Why does switching to a private mempool still leave me open to sandwich attacks?
機会
Private and encrypted mempools now route more than half of Ethereum transactions, but encrypting a transaction does not hide everything a searcher needs. The trading pair, the direction of a swap, and a coarse size range can be inferred from surrounding chain state and timing metadata even when the transaction bytes are fully encrypted. A paper accepted at NeurIPS 2026 derives exact leakage thresholds at which this partial information remains sufficient to run a profitable sandwich attack against users who believe they are protected. The result is a false safety assumption: users opt into private relays and stop worrying, while searchers adapt their strategies to the residual signal that leaks through. There is no production tooling today that lets a user measure how much information their transaction leaks before they submit it, so the protection is unauditable.
重要な理由
Encryption is not privacy if the residual observable signal is enough to attack, and the first tool that quantifies per-transaction leakage turns privacy from a claim into a verifiable property.
機会をどう評価するか
Opportunity Scoreは測定値ではなく、私自身の見解です。どれほど痛みを伴うか、どれほど頻繁に影響を与えるか、そして今日時点で解決策がいかに少ないか。スコアが高いほど、構築する価値が高いと私は考えています。
それが現れたときにどれほどの痛みをもたらすか。
実際にどれほど頻繁に人々がそれに直面するか。
今日時点で、それに対する優れたツールがいかに少ないか。