Clustering Heuristic

A clustering heuristic is a behavioral rule used to infer that several distinct addresses are controlled by a single entity.

A clustering heuristic is the engine behind entity clustering. Rather than relying on off-chain identity, it reads spending patterns and applies rules — the best known being common-input ownership and change-address detection — to decide that separate addresses share one controller. Each heuristic has a different reliability profile, so clustering is layered and weighted rather than absolute. Applied across a chain's full history, these heuristics collapse millions of addresses into a smaller set of entities, which is what makes attribution, the counterparty graph, and exposure breakdown tractable. Weaker heuristics lower attribution confidence; stronger ones raise it.