Long-Range Attack
A theoretical proof-of-stake attack in which someone acquires old private keys and uses them to construct an alternative transaction history starting from far back in the chain.
A long-range attack is a theoretical attack against proof-of-stake blockchains in which an attacker acquires private keys that controlled a meaningful share of stake at some point far in the past — potentially keys that are no longer actively used or valuable, and so might be obtained cheaply — and uses them to construct an alternative, internally valid-looking transaction history branching from that distant point in the chain's past. If a new node joining the network couldn't distinguish this fabricated history from the real one, it could theoretically be tricked into following the wrong chain.
This differs from an attack on a live, up-to-date node, since a long-range attack specifically targets the "weak subjectivity" problem (see Kaavon's "Weak Subjectivity" glossary definition) facing a new or long-offline node that has no independent, trusted reference point for which chain is genuinely correct. Proof-of-stake protocols defend against long-range attacks primarily by having new or returning nodes rely on a recent, trusted checkpoint — obtained from an already-trusted source — rather than attempting to validate the chain's entire history purely from genesis using nothing but each block's own internal signatures.
Examples
- A long-range attack specifically threatens a brand-new node syncing for the first time, which has no independent way to know which of two competing, internally valid-looking chain histories is genuinely correct.
- Relying on a recent, trusted checkpoint rather than validating a chain's entire history purely from genesis is the main practical defense proof-of-stake protocols use against long-range attacks.