Ethereum L1 as a shared sequencer with preconfirmations === *Note*: This content was discussed on [the sequencing & preconf call #1](https://www.youtube.com/watch?v=2IK136vz-PM), [the RollCall breakout on shared sequencing](https://www.youtube.com/watch?v=eycLQCaqDsk), and at [mev.market](https://www.youtube.com/watch?v=YrAlAHbgRCk). **TLDR**: We show how rollups can use Ethereum L1 for shared sequencing and preconfirmations. The simple construction significantly improves two previous designs ([here](https://ethresear.ch/t/based-rollups-superpowers-from-l1-sequencing/15016) and [here](https://ethresear.ch/t/based-preconfirmations/17353)) and does not require a hard fork. **construction** A subset of L1 proposers called "sequencers" opt into providing shared sequencing and preconfirmation services for rollups (and L2s more generally) by pledging collateral and exposing themselves to two slashing conditions described below. L1 proposers that are not sequencers are called "includers". At any given moment, the first sequencer in the beacon chain lookahead (if any) has monopoly rights to sequence and preconfirm transactions from participating rollups. Includers in the lookahead prior to the first sequencer can settle transactions preconfirmed by the sequencer, as well as include onchain transactions that the sequencer must settle by their slot. The sequencer signs preconfirmation promises to users in exchange for preconfirmation tips paid when the promise is honoured onchain. A preconfirmation promise that was not honoured onchain by the sequencer's slot is a slashable preconfirmation fault. There are two types of faults: * *liveness faults*, where the sequencer's slot was missed * *safety faults*, where the sequencer's slot was not missed **economics** Safety faults should be maximally punished as they are entirely preventable by the sequencer. The collateral amount should reflect the maximum amount a sequencer stands to make by not honouring any of its preconfirmations—1,000 ETH is probably safe given that it's larger than the largest MEV-boost bid value to date. The slashing amount for a liveness fault can be agreed with the user based on the risk of an accidental missed slot as well as the preconfirmation tip amount. Economically speaking a preconfirmation is an execution future sold by the sequencer. That future should be priced appropriately via the preconfirmation tip (whether positive or negative) to reflect the expected MEV impact (whether negative or positive) for the sequencer when the block is settled at the end of their slot. For builders to build blocks respecting preconfirmations, MEV-boost should be modified so that the sequencer can communicate constraints (e.g. top-of-block or top-of-blob constraints) to relays and builders. **discussion** The above construction shows how Ethereum L1 can be used as a shared and decentralised sequencer for rollups, with preconfirmations. The Ethereum shared sequencer mechanism is particularly exciting: * **economic security**: The mechanism enjoys the same economic security as Ethereum L1. The security assumption rollups make for settlement safety is reused for censorship resistance. * **simplicity**: Rollups do not need to use an external consensus for sequencing, and do not need an escape hatch. L1 sequencing infrastructure—including searchers, builders, relays—is reused. * **credible neutrality**: The mechanism is maximally credibly neutral for rollups and their competitors, both economically and memetically. No new token or brand is introduced. * **network effects**: The mechanism enjoys synchronous composability with the L1 EVM on sequencer slots, tapping into $0.5T of L1 TVL. * **L1 preconfirmations**: With [execution tickets](https://ethresear.ch/t/execution-tickets/17944) every rational L1 execution proposer will opt to become a sequencer so that the even the L1 EVM can enjoy preconfirmations.