Look at the version bits in that activation window. Block after block, timestamp bleeding into timestamp, each header carrying the same quiet absence: no BIP-110 flag, no upgrade intent, no commitment to the new consensus rule. Just the steady, indifferent hum of hashrate doing what hashrate does. Miner support had collapsed to below 3%. A mandatory signaling phase โ a mechanism engineered to compel the network's block producers into compliance โ had opened its gate to an empty field. The silence in the block headers was louder than any announcement.
This is the story of BIP-110, the soft-fork activation mechanism that almost forced Bitcoin's hand. Not as a footnote in the BIP graveyard, but as a case study in what happens when protocol architects attempt to legislate miner behavior without aligning miner economics. Following the ghost in the side-channel shadows, I want to re-read this failure as a crucial dataset for anyone building chain governance today. The same tension is playing out across every L1, L2, and restaking layer arguing over who gets to write consensus rules.
BIP-110 was drafted in the pre-SegWit era by Bitcoin Core-adjacent contributors, at a time when the network's governance machinery was creaking under the weight of the block size debate. It proposed something radically different from the activation mechanisms that had come before. Where BIP-9 would eventually require 95% of miner hashrate to signal readiness over a full difficulty period, BIP-110's design was adversarial in its simplicity: after a pre-agreed window, nodes running the upgraded client would simply reject any block that failed to carry the specified version bit. No miner consultation. No threshold negotiation. A client-side mandate, executed at the consensus boundary.
This was not a scaling proposal. It offered no transaction throughput gains, no fee market restructuring, no cryptographic leap. It was pure governance technology โ a weapon in the escalating conflict between the developer class and the mining industrial complex that defined Bitcoin's 2015-2017 civil war. The mandatory signaling phase was designed to answer one empirical question: when nodes impose a rule, do miners follow?
The test produced an unambiguous result. Fewer than 3% of blocks carried the signal. The miners had answered โ not with a manifesto, but with silence.
Let me be precise about what three percent means in a Proof-of-Work network. Miners are not a voting bloc in any democratic sense; they are the physical producers of canonical chain reality. When a mandatory signal requirement activates, every node executing the updated logic will reject any block missing the required version bit. At full compliance, this is a clean soft fork. At 3% compliance, it is a self-inflicted partition: the enforcing minority cannot see a valid chain, while the non-enforcing majority continues producing blocks on a fork that the enforcers refuse to recognize. The network does not split cleanly; it seizes, stutters, and produces two competing realities until one side capitulates.
This is where my pre-mortem habits kick in. Based on my experience stress-testing Lido's staking derivatives against a 40% ETH drawdown in 2022, I built a modular failure model for mandatory signaling to understand what happens when protocol designers assume compliance. The simulation logic was simple: if the mandatory window opens with low miner support, upgraded nodes orphan valid production; if the honest chain becomes the invisible chain, the enforcing minority holds an empty ledger; if the enforcing minority persists, the protocol stops being Bitcoin and becomes a hashwar without the hashpower to wage it. BIP-110's contemporaneous discussion threads confirmed the developers knew this โ the hard-fork fallback plan was already being shaped before the window fully opened. They built the retreat path before they built the offense.
I have seen this failure mode before, in a different key. During my audit of Groth16 proof verification in 2017, I found an edge-case vulnerability that only mattered if an adversary controlled the timing of node synchronization. The core devs dismissed it as irrelevant; the vulnerability was real but the conditions were unlikely. BIP-110 had the opposite problem: the conditions were guaranteed, but the failure was treated as hypothetical. Protocol designers keep making the same error โ they model adversarial intent instead of modeling indifference. The Zcash case and the BIP-110 case converge on one principle: the most dangerous assumption in any consensus design is that participants will care enough to comply.
Mapping the topology of hidden incentives: the miners' 3% support rate was not a technical verdict. It was an economic one. BIP-110 offered miners nothing: no fee uplift, no block size relief, no hardware decentralization advantage. It demanded a costly signal of subordination to node authority in exchange for zero improvements to the mining business. With electricity costs, depreciation schedules, and pool market share on the line, the expected value of adopting the signaling upgrade was negative. The rational response was not resistance; it was indifference. A sophisticated opponent would have coordinated a counter-campaign. Instead, mining pools simply declined to update their default client. No drama. No committees. A decision made in the quiet arithmetic of operational cost.
And this is the part most retrospectives miss. Interrogating the consensus of the crowd, the reflexive narrative is "miners rebuked developer authoritarianism." But at below 3% support, organized opposition is statistically implausible. Every serious mining pool failed to signal. That is what a cost-benefit calculation looks like when the mandate is not worth the candle. In governance theory, indifference is the sharpest veto: it offers no negotiating surface, no faction to persuade, no pressure point. The proposal died not from confrontation but from apathy.
There is also the second order signal โ the fallback plan itself. The production of a documented hard-fork reversal path is a confession. It means the developers could estimate the probabilistic failure of their own mechanism. So why proceed? One compelling reading: BIP-110's mandatory signaling was never intended to become a permanent activation mechanism; it was engineered as a credible commitment that node-side enforcement remained on the table. A demonstration, in production, that the client ecosystem could unilaterally impose a consensus rule โ and, more importantly, could retreat gracefully when rejected. The lesson was not lost. When SegWit activation stalled in 2017, the UASF movement revived exactly this philosophy: economic nodes could shift the network even against miner reluctance. BIP-110's ghost was walking in the UASF flags.
The eventual victory of BIP-9 โ with its 95% miner threshold, deployed for SegWit and Taproot โ was not a triumph of democratic procedure. It was the institutionalization of BIP-110's aborted trial. Bitcoin's governance system learned that miners must be negotiated with, not commanded, and then encoded that lesson into the rules of every future upgrade. The system corrected itself. That correction was not moral. It was functional.
Now the counter-intuitive angle. The 3% figure is always reported as a failure of the proposal. I argue it is a success of the system. Bitcoin processed an attempted consensus rule change through its distributed governance machine and returned a verdict: not now, not this. No permanent chain split. No lasting civil conflict. The rejection itself signified resilience โ the network absorbed a unilateral governance experiment and metabolized it without structural damage. That is not a bug report. It is the system demonstrating a defense against concentration of authority.
But the more uncomfortable transferable lesson is for today's L1/L2 ecosystem. The BIP-110 pattern โ protocol teams assuming technical authority maps to governance authority โ is being replayed everywhere. Consider rollups with forced upgrade keys, chains where the foundation controls the genesis validator set, or DA layers whose mandatory data availability rules are enforced by the same entity that produces the blocks. On those networks, a mandatory signal would simply pass. There would be no 3% moment of resistance because the separation of powers โ the structural condition that made BIP-110's failure meaningful โ does not exist. The insight is brutal: if your mandatory upgrade succeeds without friction, that might not prove legitimacy. It might prove the absence of a countervailing power. When I mapped the regulatory arbitrage of the 2024 Bitcoin ETF approvals, I saw the same dynamic: institutional frameworks neutralized decentralization not by force but by absorption. The mandate passed because no one with real capital had an incentive to refuse.
Decoding the silence between the blocks, the BIP-110 experiment leaves a governance test for any chain: watch how validators and operators respond when a mandate arrives unaccompanied by incentive. Not what they announce. What they produce. The three percent signal was Bitcoin telling its developers a truth no governance forum could deliver: consensus cannot be commanded; it can only be cultivated. When the next forced upgrade window opens โ on any network โ the first place to look is not the roadmap. It is the version bits.