The basics of chain analysis
- Author
- CRYPTO PORT Editorial
- Published
- Updated
- Reading time
- 5 min
In short
Chain analysis is the practice of inferring relationships between addresses, and traits of their owners, from public transaction data. Investigators and exchanges rely on it, yet its output is always an estimate rather than proof. For an ordinary user it is a way to understand how their own activity can be read.
Key points
- Infers relationships from nothing but the public ledger
- Used in practice by exchanges and investigators
- Its output is probabilistic, never conclusive
- A wrong match can harm someone with no involvement
Definition
A method of aggregating public blockchain data to estimate which addresses belong to the same actor and which services they passed through. Each conclusion carries its own, often limited, level of confidence.
Analysis starts from the structure of a transaction itself. When one transaction spends from several addresses, the analyst assumes the same party held all those keys. This common-input heuristic is what lets scattered addresses be grouped into clusters.
Outside information is then layered on top: deposit addresses published by exchanges, donation addresses shared openly, addresses named in press coverage of past incidents. Clusters get labels, and other addresses are scored by how many hops away they sit. This is broadly what an exchange consults when reviewing a deposit.
The limits matter. The common-input heuristic breaks down for transactions that several parties build together. Labels are third-party interpretations and can simply be wrong. A claim that some address 'belongs to' a named person is frequently unsupported guesswork, and a mistaken identification can freeze the accounts, or damage the reputation, of someone entirely uninvolved.
Watch out for
- · Findings are estimates and need corroboration before being treated as evidence
- · Mislabelling can put an uninvolved user's deposit on hold
- · Assume your own activity can be read the same way when deciding what to publish