BIP-110 nodes rejected Antpool’s block at height 961,632 and forked onto a rival chain mined by Roughnecks. The minority chain produced one more block and then stalled. The main network never
Bitcoin split into two chains on Saturday, August 8, at block height 961,632, when nodes running the BIP-110 soft fork rejected a block that lacked a required signal and started building on a competing version instead. The breakaway did not last. Backed by roughly 2.53% of the network’s mining power, the BIP-110 chain squeezed out one more block and then effectively froze, while the rest of the network kept its normal ten-minute rhythm and pulled dozens of blocks ahead. A change pitched as a year-long cleanup of “spam” became a live demonstration of how little a minority fork can do without hashrate behind it.
BIP-110, formally titled Reduced Data Temporary Softfork and authored under the pseudonym Dathon Ohm, had spent months as the flashpoint of an argument that has divided Bitcoin developers since Ordinals arrived. Its backers, led by Bitcoin Knots maintainer Luke Dashjr, wanted a one-year consensus rule to shrink the space available for arbitrary, non-financial data in transactions. The target was never hidden: Ordinals, BRC-20 tokens and Runes, all of which pay to store images, text and alternative tokens directly on the blockchain. Supporters call that data spam. Bitcoin Core developers had moved the other way only months earlier, raising the default OP_RETURN data limit from 83 bytes to 100,000 bytes, and their position was that block space goes to whoever pays the fee. Restricting what that fee can buy, in their view, breaks the network’s neutral, pay-to-play design.
The activation route was a User-Activated Soft Fork, the mechanism the 2017 SegWit standoff made famous. Because the large pools refused to coordinate, BIP-110 supporters ran modified nodes, mostly Bitcoin Knots, that would enforce the rule themselves. At block 961,632 those nodes demanded that miners signal support through version bit 4, and any block without the signal counted as invalid. Antpool mined 961,632 without it. The main network accepted the block and moved on, while BIP-110 nodes threw it out and built on a compliant version mined by a small group called Roughnecks through the Ocean pool. From that point the two chains no longer shared a history.
Timing turned a weak fork into a stranded one. Bitcoin’s difficulty adjusts every 2,016 blocks, and the next adjustment fell exactly at 961,632, lifting the target 0.99% to 127.48 trillion. The main chain absorbed that without noticing, because it still carried roughly 97% of the world’s hashrate. The minority chain did not. With about 2.53% of the mining power chasing the same difficulty, its blocks slowed to an average of roughly 6.9 hours each against Bitcoin’s usual ten minutes. Roughnecks ground out 961,633, then the chain went quiet.
MetricMain Bitcoin networkBIP-110 minority chainHashrate~97.47%~2.53%Block status at splitAdvanced past 961,680Stalled at 961,633Average block time~10 minutes~6.9 hoursStatusFully operationalFrozenMichael Saylor had sketched the outcome in advance, noting the roughly 2.6% signaling rate and predicting the branch would stall or fork into irrelevance, closing with the line since quoted back at supporters: “Bitcoin is working as designed.” Adam Back was blunter once the chain froze. “They forked off and found out,” the Blockstream chief wrote. Dashjr refused the label, conceding blocks were slow but “tolerable,” and told Bitcoin Knots users to upgrade, warning that anyone who didn’t risked what he called counterfeit coins from double-spend attacks. Dathon Ohm framed the collapse as an assault rather than a defeat.
Both chains still recognize the same transaction history up to the split, which is where the practical danger sits. A transaction signed to move coins can, in certain conditions, be valid on both sides, so anyone touching funds on the stalled chain risks a replay-style problem that drags a payment onto a network they did not intend to use. The proposal itself flags a separate hazard, spelling out narrow conditions under which pre-signed Taproot transactions could see funds frozen during the one-year deployment. Australian exchanges Bitaroo and Hardblock had published contingency plans before the block arrived. Some supporters are now discussing a change to the minority chain’s proof-of-work algorithm, the one move that could keep it producing blocks, though it would also turn the result into a separate asset with its own ticker rather than a Bitcoin fork.
BTC reclaims the 61.8% Fib. Chart: Alexander Stefanov, TradingView. Price treated the whole episode as a non-event, trading near $64,928 with a 3.06% weekly gain and a market cap around $1.3 trillion. On the 4-hour chart, a Fibonacci retracement runs from the July 21 high at $66,972 to the early-August low at $62,272, and price is now pressing the 0.618 level at $65,171. That is the line bulls need to reclaim and hold to argue the July drop was only a correction, and the last several candles have stalled just under it rather than pushing through.
Two things support the recovery: the 20-period moving average has crossed back above the 50-period one, and the 0.382 level at $64,088 lines up with that 50-period average as a floor. The catch is momentum. The RSI, a gauge of whether buying pressure is building or fading, sits near 57 and has drifted sideways for days. A 4-hour close above $65,170 that holds opens $65,966 and a retest of the $66,972 high, while a loss of $64,088 puts $63,382 back in play. Until one of those closes prints, the frozen fork leaves the harder question untouched: the datacarrier limit Core raised to 100,000 bytes is exactly where it was before block 961,632, and whether the Knots camp presses the proof-of-work button in the coming days is the thread worth tracking.
The post BIP-110 Split Tested Bitcoin’s Rules and Lost appeared first on ETHNews.
