash Fork)

Bitcoin Cash forked at block 478,558 (August 1, 2017):

  • Block size: 8 MB (later 32 MB)
  • Replay protection: SIGHASH_FORKID (new sighash type)
  • Difficulty: Emergency Difficulty Adjustment (EDA)
  • SegWit: Not included (philosophical rejection)
Block 478,558 Fork Point Aug 1, 2017 BTC 1MB + SegWit (~4MB weight) → continues BCH 8MB blocks (no SegWit) → continues
Figure 27.2: Bitcoin Cash fork from block 478,558.

27.3.1 Emergency Difficulty Adjustment

Definition 27.3 (EDA Algorithm)

The Emergency Difficulty Adjustment reduced difficulty by 20% if fewer than 6 blocks were mined in 12 hours:

if (MTP(tip) − MTP(tip − 6) > 12 hours):
    difficulty = difficulty × 0.8

The trigger compared median-time-past values (Definition 13.5), not raw timestamps.

Remark 27.4 (EDA Instability)

The EDA created oscillating dynamics:

  1. Low BCH price → miners leave → slow blocks → EDA triggers
  2. Difficulty drops → BCH more profitable than BTC
  3. Miners switch to BCH → fast blocks → difficulty rises
  4. BCH less profitable → miners return to BTC → cycle repeats

Result: more than 100,000 "extra" BCH mined ahead of schedule by the time the EDA was retired (November 2017).

Example 27.5 (DAA Replacement)

Bitcoin Cash replaced EDA with the DAA (Difficulty Adjustment Algorithm) in November 2017:

  • Adjusts difficulty every block
  • Uses 144-block rolling window
  • Targets 600 seconds per block
  • Smooths oscillations

27.3.2 BCH Economic Performance

Metric At Fork (2017) Peak (2017) 2024
Price (USD) ~$300 ~$4,000 ~$250
% of BTC price 10% 25% 0.5%
Hash rate (% of BTC) ~5% ~50-100% briefly (EDA oscillations) ~1%
Daily transactions ~10,000 ~50,000 ~30,000

27.4 Bitcoin SV (BSV)

Bitcoin SV emerged from a schism within the Bitcoin Cash community.

Example 27.6 (Bitcoin SV Fork)

Bitcoin SV ("Satoshi Vision") forked from BCH at block 556,766 (November 15, 2018):

  • Block size: 128 MB (later removed entirely)
  • Philosophy: Return to "original" Bitcoin protocol
  • Restoration: Re-enabled some disabled opcodes
  • Leader: Craig Wright / Calvin Ayre

Remark 27.5 (The Hash War)

The BCH/BSV split involved an unusual "hash war" in which both sides threatened to attack each other's chain:

  • Both chains mined at a loss to accumulate work
  • BSV side threatened 51% attacks
  • Exchanges suspended trading during uncertainty
  • Eventually, both chains coexisted with replay protection
BCH Nov 15, 2018 BCH (ABC) 32MB blocks BSV 128MB+ blocks → BCH → BSV 2020: eCash fork Delisted from major exchanges
Figure 27.3: BCH → BSV fork and subsequent developments.

27.4.1 Technical Differences

Feature BCH BSV
Block size 32 MB Unlimited (4GB tested)
OP_RETURN limit 220 bytes Unlimited
Script opcodes Some re-enabled More re-enabled
Development model Multiple implementations nChain-led
Major exchanges (2024) Listed Largely delisted

Remark 27.6 (BSV Controversies)

BSV faced unique challenges:

  • Association with Craig Wright's disputed Satoshi claims
  • Legal threats against critics
  • Delisting from Binance, Kraken, ShapeShift
  • Multiple 51% attacks (2021)
  • Centralization concerns (few large miners)

27.5 Bitcoin Gold (BTG)

Bitcoin Gold represents a different type of fork: changing the mining algorithm.

Example 27.7 (Bitcoin Gold Fork)

Bitcoin Gold forked at block 491,407 (October 24, 2017):

  • Algorithm: Equihash (ASIC-resistant) instead of SHA-256d
  • Goal: Democratize mining (GPU-friendly)
  • Block size: 1 MB (unchanged)
  • Premine: 100,000 BTG for development

Remark 27.7 (Limits of ASIC Resistance)

Changing the proof-of-work algorithm to be "ASIC-resistant" provides only temporary protection:

  1. No fixed algorithm is known to resist ASIC implementation indefinitely
  2. GPU-mineable algorithms have historically attracted botnet hash power (stolen compute)
  3. Algorithm changes cause disruptive hard forks
  4. Smaller hash rate increases 51% attack vulnerability

Example 27.8 (BTG 51% Attacks)

Bitcoin Gold suffered multiple 51% attacks:

Date Double-Spend Reorg Depth
May 2018 ~$18 million ~20+ blocks
January 2020 ~$72,000 14-15 blocks (two reorgs)
July 2020 none (thwarted) ~1,300-block secret chain rejected via checkpoint at 640,650

These attacks demonstrated the security cost of low hash rate.

27.6 Other Notable Forks

27.6.1 Bitcoin Diamond (BCD)

Example 27.9 (Bitcoin Diamond)

Bitcoin Diamond forked in November 2017 with:

  • 10× coin supply (210 million)
  • 8 MB blocks
  • X13 proof-of-work algorithm
  • Encrypted transaction amounts (claimed)

Remark 27.8 (Bitcoin Diamond Reception)

Contemporary press coverage characterized Bitcoin Diamond as a "fork for profit"; development activity declined to minimal levels within a few years of launch.

27.6.2 Bitcoin Private (BTCP)

Example 27.10 (Bitcoin Private)

Bitcoin Private was a fork-merge combining Bitcoin and Zclassic (February 2018):

  • zk-SNARK privacy features from Zcash
  • Combined UTXO sets of BTC and ZCL
  • Equihash algorithm

A CoinMetrics analysis (December 2018) identified approximately 2.04 million covertly premined coins.

27.6.3 eCash (XEC)

Example 27.11 (eCash)

eCash forked from Bitcoin Cash in November 2020:

  • Avalanche consensus layer (post-consensus)
  • Redenomination (1 BCH = 1,000,000 XEC)
  • Development funded by block reward
  • Led by Bitcoin ABC team

Remark 27.9 (The Long Tail of Forks)

Dozens of Bitcoin forks have been created. Most share characteristics:

  • Minimal development post-fork
  • Value decay relative to BTC
  • Low liquidity and exchange support
  • Security vulnerabilities from small hash rate

The "free coins" distribution model proved insufficient for building sustainable networks.

27.7 The 2017 SegWit2x Attempt

The most consequential fork attempt is one that did not happen: SegWit2x.

Example 27.12 (New York Agreement / SegWit2x)

The New York Agreement (May 2017) was a compromise signed by major companies and miners:

  1. Activate SegWit (soft fork)—achieved August 2017
  2. 2 MB base block size hard fork within 6 months

The second part (SegWit2x) was abandoned days before activation.

Remark 27.10 (SegWit2x Failure Factors)

SegWit2x failed despite ~90% miner signaling because:

  1. No replay protection: Intentionally omitted, seen as aggressive
  2. Core developer opposition: None of Bitcoin Core signed NYA
  3. User opposition: Futures markets showed BTC premium over B2X
  4. Business defections: Key signatories withdrew support

Remark 27.11 (Lessons from SegWit2x)

The episode is widely read as showing:

  • Miner signaling is necessary but not sufficient
  • Users (economic nodes) have effective veto power
  • Futures prices provided one public signal of relative valuation (though the B2X futures market was thin)
  • Hard forks require broader social consensus than soft forks

Its failure strengthened the "User Activated" model of consensus.

27.8 Comparative Analysis

A systematic comparison reveals patterns in fork outcomes.

Fork Type % BTC Value (Peak) % BTC Value (2024) Status
Bitcoin Cash Ideological ~25% ~0.5% Active
Bitcoin SV Ideological ~5% ~0.1% Marginal
Bitcoin Gold Technical ~5% ~0.02% Marginal
Bitcoin Diamond Commercial ~1% ~0.0002% Inactive development
Bitcoin Private Commercial ~1% N/A Defunct
SegWit2x Political N/A N/A Never launched

Remark 27.12 (Fork Value Decay)

Empirically, Bitcoin forks exhibit value decay relative to BTC following a pattern consistent with:

Vfork(t) / VBTC(t) ∝ e^(−λt)

where λ varies by fork; the endpoint ratios imply average decay rates of approximately λ ≈ 0.4/year for BCH and λ ≈ 0.7/year for BTG.

This suggests that network effects compound over time, increasingly favoring the dominant chain.

27.9 Lessons and Principles

Remark 27.13 (Fork Success Factors)

Analysis of historical forks suggests success factors:

  1. Clear value proposition: Must offer something BTC cannot
  2. Technical competence: Must maintain security and development
  3. Community: Must build independent user base
  4. Timing: Early forks captured more value
  5. Replay protection: Essential for user safety

Remark 27.14 (Why Hard Forks Rarely Succeed)

The historical record suggests hard forks face structural disadvantages:

  • Network effects strongly favor incumbents
  • Developer talent tends toward dominant chain
  • Liquidity fragmentation harms both chains initially
  • User confusion degrades trust
  • Minority chain faces security challenges

Soft forks, by maintaining compatibility, avoid these problems.

Remark 27.15 (The Fork Right)

Despite their typically poor outcomes, the ability to fork remains important:

  • Provides exit option, limiting majoritarian abuse
  • Creates natural experiment for different approaches
  • Demonstrates that Bitcoin is not controlled by any group
  • Historical forks strengthened BTC's Schelling point

The ability to fork, though rarely exercised successfully, serves a structural function: it bounds what any majority can impose.

Exercises

Exercise 27.1

Calculate the cost (in USD) to perform a 51% attack on Bitcoin Gold for 1 hour, given that BTG hash rate is approximately 2 MS/s (2,000 kS/s) and Equihash ASIC rental costs $0.10 per kS/s per hour. Compare to the cost of attacking Bitcoin.

Exercise 27.2

Design a replay protection mechanism that works without changing the transaction format. What are the tradeoffs of your approach?

Exercise 27.3

Model the "difficulty death spiral" mathematically. Given a minority chain starting with 10% hash rate, Bitcoin's 2016-block adjustment period, and miners who switch chains based on profitability, simulate the first month of the minority chain. Under what conditions does it survive?

Exercise 27.4

Analyze the SegWit2x futures markets. Given that B2X futures traded at 0.15 BTC before cancellation, what did this imply about market expectations? How do futures markets serve as coordination mechanisms?

Exercise 27.5

Research the BCH "hash war" of November 2018. Both sides mined at a significant loss. Model this as a game theory problem: what were the payoffs, and why did both sides eventually accept coexistence?