How Application-Specific Ordering Redistributes MEV

Introduction

Blockchain transaction ordering is a complex, high-stakes process. Block producers don't simply list transactions in the order they appear; instead, they solve a problem similar to the knapsack problem, selecting the mix of transactions that maximizes fee revenue while staying within block size or gas limits. Once this subset is chosen, the order in which transactions are executed can influence market fairness, expose vulnerabilities like front-running, and even allow block producers to extract extra value.

The Current Landscape

The industry's intense focus on MEV retention has arguably overindexed on capturing immediate, extractable profits at the expense of broader systemic value. This narrow focus tends to overlook the transformative potential of generalised sequencing strategies. By shifting the lens from pure fee maximization to application-specific sequencing, protocols can be designed not only to mitigate exploitative practices like front-running, but also to return meaningful value directly to decentralized applications. This approach leverages a more nuanced understanding of market dynamics, one that aligns long-term system stability with the genuine needs of users and developers, rather than simply rewarding those best positioned to extract MEV.

Several projects are now innovating with new application-specific sequencing solutions. FastLane's Atlas Protocol offers customized Order Flow Auctions, Sorella Labs' Angstrom implements permissionless sequencing hooks for Uniswap V4, and Vertex Protocol employs off-chain sequencers to optimize performance while minimizing MEV risks. 

Novel Approaches to Transaction Sequencing

Given these challenges, new protocols are emerging that reimagine how transactions are sequenced. By applying principles from auction theory, cryptography, and game theory, these proposals aim to create more predictable and equitable transaction processing. Techniques such as batch auctions, randomized ordering, and cross-transaction matching are being explored to minimize arbitrage opportunities and align incentives more closely with overall system stability. We organize these proposals into four logical groupings:

Fair Pricing:

This group addresses the fundamental challenge of establishing a level playing field for price discovery while eliminating advantages for early movers, these mechanisms not only deter predatory strategies but also foster a market environment where every participant receives fair treatment regardless of their ability to bid aggressively.

Batch Auctions: Instead of processing swaps sequentially, a DEX sequencer could batch all swap orders within a time interval and clear them at a single uniform price. This turns continuous trading into discrete auction ticks, creating an instrument similar to a periodic call auction. All participants in the batch get the same fair price, eliminating the first-mover advantage. Mathematically, if buy and sell orders for an asset are aggregated, the sequencer can solve a clearing price P where aggregate demand equals supply. This P is then applied to all trades in that batch, reducing MEV from arbitrage within that interval. Batch auctions can improve price discovery and remove the incentive for high-frequency ordering races.

Deterministic Slotting vs. Auctions: Today’s public mempools essentially create fee auctions for priority – which inherently favor the richest or fastest. An alternative is deterministic slotting, where the block has predefined “slots” or order groups for certain types of transactions or users. For example, a block might be structured so that: first slot goes to a randomly selected normal transaction, second slot to the highest-fee transaction, then repeat. This way, even low-fee transactions have a chance to appear early , and high-fee ones still get in but not monopolizing the top. Such a mechanism can be described mathematically: if we label transactions $t_i$ with fees $f_i$, the current approach sorts by $f_i$. A mixed-slot approach might sort by $f_i$ within half of the slots, and in the other half use a different metric. This balances fairness and efficiency, high fee payers can’t completely crowd out others. Implementing this would require the sequencer to explicitly follow the slot pattern when building blocks.

Order Matching:

This grouping focuses on the integration of order matching directly within the sequencing process, thereby optimizing efficiency and tailoring the transaction order to specific application needs.

Cross-Transaction Matching: A sequencer with insight into multiple user intents could directly match complementary transactions. For instance, if Alice wants to swap ETH for DAI and Bob wants to swap DAI for ETH, a custom sequencer could detect this and match them P2P at a fair rate within the same block, bypassing an AMM pool. This effectively creates a decentralized order matching engine as a layer on top of the blockchain. The outcome is akin to a peer-to-peer trade or a small OTC deal embedded in the block sequencing. By internalizing such order matching, the sequencer creates an ad-hoc direct swap contract that wouldn’t exist under standard FIFO ordering. This improves efficiency and keeps value inside the user ecosystem rather than leaking to arbitrage bots.

Turn-Based Sequencing: A sequencer can enforce a round-robin ordering among a set of known players. For example, in a turn-based strategy game on-chain, rather than relying on each player to submit a move and hoping miners include them timely, the sequencer could explicitly order moves: Player1’s move, then Player2’s, cycling. If a player fails to submit a move within their allotted slot , the sequencer could either skip them or allow a default action. This prevents any player from doing two moves in a row by bribing the miner, and ensures fairness akin to traditional turn order. It basically requires the sequencer to have game state context. On an app-specific chain or rollup for the game, that is feasible. It’s a clear example where application-specific sequencing directly mirrors game mechanics to preserve intended fairness.

Priority for Repayments: To encourage healthy lending markets, a sequencer might give slight priority to certain beneficial actions, like loan repayments or collateral top-ups, over new risky borrowings. For example, if a user is in danger of liquidation but sends a repayment transaction around the same time as another user trying to seize collateral, the sequencer could choose to sequence the repayment first. This kind of rule-based prioritization can optimize for system stability. It is also possible to prioritise order cancellations. While this deviates from pure time-ordering, it demonstrates an application-specific choice: valuing the long-term protocol integrity by sequencing critical safety transactions earlier.

Anti-Spam, Security, and Integrity Commitments:

Security is paramount in transaction ordering. This group explores protocols designed to safeguard against manipulation, spam, and collusion. These proposals collectively aim to enhance the robustness and integrity of the sequencing process. By embedding security measures and limits directly into the ordering protocol, the system can better resist manipulative tactics and maintain trust among participants.

Randomized Transaction Ordering: Another approach is to introduce a degree of randomness in the ordering of transactions with similar timestamps or fees. For example, if multiple swaps are submitted within the same block interval, the sequencer could randomly shuffle their execution order rather than strictly sorting by gas price. Economically, this turns the adversary’s problem into a probabilistic one, even if they bid high fees to sandwich, there’s a chance their transactions won’t get the advantageous ordering. Over time, this random sequencing can deter bots because the expected value of attempting a sandwich attack is reduced. This idea connects to game theory: if bot operators face uncertain outcomes, the equilibrium might shift away from always attempting to front-run. However, purely random ordering must be used carefully to avoid harming honest users; typically it could be applied among transactions that arrive nearly simultaneously to avoid violating first-come-first-served too much. 

No “Peeking” Sequencers: Even without fancy crypto, L2 rollups like Optimism and Base have implemented a rule where their sequencer is supposed to order transactions by the fee and arrival, without “peeking” into transaction content or censoring​. If such rules are verifiable or enforced by consensus, users gain security. One could imagine a sequencer design where the sequencer must commit to an ordering for each batch before processing them, under threat of slashing if they deviate. For instance, the sequencer publishes a hash of the transaction sequence it will include next, binding it to a specific order, then executes it. Any attempt to insert a new TX out of order (say, after seeing a profitable one in mempool) would break the commitment and could be punished. This is a game-theoretic commitment: the sequencer’s optimal strategy is to follow the committed order or lose its security deposit. Such mechanisms align sequencer incentives with honest ordering, thus reducing front-running opportunities stemming from sequencer misconduct.

Rate Limiting per Address: A simple sequencing rule to hurt spam bots is: limit the number or frequency of transactions from the same address within a single block or short time frame. For instance, the sequencer could enforce that a given address can only have at most 1 transaction in the “first half” of a block, or insert at most X gas worth of TXs per block. This prevents a single bot from dominating a block even if it sends a flurry of TXs. Real bots can use multiple addresses, but then they incur overhead and possibly need to distribute funds across them. The network effect is that casual spamming sees diminishing returns. Critically, this must be balanced with liveness: honest users sometimes legitimately send multiple TXs. The rule might be relaxed or applied probabilistically. For example, after an address’s first transaction, each subsequent one could face an increasing “virtual delay” in sequencing unless it pays significantly higher fees. This creates a diminishing priority model: the more you send rapidly, the less priority each additional TX gets, discouraging bot spam. 

Throttling Low-Reputation Actors: Conversely, new or low-reputation addresses might be limited in how aggressively they can influence ordering. A sequencer might deliberately delay a transaction from a brand-new address by a small random offset or require a higher minimum fee until the address builds some history. This can thwart throwaway bot addresses that are created to spam – each new address starts at the “back of the line” unless it pays a premium. One could formalize this: e.g., reputation score 0 might have a default +5 second delay applied in the sequencing algorithm. Over time as the address completes valid transactions, this handicap is removed. The effect is that Sybil attacks become less effective, because none of the new addresses get immediate top priority. It’s important though that genuine new users aren’t alienated, the delays or fee uplifts should be small and maybe one-time. Perhaps a user making a couple of successful TXs automatically graduates to normal priority.

Lottery or Random Sequencing for Mints: Instead of first-come-first-served minting, the sequencer could accept mint transactions during a time window and then order them randomly in the block that finalizes the drop. This effectively turns the drop into a lottery: everyone who submitted within, say, a 1-minute window has an equal chance of being early in the sequence to get an NFT until supply runs out. Some NFT launches have tried off-chain lotteries or pre-registrations; doing it at the sequencer level makes it trustless. The randomness could be derived from a verifiable random function or the block hash, to ensure it’s not tampered with. This sequencing approach removes the incentive for bots to spam at the exact launch time, since 100 bot entries have the same chance as 1 human entry if duplicates are filtered or not advantaged by time. It preserves liveness by still processing all transactions, just shuffling their order for fairness.

Premium Services and Dynamic Incentive Structures:

As decentralized systems mature, there is increasing demand for flexible, application-specific execution services. This grouping explores models that introduce service tiers and dynamic incentives for ordering. By offering differentiated services and dynamic incentives, these models provide users with choices tailored to their needs, ranging from cost efficiency to execution certainty, thereby fostering a more responsive and market-driven sequencing ecosystem.

Value-Added Ordering Services: An app-specific sequencer could offer premium ordering services for complex user needs. For example, a DeFi platform might say: if you want to execute a complicated multi-step transaction with guaranteed ordering, essentially a private block space, you pay a higher fee or a service charge to the protocol. This is similar to Ethereum’s concept of private order flow but directly provided by the protocol for a fee. The user pays for the guarantee that their transactions will be executed in a certain relative order or at a certain block time without competition. This could be packaged as an enterprise service for institutional users of a DEX who might want to do large trades without being front-run, they could pay the DEX’s sequencer a fee to handle it in a special block or shard. This pay-for-ordering service is a new revenue stream, and it competes on quality of execution rather than just lowest fees. The economic model is akin to selling execution quality as a product. In traditional finance, dark pools and exchanges charge for order types and execution guarantees; here the decentralized app can do something similar through its sequencing control.

Quality of Service Tiers: The blockchain could offer tiers like Standard, Gold, Platinum users. Standard is the normal public mempool treatment. Gold subscribers might get their transactions routed through a special high-priority lane in the sequencer. For example, the sequencer could always include at most one Gold TX per block at the front. Platinum could get even more guarantees, like always first in block or the ability to pre-schedule a transaction at a future block height. This is a big departure from egalitarian blockchain philosophy, but it introduces a predictable cost for priority rather than the uncertainty of bidding gas each time. Subscribers pay for assured performance. The revenue from subscriptions would go to the protocol or validators. Economically, this converts what is typically an ad-hoc per-transaction auction into a steady stream. It could attract power users. The key is that non-subscribers are still served, but subscribers get a slight deterministic edge.

Deterministic Priority Auctions: Instead of the current blind bidding, the sequencer could run a quick second-price auction each block for the top slots. For example, it could take the top N fee bids, then sort them, but charge each of those N transactions the fee of the (N+1)th bidder. This encourages truthful bidding for priority. The sequencer then includes those N high-priority TXs first. The revenue difference could be burned or given to the protocol, depending on the model. This is a better monetization than first-price because it can extract more willingness-to-pay without overcharging, and it’s fairer to users. The economic model here is aligning the priority fee with users’ true utility for being first, which maximizes total surplus. The sequencer can regularly schedule such auctions. Since it’s custom, it could even allow certain types of priority essentially multiple priority markets. This fine-grained auction approach is only possible with an intelligent sequencer that knows the context of TXs.

Conclusion

The evolution of transaction sequencing from a narrow focus on MEV retention to a broader, application-specific approach represents a pivotal shift in blockchain design. While traditional fee-based ordering has its merits, its overemphasis on extracting immediate value risks undermining long-term system integrity and equitable market participation. By embracing generalised sequencing strategies, where fairness, efficiency, and tailored value delivery are paramount, blockchains can transition towards architectures that genuinely serve the diverse needs of decentralized applications. This forward-thinking yet cautious approach not only curbs exploitative practices like front-running but also paves the way for innovative revenue models and enhanced user experiences, ensuring that transaction ordering remains a robust, scalable, and integral component of the blockchain ecosystem.

Join the Sei Research Initiative

We invite developers, researchers, and community members to join us in this mission. This is an open invitation for open source collaboration to build a more scalable blockchain infrastructure. Check out Sei Protocol’s documentation, and explore Sei Foundation grant opportunities (Sei Creator Fund, Japan Ecosystem Fund). Get in touch - collaborate[at]seiresearch[dot]io