Two mining pools now decide whether four days of transaction history survive, after the project team's request for a closer recovery point was refused.
Ravencoin's proof-of-work chain has forked around a header field the network never actually validated.
The first invalid block slipped past vulnerable nodes at height 4,487,776 on 7 August at 15:44 UTC. Ravencoin uses KAWPOW, the memory-hard variant of ProgPoW that replaced X16R in 2020, and every block header carries a field called nHeight that records where the block sits in the chain. That value was never compared against the block's actual position. An attacker who understood the gap could set nHeight to whatever value routed the header through a validation path that skipped KAWPOW entirely. The proof-of-work never had to be done. Forged blocks appeared as if they had earned their place.
Once the mechanism was proven on mainnet, others copied it. The invalid branch grew.
Two mining pools now decide how much of the last four days of history survives. 2Miners and RavenMiner, which between them command a majority of the network's hashrate, are mining a separate chain that treats block 4,487,775 as a checkpoint and rejects any header with a mismatched nHeight from that point forward. 2Miners shipped the patch as version 4.6.1.1-hf1. If the recovery chain overtakes the exploited one, the reorganisation will span roughly three days.
The Ravencoin project account, which speaks for the developer team but not the pools, posted a notice on 10 August spelling out what that means for users. Any transaction confirmed after block 4,487,775 should be treated as at risk. Some may return to the mempool and be mined again. Others will not. Deposits, withdrawals and payments made in that window may simply disappear from whichever ledger the pools decide is authoritative.
The notice contains a line that matters more than the technical detail. The team asked the two pools to move the recovery checkpoint closer to the present so fewer users would be affected. The pools said no. That refusal is the story of Ravencoin's governance in one sentence: the code was written by developers who no longer control what happens when it breaks. The pools ship the patch and pick the checkpoint. Whichever four days of history the network keeps is now their call.
This is the third consensus failure in Ravencoin's history. In 2020 an attacker minted around 1.5% of the total supply through a bug in asset creation logic. In June this year a separate bug in forked asset code was patched by an outside developer through a BIP9 soft fork. The current incident is the most severe of the three because it touches the base layer directly and requires a reorg to fix, not a soft fork that activates on future blocks.
RVN dropped roughly 19% on the news, trading around $0.00288 and pushing the market cap below $48 million. Upbit and Bitget suspended deposits and withdrawals. Other venues had not immediately followed by 11 August, though the project's notice recommended they should. An exchange that keeps RVN flowing through a possible three-day reorg is exposing itself to reversed deposits and orphaned withdrawals, with no automatic recovery for either.
The technical fix is small. Ravencoin's KAWPOW validation should have checked the header's declared height against the block's real position at the point of proof-of-work verification and refused to hash any header where the two disagreed. It is a one-line consistency check that never made it into the reference client. What that says about the depth of review the codebase has received since the project's original founders stepped back is a separate question, though not a difficult one to answer.
Ravencoin has always sold itself as a fair-launch cypherpunk asset chain with no premine and no foundation governing issuance. That model works while nothing breaks. When something does break, the absence of a coordinating body means whoever controls the majority of the hashrate becomes the coordinating body by default. 2Miners and RavenMiner are cooperating this time. Nothing in the network's structure guarantees they always will.
Every proof-of-work chain smaller than bitcoin lives with this. A network under $50 million in market cap and a handful of pools between it and a rewrite of its recent history is not really a chain in the sense bitcoin is one; it is a working assumption that everyone with power will keep behaving as they have. The nearest parallels are Ethereum Classic's repeated 51% attacks and Horizen's earlier defence built on checkpoints imposed after the fact rather than the security budget itself.
Ravencoin's users find out in the next few days which chain wins.
