Ethereum defi benchmark reveals challenges in protocol and contract shifts
ETH-TraceBench: A Large-Scale Event-Stream Benchmark for Ethereum DeFi under Temporal, Protocol, and Contract Shift
Machine Learning
Summary
Ethereum's DeFi system records huge amounts of time-stamped transaction data, but this data is tricky to analyze because similar symbols can mislead machine learning models. The authors created ETH-TraceBench, a large benchmark dataset to test how well models handle changes over time, new protocols, and new contracts. They found that simple models work well on known data but struggle when encountering new protocols or contract types. The study shows that popular modeling techniques don't always gain advantages from event order, and recommends focusing on difficult test conditions for improving DeFi transaction analysis.
What this means in practice
- •For blockchain developers: Evaluate and improve machine learning models for detecting and classifying DeFi transactions under changing protocols and contract conditions.
- •For financial risk teams: Use benchmark results to better anticipate how new DeFi protocols might impact monitoring and compliance systems.
Authors
Kemal Kirtac, Carsten Maple
Abstract
Ethereum decentralized finance (DeFi) provides a public, time-stamped record of transaction-level event streams, but the same public symbols can create strong machine-learning shortcuts. We introduce ETH-TraceBench, a benchmark for evaluating Ethereum DeFi representations under temporal, protocol, pool/infrastructure, and symbolic shift. The raw event universe covers January 2021-December 2025 and contains 1.35 billion transactions with logs and 5.01 billion raw log rows. Model evaluation uses a fixed 911,267-instance supervised sample, training on 2021-2024, selecting models on 2025H1, and testing on 2025H2. Simple models perform strongly on the aggregate temporal test: TraceStats-GB reaches 0.953 macro-F1 and TopicEmitterHashMLP 0.959 on the canonical DEX test set. Performance drops sharply under protocol novelty, with macro-F1 of 0.794, 0.743, and 0.766 for TraceStats-GB, TopicEmitterTrace-SGD, and TopicEmitterHashMLP, while strict unseen-pool scores remain 0.927, 0.897, and 0.935. Uniswap v4 and Ekubo v1, both absent from supervised training, are materially harder than the full test. Jointly masking emitter and topic identity reduces DEX macro-F1 to 0.916 and liquidation macro-F1 to 0.774 for TopicEmitterTrace-SGD. A standard Transformer over log-index-ordered events provides no consistent advantage over a deterministic shuffle of the same events, indicating that high aggregate scores can arise without sophisticated chronological modeling. A natural-prevalence audit estimates 2025H2 DEX prevalence among logged Ethereum transactions at about 22.5%, and a deterministic 400-transaction audit finds complete agreement with task label sources and independently re-queried raw-log counts. ETH-TraceBench therefore treats difficult transfer and controlled-input conditions, rather than a single aggregate score, as the main evaluation target.