The Origins of MEV: Systematic Attribution of Arbitrage Opportunity Creation at Scale

Abstract: Maximal Extractable Value (MEV) represents billions of dollars in extracted value that fundamentally shapes blockchain network dynamics and participant incentives. While research has focused on MEV extraction and mitigation, we lack systematic methods to attribute MEV opportunities to their on-chain origins. This paper formalizes the MEV opportunity attribution problem and introduces a systems framework for identifying which transactions create arbitrage opportunities and quantifying their contributions. We design and evaluate four attribution methods for atomic arbitrage on EVM-compatible networks: bot-data-driven, simulation-based, coefficient-based, and Shapley-based approaches. Through large-scale retrospective analysis spanning over one million blocks on Polygon, we demonstrate that the majority of atomic arbitrage opportunities can be traced to single source transactions, validating our central hypothesis about competitive MEV markets. We quantify a highly concentrated distribution of MEV creation, where a small subset of protocols generates most opportunities, and provide comparative analysis of method trade-offs in accuracy, cost, and scalability. Our findings offer insights for protocol designers reducing MEV leakage, validators optimizing transaction ordering, and analysts measuring ecosystem health through opportunity creation.

• The paper presents a framework that rigorously attributes atomic arbitrage MEV creation to distinct, causally linked on-chain transactions.
• It integrates bot-data, simulation, coefficient, and Shapley methods, achieving 91.7% accuracy and high coverage for identifying transaction sources.
• Findings indicate over 96% of atomic arbitrage events stem from single transactions, revealing concentration among key DeFi protocols.

Systematic Attribution of MEV Opportunity Creation in Atomic Arbitrage

Introduction and Problem Statement

This paper addresses a fundamental gap in MEV (Maximal Extractable Value) research: the systematic attribution of arbitrage opportunity creation within EVM-compatible blockchains (2604.27979). While previous works have focused on measuring extracted MEV, identifying searcher strategies, or designing mitigation protocols, this study formalizes the problem of attributing the creation of MEV opportunities — specifically atomic arbitrage — to concrete on-chain actions. The research advances a systems-level perspective and develops methods to assign causality and quantify contribution among preceding transactions, enabling protocol designers, validators, and analysts to better understand and potentially influence MEV dynamics.

Methodological Framework

The core contribution is a modular attribution framework that leverages the deterministic state machine model of blockchains to enable counterfactual analysis. Given a detected atomic arbitrage transaction, the system reconstructs which prior transaction or set of transactions materially caused the transient price imbalance that was exploited. The authors limit scope to atomic arbitrage (multi-pool swaps with zero execution risk and fully on-chain observability), exploiting the property that EVM execution is reproducible for precise causal inference.

Four attribution methods are designed and comparatively evaluated:

1. Bot-Data-Driven Attribution: Utilizes real-time predictions and bidding logs from competitive MEV searchers (reinforcement learning agents with GNN and MLP predictors) as an external validation signal for opportunity source identification.
2. Simulation-Based Attribution: Counterfactually replays the transaction history to isolate the marginal effect of each candidate transaction on arbitrage profit. The transaction exhibiting the maximal positive impact is credited as the source. This approach provides high-fidelity, reproducible attributions.
3. Coefficient-Based Attribution: Employs analytical models based on price coefficient (k-value) differentials to rapidly estimate the transaction that most amplified the arbitrage opportunity, trading some accuracy for computational efficiency.
4. Shapley-Based Attribution: Applies cooperative game theory (Shapley value) to fairly distribute credit among multiple contributing transactions, using either exact calculations for small candidate sets or Monte Carlo approximation for larger ones.

The pipeline architecture supports hybrid workflows—using fast methods for broad coverage, simulation for primary attribution, and Shapley-based inference as a theoretical ground truth for ambiguous or contested cases.

Empirical Analysis and Numerical Results

Two primary datasets from Polygon (EVM chain) underpin the analysis: a comprehensive March 2026 set (1 million blocks, 360,026 atomic arbitrage events) and a focused February window for ground-truth validation. The authors' findings significantly shape understanding of MEV supply-side concentration and attribution reliability.

• Single-Source Dominance: 96.7% of atomic arbitrage events are attributable to a single transaction, with the responsible action accounting for over 70% of opportunity creation as measured by Shapley value. Only 3.3% require multi-source attribution, most commonly arising from last-in-block arbitrage and complex interdependencies.
• Protocol Concentration: MEV creation is highly concentrated. The top 5% of protocols are responsible for 73% of opportunities; Uniswap V3 and Algebra, both concentrated liquidity AMMs, account for the majority of attributed opportunities, outpacing higher volume but less capital-efficient designs such as Uniswap V2.
• Arbitrage Market Structure: The top 1% of arbitrageurs capture ~80% of extracted value; similarly, the top 1% of opportunity-creating transactions generate a comparable share. For each opportunity, an average of only 1.6 arbitrage transactions are executed, reflecting competitive yet not excessively fragmented extraction behavior.
• Methodological Performance: Simulation-based attribution achieves 91.7% accuracy (relative to Shapley ground truth) with 99.1% coverage and 12.3 ms/event computational cost, establishing it as the practical default. Coefficient-based filtering is faster (0.8 ms/event) but less accurate (77.2%). Shapley-based methods provide perfect ground truth within resource limits, converging to stable values with 1000 Monte Carlo samples for candidate sets up to ~20 transactions.

Practical and Theoretical Implications

The systematic attribution of MEV opportunity origins has several noteworthy implications:

• Protocol Design: Accurate identification of high-risk pools and transaction patterns enables designers to implement mechanisms to reduce MEV leakage, adjust fee structures, or alter liquidity allocation to minimize exploitable imbalances.
• Validator Strategy: Insights into MEV creation permit validators to optimize block construction policies, potentially aligning incentives towards MEV-aware or MEV-mitigating ordering.
• Ecosystem Analytics: The ability to aggregate attribution metrics by protocol, address, or activity category facilitates health assessments, regulatory compliance, and detection of emergent MEV hotspots.
• Research Directions: The formalism and empirical findings motivate the extension of causal attribution to more complex MEV categories (liquidations, sandwich attacks, cross-chain arbitrage) and to real-time or multi-chain analytical regimes.

The high concentration of MEV opportunity creation and extraction signals persistent economic centralization, despite protocol efforts towards democratization. The validation of the single-source hypothesis simplifies attribution in competitive environments and points to the importance of rapid arbitrage securing state synchronization across liquidity venues.

Limitations and Future Prospects

Current limitations include reliance on triangulated ground truth (bot consensus, Shapley value, and manual review), the restriction to Polygon, and the focus on atomic arbitrage. Generalization to other chains, inclusion of additional MEV types, and adaptation to high-throughput or sharded architectures (with more intricate causal graphs) remain as compelling areas for future expansion. Automated, real-time attribution and the integration of off-chain routing intelligence (e.g., cross-DEX and off-chain orderflow) are also open challenges.

Conclusion

This paper introduces a rigorously formalized and empirically validated framework for attributing MEV opportunity creation in EVM-compatible blockchains at transaction-level granularity (2604.27979). It demonstrates that nearly all atomic arbitrage opportunities can be causally linked to a single source transaction, and that MEV opportunity creation is highly concentrated among a subset of DeFi protocols, particularly concentrated liquidity AMMs. The methodology establishes both a foundation for future attribution in broader MEV contexts and practical tools for risk mitigation and ecosystem monitoring. Deployment of such attribution systems promises to inform fairer, more transparent, and MEV-aware blockchain market designs.