The BIP-110 signaling monitor — the same chart the proposal’s author cites as evidence of momentum — records 47 blocks carrying versionbit 4 out of 1,806. That is 2.60 percent. Four hundred blocks earlier, the rate was 2.70 percent. In between, eight new flags appeared across 392 blocks. The trend line, over the only meaningful sample available, is not rising. It is slipping.
Four public figures read this data set and produced incompatible conclusions. Dathon Ohm, the pseudonymous author of BIP-110, thanked the miners he says are signaling readiness and instructed every Bitcoin user to abandon Bitcoin Core. Luke Dashjr, the Bitcoin Knots maintainer, says the outcome is already settled. Michael Saylor reads the same numbers and tells BIP-110’s backers to stand down. Adam Back calls the proposal a stupid idea that completely fails at technical consensus.
They cannot all be reading the same chart. At least one of them is not reading it at all.
This is the state of Bitcoin at the approach to block 961,632, one of the most peculiar soft fork deadlines in recent memory. A proposal whose own data proves it cannot activate. A contributor who says the alternative — running Core — becomes insecure when mandatory signaling begins. A maintainer who says the result is settled while his own monitor shows 97.4 percent of the network declining to participate. Truth is found in the block header, not the press release. The split everyone is debating is not the interesting one. The interesting split is between what the signaling data shows and what each camp claims it means.
Context: The Fork That Cannot Lock In
BIP-110 is a temporary rule change that caps the amount of data a Bitcoin transaction can carry. In practice, that means tighter limits on the data-bearing fields of a transaction — a policy position, not a protocol defect. It ships in Bitcoin Knots, the smaller alternative client maintained by Dashjr. Its activation mechanism is an unusual hybrid: under its rules, when mandatory signaling begins at block 961,632, miners must set versionbit 4 in every block header. Blocks without the flag become invalid to BIP-110 nodes.
The change is not pointed at a known exploit and not at a researched vulnerability. The project’s own website frames the proposal as a curb on arbitrary data — the long-running blockspace spam debate. Ohm, however, described it as a fix for critical vulnerabilities. That claim does not appear on the proposal’s website. The gap between those two framings is the first thing a careful reader should audit.
Activation requires 1,109 signaling blocks within a single 2,016-block period. At 55 percent of the window, that threshold is deliberately low — a departure from BIP-9’s historic 95 percent requirement, designed to avoid the hostage scenarios that plagued earlier activation fights. The threshold was lowered to make activation easy. It has not been approached once in eight months.
Since December, no completed two-week period has finished above 1.29 percent signaling. The current period is running at roughly double that baseline, which is the source of the modest optimism in Ohm’s acknowledgments. But the absolute math is unforgiving. With 217 blocks remaining in the activation window, even a perfect finish — every remaining block carrying versionbit 4 — produces a period total of 263. The highest reachable total is 263. That is 23.7 percent of the 1,109-block threshold.
The fork cannot lock in. The question is no longer whether BIP-110 activates. It is what happens at block 961,632 to the people who believe it already has.
The constituency watching this unfold is not limited to miners and node operators. Exchanges, custody providers, and institutional market makers are observing from the sidelines, preparing the same internal memos they prepared for every contested activation since SegWit: monitor block 961,632, identify the minority chain, and refuse to credit deposits on either chain until the hashrate picture stabilizes. In a sideways market, where attention is scarce, this is the kind of technical event that separates teams that know their stack from teams that copy announcements.
Core: The Last 217 Blocks
Let me be precise about the arithmetic, because both camps have an incentive to be sloppy with it.
The current window opened 1,806 blocks ago. To lock in early — the only viable path to activation, since the full window threshold was never plausible — the fork needs 1,109 signaling blocks by period close. It has 47. There are 217 blocks left in which to find another 1,062 signals. Every remaining block would need to carry the flag, and that would still land the total at 263. No miner defection, no eleventh-hour capitulation, no coordinated final stand can close a gap of 846 signatures. The early lock-in is mathematically unreachable.
The more interesting number is not 263. It is the implied composition of the signaling hashrate. At 2.6 percent of the network over 1,806 blocks, 47 blocks represents roughly two to three mining entities working consistently. This is not a user movement, and it is not a mining majority finding its conscience. It is a small cluster of operators — likely a single pool and its affiliates — running Knots and setting the bit on schedule. The rest of the network is declining by omission.
That composition matters because of what it reveals about the ceiling of any signaling campaign. Based on my audit experience with contested activation mechanisms, the final two weeks provide the only honest signal. Committed advocates surge toward the deadline: the monitor ticks upward, the language inflates, the retweets accelerate. The current period’s 2.60 percent — roughly double the previous 1.29 percent ceiling — is exactly such a surge. It is real. It is also the exhausted final form of the campaign. The committed cluster is already signaling at full capacity, and full capacity is 2.6 percent.
Now run the orphan mathematics nobody in the public argument has printed. A miner signaling at 2.6 percent of network hashrate finds a block, on average, every 384 minutes. Once mandatory signaling begins, the non-signaling majority continues extending the chain that the rest of the network treats as canonical. The minority chain’s blocks are orphaned whenever the majority chain extends past them — a constant condition, because the majority mines 97.4 percent of all blocks. A miner entering this arrangement at the deadline is choosing a lifetime expected orphan rate near 97 percent. Every orphaned block is a burned coinbase plus burned fees. The expected value of signaling is a near-total loss of mining revenue.
The pro-fork camp can argue that BIP-110 nodes regard the majority chain as invalid, making the minority chain the “real” one. That is the argument of a protocol that has confused its own rulebook with the physics of proof of work. The majority chain does not require anyone’s recognition to exist. It requires hashrate. It has 97.4 percent.
This is the mechanism behind Ohm’s warning that miners obtaining templates from Core may produce invalid blocks on “an incoherent chain that keeps being wiped out, along with any earnings.” He is describing the minority chain precisely. He is aiming at the wrong miners. The chain that gets wiped out is the one carrying versionbit 4. Its blocks are rare, orphaned constantly, and their rewards are absorbed by the longer chain’s reorgs. There is a reason the threshold is not being reached. It is not a failure of messaging. It is a rational, network-wide response to a proposal whose expected value for miners is catastrophically negative.
Core: The “Insecurity” Claim Is Backwards
Now the claim that Bitcoin Core becomes insecure when mandatory signaling begins. Nothing on the BIP-110 website says this. Ohm said it on social channels, and it deserves the same scrutiny as a code diff.
Core becomes, at block 961,632, non-compliant with a consensus rule enforced by BIP-110 nodes. Non-compliance is not insecurity. Core does not lose funds. It does not expose keys. It does not accept invalid signatures. It simply does not set versionbit 4. If miners using Core templates produce blocks without the flag, BIP-110 nodes reject them. The risk to those miners is real — it is the orphan risk described above — but it is manufactured by BIP-110 nodes choosing to enforce a minority rule. It is not a vulnerability in Core.
The more revealing discrepancy is the “critical vulnerabilities” framing itself. A flaw severe enough to justify an emergency soft fork would not be handled this way. It would be coordinated across implementations. It would move release schedules. It would be documented prominently on the proposal’s own site. None of that has happened. Dashjr told reporters that shipping a Knots release without BIP-110 would not change the schedule. Security fixes change schedules. They are the one category of code that cannot wait on versionbit windows.
Code does not lie, only the architecture of intent. When a proposal begins as a policy argument — a curb on arbitrary data, a position in the blockspace debate — and ends its campaign as an emergency security intervention, the code is the same; only the framing moved. The “critical vulnerabilities” language arrived on the eve of failure, after eight months of signaling rates that never crossed 1.29 percent. That is not a technical finding. It is a dispatch from a project that ran out of arguments but not out of announcements.
This is also why the conduct of the final days matters. Miners are being pressured to signal a rule change using a security justification its own website does not make. Users running Knots at this point are not running the version that fixes a critical vulnerability — they are running the version that loses the orphan race. Neither instruction survives contact with the activation math. The same monitor that shows 47 signals is the monitor Dashjr cites when he says there is no material opposition. Those two statements cannot both describe the same chart honestly.
Core: What Conviction Costs
Adam Back’s dismissal of BIP-110 points at the real distinction. SegWit had technical consensus and ecosystem consensus — modulo a later-stage attempt to abuse its own activation process to veto what had been agreed. BIP-110 has neither technical nor ecosystem consensus, and it is failing without even the drama of an activation fight. The absence of drama is itself the data. The last contested fork in Bitcoin’s history engaged the entire ecosystem. This one is being contested by a monitoring site and a social media account.
The cost of conviction, meanwhile, is being paid in uncounted form. The miners who signal through 961,632 are not expressing a preference. They are burning capital. Each block of the minority chain that falls to a reorg is a real loss no market dynamic will refund. The pools that continue signaling past the first reorg wave will be making a political donation to a proposal that cannot mathematically activate. At 2.6 percent hashrate, the minority chain cannot sustain meaningful economic throughput; it cannot settle exchanges, and it cannot time-stamp the volume that would give its coinbase value. The rest of the network’s silence is not ignorance, and it is not conspiracy. It is the ordinary arithmetic of a cost-benefit analysis BIP-110’s advocates have refused to print.
Technical Appendix: The Activation Math
- Deadline set by Ohm: mandatory signaling begins 290 blocks after the August 6 announcement (~48 hours); at publication, the window has 217 blocks left.
- Current window: 1,806 of 2,016 blocks elapsed
- Signals logged: 47 (2.60%)
- Prior completed windows since December: all below 1.29%
- Required for early lock-in: 1,109
- Maximum achievable total: 263 (23.7% of threshold)
- Implied signaling entities: 2–3 pools
- Average time to a minority-chain block at 2.6% hashrate: 384 minutes
- Expected orphan rate for a continuously signaling miner after mandatory signaling: ~97%
Contrarian: Everyone Is Reading the Wrong Chart
Dashjr says there is no material opposition to BIP-110 and the outcome is settled. Both statements deserve scrutiny. The outcome is settled: the fork cannot reach its threshold. But it is settled precisely because there is material opposition. 97.4 percent of the network declining to signal for eight months is the definition of material opposition. Confusing a favorable outcome with an agreeable one is a category error. It is the same error as thanking the many miners who are signaling while reading a 2.60 percent monitor.
Saylor’s advice — that Bitcoin continues normally while BIP-110 stalls or forks into irrelevance — is mostly right and partially blind. Bitcoin does continue. The minority chain is a ghost. But the ghost is not harmless. Users who followed the instruction to run Knots, and miners who continue signaling after 961,632, will experience a chain that confirms slowly and reorgs violently. Transactions paid on the minority chain are reversed when the majority chain extends. For a holder, the correct posture is not “continue normally.” It is: do not transact during the activation window, verify which chain your confirmations actually live on, and treat any flow that references versionbit 4 with suspicion.
And Ohm’s own warning deserves to be quoted back to him in full: miners using Core templates may produce invalid blocks on an incoherent chain that keeps being wiped out. It is the most accurate sentence written in this entire episode. The only correction is the attribution. The incoherent chain is the one bearing versionbit 4.
The derivatives market will price this before the mempools do. Perpetual funding and basis will diverge for any instrument that references the versionbit-4 chain, which is exactly when the minority chain’s economics turn from bad to worse. Anyone who cannot measure that spread should not be choosing sides; they should be holding the one asset that is not ambiguous — the chain with 97.4 percent of the work.
The deeper error on both sides is framing this as a contest between Bitcoin Core and Bitcoin Knots. It is not. It is a contest between a proposal with no hashrate and a chain with 97.4 percent of it. Bitcoin neutrality is not a policy position; it is a probability distribution. Hedging is not fear; it is mathematical discipline. The correct hedge requires no opinion on the merits of BIP-110’s data caps. It requires only the acceptance that a 2.6 percent chain is not a bet. It is a donation.
Takeaway: What the Next Fork Will Look Like
BIP-110’s activation window closes with a maximum reachable total of 263 — 23.7 percent of its own threshold. Block 961,632 will arrive, and the versionbit-4 chain will either never begin or collapse within its first reorg cycle. The casualties will be the miners who believed their own monitor, and the users who moved to Knots on the strength of a security claim no website was willing to make. For holders, the correct posture is simple: do not transact across the activation boundary, hold whichever chain the market prices, and do not build a position on a 2.6 percent chain.
The playbook, however, is now public record. A policy fork dressed as an emergency. A low activation threshold its supporters still cannot reach. A monitor showing the truth while the announcements describe something else. The next proposal will not wait for 95 percent hashrate. It will come with its own website, its own versionbit, and its own rehearsed urgency. History is a dataset we have already optimized. The only open question is whether the next one will be wearing the same architecture of intent — and whether this time, the people being asked to run the minority client will check the numbers before they check out.