Recovering a supermajority
Ethereum finality depends on a sufficient share of active stake participating correctly in checkpoint voting. If too much stake becomes inactive, blocks may continue to appear while finality stalls. The inactivity leak gradually reduces the balances of validators that are not making the required contributions, allowing the participating set eventually to regain the necessary weight. It is an availability-recovery mechanism, not a rule that every missed attestation constitutes malicious behavior.
2 sources for this section
An illustrative outage
Suppose a substantial group of validators loses connectivity while another group continues operating. Without a recovery mechanism, the online group could remain below the finality threshold indefinitely. The leak changes relative stake weights over time. Recovery speed and losses depend on the state and duration of the event, so a fixed penalty quoted from one incident should not be presented as the cost of every outage. Network partitions also complicate what each group observes.
2 sources for this section
Distinguishing penalties from slashing
Ordinary missed duties, inactivity-leak penalties, and slashing address different conditions. Slashing concerns provable violations such as conflicting signed messages; the inactivity leak is triggered by prolonged failure to finalize. Operators should examine client diversity, time synchronization, connectivity, and safe recovery procedures rather than assuming every penalty means keys were compromised. For users, a stalled-finality event is a reason to distinguish a newly proposed block from a finalized checkpoint when evaluating settlement assurances.
2 sources for this section
The source notesEvidence & further reading3 sources
- Proof-of-stake rewards and penalties Ethereum.org · Primary source · accessed 2026-09-21
- Ethereum proof of stake Ethereum.org · Primary source · accessed 2026-09-21
- Gasper: fork choice and finality Ethereum.org · Primary source · accessed 2026-09-21