BIP-110's Split: 59 Blocks, Zero Signals, and the Forensic Truth of a Two-Block Orphan

At 06:34 UTC on Aug. 9, Bitcoin's dominant chain sat at block 961,690. The BIP-110 enforcing branch stood at 961,633. Fifty-seven blocks behind. Eight hours and forty-five minutes old. The code whispered secrets the audit missed. This is not a price movement. This is a consensus split. A mandatory signaling window intended to produce a binary choice has instead produced two chains: one with all the hashrate, the other with a two-block orphan and a belief system. I have dissected enough protocol failures to know what this means. We are about to examine a fork that was mathematically inevitable. BIP-110 is a temporary soft fork that proposes to restrict several methods of placing arbitrary data inside Bitcoin transactions. Proponents say it keeps the chain focused on money. Critics say filtering valid transactions weakens neutrality. I do not care about the politics. I care about the state machine. The deployment mechanics are precise. If a node enforces BIP-110, it will only accept blocks carrying version bit 4 during the signaling window from height 961,632 through 963,647. The threshold is 55%, or 1,109 blocks out of 2,016. A compliant chain that reaches 963,648 enters LOCKED_IN. The restrictions become ACTIVE only at 965,664, one full retarget period later. That is the design. It is not a referendum in the Twitter sense; it is a hard threshold of mining power. From my experience stress-testing consensus mechanisms, I have seen a patient, silent divergence before—but rarely this pure. The divergence began at 961,632. Enforcing nodes looked at the next block header. No bit 4. They rejected it. The dominant chain mined it anyway. At that instant, enforcing nodes left the canonical chain. They did not do so by choice. They did so because the rule set they chose to run dictated it. That is the nature of a soft fork with enforcement: it is no longer soft for the enforcer. Now, the forensic data. The dominant chain produced 59 consecutive blocks after the window opened. I sampled the first 59 block headers. Zero contained bit 4. Not one. A 59-block streak of silence. The BIP-110 enforcing branch, meanwhile, produced exactly two blocks—961,632 and 961,633—both from OCEAN, both carrying the required signal. Then nothing. The enforcing branch stopped. It did not get 51% of hashrate. It got a proof-of-concept and left it at that. Let me unpack that silence with a little statistics, because this is where the narrative collapses. If the true signaling rate were a modest 5%, the probability of seeing zero signals in 59 independent blocks is (0.95)^59, roughly 4.8%. That is a rare event. If the true rate were 10%, the probability is (0.90)^59, about 0.2%. In other words, a 59-block zero count is a strong statistical rejection of any meaningful signaling effort. It is not a random pause. It is an effective boycott. But do not call it silent. The data screamed. The absence of bit 4 in every header is not a void; it is a measured, monotonic sequence of refusals that any node can verify. The proof-of-work chain is itself a voting machine, and the first 59 votes were unanimous. I do not trust sentiment. I verify block headers. The headers say what the mining industry will not say in public. The enforcing branch's fate is instructive. Two blocks, both from OCEAN, both stale within minutes. For enforcing nodes, they are now 57 blocks behind the dominant branch. That means their chain is eight hours and 45 minutes old at the snapshot. A chain that exists but does not grow is not a chain; it is a bookmark. It is vulnerable to reorganization, to stuck transactions, to an eventual decision to abandon the branch. The cost of running an enforcing node is not electricity; it is isolation. Consider the economic incentives from the miner's side. Blocks attributed to Foundry, F2Pool, AntPool, ViaBTC and MARA appeared on the dominant branch during the entire sample. No observable major-pool switch followed the window opening. Coinbase tags are not formal policy, but they are the only public signal available, and the signal is uniform. The major pools did not switch because switching would have meant throwing capital into a sea of orphaned blocks. OCEAN's ideological gesture produced two orphaned blocks. Collateral is a lie; math is the only truth. And the math of unsupported signaling is liquidation. Now, the exchange status feeds. Coinbase and Kraken reported their Bitcoin-related systems operating normally. That is a bounded check. It tells you they did not see a liquidity crisis. It does not tell you whether the enforcing branch is stable, whether any merchant or wallet has processed a transaction on that branch, or whether the next 1,957 blocks will contain a single signal. The sample excludes the entire decentralized periphery—self-hosted nodes, lightning routers, custody clients, and every wallet that has not posted a public status page. I do not trust reputation. I verify hashes. Until every node directory publishes its fork status, "normal operations" is a placeholder, not a proof. One layer deeper. A soft fork of this kind does not need to lock in to create damage. The damage is already here. The moment enforcement nodes rejected block 961,632 without bit 4, they created a partition in the Bitcoin network. Transactions mined on the dominant chain are invisible to them. Their mempools diverge. Their UTXO sets diverge at 961,632. If they ever reconnect, the reorg will erase transactions they thought were settled. This is exactly why I spent weeks in 2022 dissecting UST's depeg: the failure mode was not a single exploit, it was a slow, accelerating divergence of incentives. The BIP-110 split is smaller, but it is the same disease. That brings me to the contrarian angle, because there is one. For all the noise, BIP-110 has not broken Bitcoin. On the contrary, it has demonstrated Bitcoin's immune response. The dominant chain absorbed the challenge and moved forward. Proof-of-work is the ultimate judge, and in this test, the judge chose the chain without arbitrary-data restrictions. That is not an endorsement of BIP-110's critics; it is an economic decision, not a moral one. Miners followed the most profitable chain. OCEAN chose ideology and paid the price. The network's resilience is real. The market's verdict is separate. But here is the uncomfortable conclusion that both sides will hate: this failed rollout is actually a successful stress test. It showed that a small, determined group can enforce a rule set that diverges from the majority, and that the majority can simply produce a longer chain that ignores them. No double-spend. No exchange collapse. No lost funds beyond what the enforcers willingly sacrificed. The protocol's security model did exactly what it was designed to do: it absorbed a minority fork with negligible cost. The code whispered secrets the audit missed; we should be glad the mechanism survived the secret. Does that mean BIP-110 is safe? Absolutely not. A successful stress test does not validate the fork; it validates the substrate. The proposal remains a political product. It restricts transaction data types based on a view of what Bitcoin should be. That is a governance choice, not a cryptographic requirement. And governance by miner signaling is a fiction. The 0.42% signaling rate since May 1, the 0.0% signaling rate in the first live window, and the two-block orphan are all evidence of the same reality: miners vote with hashrate, not with headers. What remains is not a technical fight. It is a governance question. BIP-110 will almost certainly fail to reach LOCKED_IN. The 1,957 remaining blocks are now just a countdown, a tombstone with a timer. But the next contentious proposal will come. And the next one. The signaling mechanics have not changed. The gap between bit 4 in a header and hashrate in a chain is still absolute. Between the lines of bytecode lies the trap; today, the trap is empty. Do not mistake an empty trap for an absent adversary. The proof is incomplete. The doubt is not obsolete.