MEV Profitability in 2026: A Reproducible Evaluation

Answer first — MEV profitability cannot be stated as a reliable monthly percentage or income range. It is the residual after an executed opportunity is confirmed and every direct and indirect cost is recorded. Market state, competition, route selection, fees, failed attempts, infrastructure, and inventory risk can change that residual from positive to negative. A credible evaluation therefore publishes its method and observation window, not a projected return.

Define the result before measuring it

Use one definition across every strategy:

verified net outcome = settled proceeds - input inventory change - execution fees - relay or validator payments - infrastructure allocation - financing costs

The calculation must be based on settled transactions. A simulated opportunity, accepted bundle, or transaction signature is not yet a verified outcome.

Build a complete cost ledger

An incomplete ledger is the most common source of inflated performance claims. Record:

• Network fees, priority fees, builder payments, and Jito tips.
• Reverted, expired, replaced, or otherwise unsuccessful attempts.
• RPC, archive-node, WebSocket, server, monitoring, and storage costs.
• Inventory changes caused by partial fills or non-atomic cross-venue execution.
• Borrowing, exchange, bridge, withdrawal, and rebalancing costs where applicable.
• The FRB fee: 20% only on qualifying net profitable executed trades where the fee logic applies.
• Gas, slippage, failed transactions, liquidity changes, and market losses, which remain the user's responsibility.

Do not combine simulated values with settled values in one performance total.

Use a reproducible observation record

Every result set should declare:

Results from different windows or configurations are not directly comparable unless the differences are normalized and disclosed.

Why private submission changes risk, not certainty

The Flashbots documentation provides private transaction and bundle infrastructure for Ethereum. The Jito documentation describes atomic Solana bundles and explicitly states that receiving a bundle ID does not guarantee on-chain inclusion.

Private submission can reduce public pre-trade visibility and can make some multi-transaction paths atomic. It does not make inclusion certain, eliminate competition, prevent state changes, or make an opportunity profitable.

Chain-specific questions to test

Ethereum

• Was the full bundle simulated against an appropriate target block?
• Are gas policy, builder payment, and block expiry bounded?
• Does the selected route have a documented privacy and fallback policy?
• Is inclusion verified on the canonical chain?

Ethereum uses 12-second proof-of-stake slots, as documented by ethereum.org. The available decision window does not remove state or builder competition.

Solana

• Is the blockhash fresh and is account state current?
• Are compute limits estimated rather than maximized blindly?
• Are writable-account conflicts and priority fees recorded?
• Is Jito bundle status separated from transaction confirmation?

The Solana fee documentation explains the base and prioritization fee model. Those inputs must be captured for each attempt.

L2 and sequencer-based networks

• What ordering and preconfirmation guarantees does the sequencer publish?
• Is there a private path, or only direct submission?
• How are L1 data fees and delayed settlement handled?
• Does the strategy remain valid during sequencer downtime or reorganization?

Do not transfer assumptions from Ethereum mainnet to an L2 without documentation.

A safe evaluation workflow

1. Write the strategy trigger and invalidation conditions.
2. Run simulation without moving funds.
3. Save every candidate, including rejected and failed candidates.
4. Create a cost ledger before considering live execution.
5. Use a restricted wallet and hard limits for any controlled canary.
6. Confirm outcomes on-chain and reconcile every fee.
7. Stop when data quality, provider availability, or risk limits fail.
8. Publish only the measured window and method; never generalize it as an expected return.

FRB Agent is a non-custodial Windows tool and keeps signing keys local. Simulation is a safety gate, not a forecast. It can identify invalid transactions and estimate configured costs, but live state and ordering can still differ.

Frequently asked questions

What is a realistic MEV return?

There is no responsible universal figure. A return is meaningful only for a declared strategy, configuration, observation window, cost ledger, and set of settled transactions.

Can a private bundle lose money?

Yes. A private route can reduce some exposure and avoid certain public-broadcast failure costs, but fees, adverse state changes, partial cross-venue execution, and market movement can still produce a loss.

Does simulation prove profitability?

No. Simulation checks a transaction against a referenced state. It does not prove future inclusion, ordering, settlement, or repeatability.

What evidence should a vendor provide?

Ask for methodology, timestamps, transaction references, cost definitions, exclusions, and a way to reproduce or independently verify the sample. Reject screenshots or aggregate percentages without source data.

Primary sources and next steps

• Ethereum proof-of-stake documentation
• Flashbots documentation
• Solana fee documentation
• Jito transaction and bundle documentation

Review the MEV risk disclosure, test the simulation workflow, and use the vendor due-diligence checklist before trusting performance claims.