October 29, 2026 · 6 min
Every sybil filter is a trade-off
One L2 review flagged about 40% of airdrop-eligible addresses. Strict filters catch farmers — and real users with them.

When Linea reviewed its airdrop list with Nansen, 516,960 of 1,297,203 eligible addresses were flagged as sybils — about 40% (CMC Academy). LayerZero's preliminary list named 803,093 addresses (Cointelegraph).
Those are big numbers, and they are still not "every farmer". zkSync defended a looser filter because a stricter one would have excluded real users; it filtered clusters with more than 20 wallets (ZKsync docs).
How we handle it
- The heuristic is published: common funder within two hops, actions inside one short window, cluster size.
- Every flag carries a confidence level.
- Flagged wallets are excluded only where you ask — a toggle in Retention, a segment in Funnels.
We will never claim 100% detection. We will show you what was flagged and why.