The number was so low it read like a typo. Under three percent. That's how much miner support Bitcoin's BIP-110 had gathered when it entered its mandatory signaling phase — a phase designed to compel. Not persuade. Not negotiate. Compel.
I've seen that number before in my own work. Auditing ERC-20 token contracts in 2017, I spent four months tearing through implementations for three Cape Town projects, and I learned that the worst vulnerabilities never live in the functions. They live in the assumptions. And assumption failures always show up in the quiet numbers — the ones nobody quotes in the press release.
Less than three percent is not a vote. It is a whisper that says: this change has no constituency. And yet the protocol moved forward anyway. That disconnect between what the code demanded and what the network actually wanted turned BIP-110 into one of Bitcoin's most instructive governance artifacts.
Here is the history most people have forgotten. There are two ways to activate a soft fork on Bitcoin. You can ask miners to signal their willingness over a difficulty period, as BIP-9 eventually standardized, requiring 95 percent of hashing power to signal before activation. This is negotiation — slow, transparent, and forgiving. Or you can do what BIP-110 attempted: set a deadline after which nodes simply reject blocks lacking a designated version bit. No discussion. No opt-out. Just enforcement.
BIP-110 was an early exploration of that second path. It belongs to the same philosophical family as user-activated soft forks, the idea that full nodes — not miners — should hold ultimate power to enforce protocol rules. On paper, this sounds elegant. Nodes are the validators. They check everything. Why should the people who burn electricity get to dictate the rules? That question, it turned out, had an answer written in hashpower. And the three percent figure spelled it out plainly.
When a protocol's enforcement mechanism and its production layer diverge, you don't get an upgrade. You get two competing views of the same chain. In BIP-110's framework, after the mandatory window, a node running the new software would reject any block that lacked the version bit. Meanwhile, miners producing blocks without it were still operating honestly from their perspective, building on what they believed was the canonical chain. If both sides persisted, the network would fragment. Transactions confirmed on the miner-supported chain would be invisible to enforced nodes, and vice versa. That is not a soft fork. That is a civil war expressed in block headers.
The existence of a hard fork rollback plan — buried in the proposal's contingency language — revealed that the developers themselves knew the odds. You don't write a rollback plan for a change you are confident about. You write one because you expect to lose. And losing, in this context, meant one of two outcomes: the proposal silently dies, or the network cracks open. The rollback plan was the acknowledgment that both outcomes were possible.
Let me walk through what the low support number actually means at the protocol layer, because I believe most commentary has misread it entirely.
First, the three percent figure represents a signal failure, not an intellectual rejection. In proof-of-work systems, miner behavior is driven by incentive clarity. If an upgrade does not offer block-space advantages, fee improvements, or economic rewards, most hashing power will simply ignore it. This is not malice. It is economics. The majority of miners never evaluated BIP-110 as a technical proposal. They ran whatever client their pool shipped, and their pool software did not signal. That distinction matters: a quiet default is not a coordinated boycott, but it functions identically at the network level.
Second, the mechanism itself created a new class of consensus risk. In a scenario where mandatory signaling activated with less than three percent cooperation, any block produced by nonsignaling miners — more than 97 percent of production capacity — would be rejected by signaling nodes. The chain would stall. In extreme cases, you would see a period of dead air: no valid blocks, no confirmations, followed possibly by mass reorganizations. For a network whose core value proposition is settlement finality, this is existential. The source analysis flagged this as the central risk — chain stall and potential network split — and that assessment holds up even from a distance of years.
Now compare BIP-110 to what Bitcoin actually adopted. BIP-9 requires 95 percent of hashing power to signal version bits over a retarget period. If the threshold is not met, the soft fork simply does not activate. No conflict. No rival chains. No drama. It is a governance mechanism designed to fail quietly. BIP-110 was designed to fail loudly — or succeed by force. The contrast is not merely technical. It is philosophical. BIP-9 encodes the belief that miners are stakeholders whose consent must be courted. BIP-110 encodes the belief that miners are executors whose behavior can be commanded. Bitcoin's subsequent history — SegWit via BIP-9, Taproot via BIP-9's successor — tells you which belief survived contact with reality.
There is something else hiding beneath the surface that I want to pull out. The mining pools, not individual miners, are the real coordination layer. A support rate below three percent means the major pools never bothered to upgrade their software or signal enthusiasm. This was not a grassroots miner uprising. It was a structural indifference cascading up from pool-level defaults. In decentralized protocols, silence aggregates quietly. By the time it becomes visible, it is already decisive. This is the hidden information in every governance failure: what looks like a protest is almost always just a default.
And here is where I add something from my own experience. When I ran my DeFi education workshops in Cape Town during the summer of 2020, I taught over two hundred local residents how liquidity pools work. The most common failure I saw was not technical ignorance. It was people entering positions without understanding the incentive alignment — or misalignment. Yield farmers jumped into pools because the APY looked big, then discovered impermanent loss when the market moved against them. The panic that followed was not a code bug. It was an assumption bug.
BIP-110 was the same story at the consensus layer. Its authors assumed miners would fall in line because the developers held the keys to the software. What they underestimated was the simple, immovable truth of economic self-determination. Miners are not employees. They are counterparties. You cannot enforce a contract on a counterparty who never signed it. Every line of code is a hand extended in trust, and when the other side does not extend its hand back, you withdraw. Gracefully. That is what happened with BIP-110.
Now comes the contrarian angle, and I am aware it will irritate people on both sides of the old debate. I am going to argue that BIP-110's failure was, in the long arc of Bitcoin's governance history, a success. And this is the angle I believe matters most for the current bull market.
The mandatory signaling phase of BIP-110 functioned as a pressure test — and the entire ecosystem passed. Miners demonstrated, through quiet nonparticipation, that they hold a structural veto over protocol changes. Nodes demonstrated that they could propose enforcement but not unilaterally win a legitimacy fight. The result was a resolution mechanism that required neither a contentious chain split nor a devastating reorg. In the history of governance failures, this is a remarkably graceful outcome. The proposal was abandoned because the economics said so. That is a governance system working. Not because it was pretty, but because it was accurate.
For anyone skeptical of this reading, consider what came next. The lessons from BIP-110's collapse informed BIP-9's design philosophy: threshold-based activation, no enforcement without consent, no timeline pressure. SegWit activated after a long and bruising debate, but it activated without a chain split. Taproot followed with almost no controversy because the mechanism had matured. The three percent figure did not need to become a confrontation, because the network's economic layer had already communicated its answer. The code listened.
This is the blind spot that the current bull market keeps hitting. We are watching projects launch with governance tokens nobody genuinely controls, liquidity programs designed by venture capital, and community votes that resemble participation theater. The same people who mock Bitcoin as slow and conservative are shipping protocol changes with executive authority. When their communities resist, they fork or deploy new chains, treating consent as a technical inconvenience rather than the core constraint it actually is. Tracing the code back to the conscience behind it was always the heart of open-source legitimacy, but BIP-110 proves something more concrete: the market has a veto, and it will exercise that veto through indifference long before it resorts to conflict. Open source is not a license; it is a promise. And the promise is that change must be earned, not mandated.
So what does this mean for you, reading this in a bull market where every project claims to be decentralized? There is a practical question embedded in BIP-110's story. Are you earning your consensus or assuming it? If your governance mechanism has an enforcement clause but no organic constituency, you are rebuilding BIP-110 without the humility that its rollback plan implied. And the three percent outcome is not the worst-case scenario. The worst-case scenario is the one where the third percent rises to forty, your network splits, and both sides spend years pointing at each other as the origin of the schism.
I think about the governance signal in BIP-110 every time I see a project announce an upgrade by fiat. The source analysis marked the miner support number as a voting participation rate, and that framing gets to the heart of it. In decentralized systems, participation is the only power. Everything else is just code waiting for permission. Education is the only true decentralized currency, and BIP-110 was, in its strange way, an education. It taught a generation of protocol designers that the math of consensus is not technical. It is social. We build bridges, not just blocks, between people.
So the next time someone tells you that a protocol upgrade is inevitable, or that the network will simply adapt because the code says so, remember that number. Less than three percent. It was all anyone needed to know. The question is whether the next project making that mistake will be humble enough to include a rollback plan.
