What a wallet risk score actually measures

A single number is easy to act on and easy to misuse. A risk score of 82 out of 100 tells you almost nothing on its own — what matters is which signals produced it, because different combinations of signals justify different levels of action.

Under the hood, a score is a weighted rollup of individual signals, each with its own confidence level: shared fee payer with other wallets, funding-window overlap with a known cluster, prior interaction with a labeled address, absence of on-chain history before a funding event, and several others. None of these is scored as binary yes/no — each carries HIGH, MEDIUM, or LOW confidence depending on how corroborated it is.

This is why two wallets can carry the same score for different reasons — one driven by strong cluster membership, the other by a handful of weaker signals that happen to add up to a similar total. Treating the score as interchangeable across those two cases is where scoring systems go wrong: the number is a summary, and summaries lose information by design.

The practical takeaway: use the score to triage what needs a closer look, but make the actual decision — exclude, flag, escalate — based on the signal breakdown underneath it, not the top- line number alone.