Skip to content
D23E Research
Sanctions Evasion
Measuring what works, what doesn’t, and how to track funds reliably on-chain.

On-chain Forensics & Compliance

Evasion Under Blockchain Sanctions

ForensicsBy Liyi Zhou

Sanctioning blockchain addresses is now a common regulatory response to malicious activity. But enforcement on permissionless blockchains is messy: real-world funds move through long, branching transaction graphs and increasingly sophisticated obfuscation services.

In our paper, we use Tornado Cash as a case study to quantify the practical impact of U.S. OFAC sanctions over a 957-day period (covering 6.79M Ethereum blocks and 1.07B transactions). The core message is nuanced: sanctions change behavior substantially, but they do not eliminate illicit reliance on mixers.

  • Sanctions reduced mixer usage, but not to zero. We observe a 71.03% drop in Tornado Cash deposit volume (to ~ $2B), yet attackers still relied on Tornado Cash in 78.33% of Ethereum-related security incidents.
  • Three structural limitations show up in practice. Binary “sanctioned / not sanctioned” labels are brittle under dusting, block producer censorship is fragmented, and obfuscation services create complex flow patterns that are hard to reason about with simple heuristics.
  • A practical scoring approach for tracking. We introduce an algorithm grounded in quantitative impurity to score and track flows. In our evaluation on the Bybit exploit, it reaches 97.61% precision and 74.08% recall, while processing blocks in 0.07 ± 0.03s on average.

Practical takeaways

If you are building monitoring, compliance, or incident response for on-chain systems, consider:

  • Prefer graded risk scoring over binary labels; it’s more robust to dusting and “partial taint” edge cases.
  • Treat block-level censorship as heterogeneous; assumptions about uniform enforcement often break.
  • Build investigation tooling around speed: the time window between discovery and cash-out is often short.

Need help tracing stolen funds, understanding complex transaction flows, or building monitoring pipelines? Contact us at [email protected].

Reference