21wiki

OXT Research: Understanding Bitcoin Privacy

5 min read

OXT Research: Understanding Bitcoin Privacy

Authors: Samourai Wallet team (TDevD and others) | Parts: 4 | Source: raw/Theory/privacy/oxt/ | Tags: series, privacy, blockchain-analysis, chain-surveillance

Overview

A 4-part series written by the developers of Samourai Wallet explaining blockchain analysis from the attacker’s perspective — so that Bitcoin users can understand and defeat it. Also available as a video playlist.

The series represents the deepest treatment of Bitcoin privacy in the 21ideas.org library. Understanding the attacker’s view is the foundation of meaningful privacy practice.

“The reality of Bitcoin privacy… is somewhere between complete anonymity and complete traceability.” — Part 1

Source files: oxt-1.md through oxt-4.md

Part 1: Blockchain Analysis and Transaction Privacy

Core claim: Bitcoin is not anonymous — it is pseudonymous. The common belief in 2011 that Bitcoin offered “anonymous payments” was dangerous and incorrect. Blockchain data is permanent; surveillance capabilities in 2011 were minimal but have since become sophisticated law enforcement tools.

What blockchain analysis is:

  • Examination of the public transaction graph to identify patterns, link addresses, and attribute transactions to real-world entities
  • Used by: law enforcement, chain analysis firms (Chainalysis, Elliptic), exchanges (for compliance), governments
  • Enables: linking KYC-exchange withdrawals to subsequent transactions, identifying when coins from different sources are combined, building probabilistic identity clusters

Key insight from Part 1: The blockchain is forever. Surveillance capabilities are increasing. Privacy choices made today affect the interpretability of past and future transactions.

Part 2: Key Concepts of Chain Analysis

CIOH (Common Input Ownership Heuristic)

The most powerful blockchain analysis heuristic: if multiple UTXOs appear in the same transaction, they likely come from the same wallet.

Why: to spend multiple UTXOs, you must provide signatures for each — which requires controlling all the corresponding private keys. Standard wallets only have one user’s keys.

Consequence: An analyst seeing UTXOs A, B, C combined in one transaction infers A, B, C belong to the same entity. If A was identified (via KYC or other means), B and C are now also identified — even if they came from separate, previously unlinked sources.

CoinJoin defeats CIOH: When 5 users combine inputs in a single transaction with equal outputs, the CIOH cannot determine which input belongs to which user.

Change Detection

When a transaction has two outputs and one is a round number while the other is an odd change amount, analysts infer which is change and which is payment. Additional heuristics:

  • Address type consistency (change often returns to same address type as input)
  • Address reuse (if you’ve received to an address before, a second output to that address is likely change)
  • Round number bias

Defense: Use equal-output CoinJoin (Whirlpool), avoid address reuse, use coin control.

Transaction Graph Analysis

Following the flow of funds through multiple hops. Even after N transactions, if coin history can be traced back to an identified source, the current UTXO is “tainted.”

Part 3: Defenses Against Chain Analysis

The third part provides the practical defensive toolkit:

CoinJoin (Whirlpool):

  • Equal-output transactions break CIOH and make input-output linking impossible
  • 5 parties in, 5 equal outputs out → analyst cannot determine which input paid which output
  • “Toxic change” concept: the unequal pre-mix change retains prior history; treat separately from post-mix coins

Address reuse avoidance:

  • Using a fresh address for each receive breaks the clustering analysts rely on
  • BIP47 / PayNym provides a reusable “identity” that generates unique per-sender addresses (ECDH-derived)

UTXO management / coin control:

  • Manually selecting which UTXOs to spend prevents accidentally mixing KYC and no-KYC coins
  • Cross-contamination: combining a KYC UTXO with a no-KYC UTXO in one transaction links them permanently via CIOH

Running your own node:

  • Wallet queries to third-party servers reveal your address set
  • Connect to your own Dojo/Electrum server to eliminate this leak

Part 4: Applying Chain Analysis Concepts to Improve Privacy

The final part is practical: applying the attack model to real transaction histories to identify and remediate privacy leaks.

Steps:

  1. Analyze your own transaction history using OXT.me (the block explorer by Samourai)
  2. Identify: address reuse, CIOH violations (UTXOs you combined), timing correlations
  3. Remediate: use Whirlpool to break history on coins you want to protect
  4. Plan future: coin control, fresh addresses, no-KYC acquisition going forward

The OXT.me Tool: A blockchain visualization and analysis tool built by the Samourai team. Allows transaction graph visualization, entity clustering, and privacy scoring of your own transactions.

Significance

This series establishes that:

  1. Bitcoin privacy is not free — it requires deliberate action
  2. The tools to compromise privacy are widely deployed and improving
  3. The tools to protect privacy (CoinJoin, coin control, self-hosted nodes, BIP47) are available but require effort
  4. Past transactions are permanent — remediation is possible but not retroactive

The OXT Research series is the analytical foundation for the privacy practice guides in privacy-practice].

Note: Samourai Wallet developers (Keonne Rodriguez and William Hill) were arrested by the DOJ in April 2024. The OXT research and Whirlpool tools remain, though the ecosystem is under pressure.

Sources

Glossary | privacy | UTXO | privacy practice | no-KYC buying | cypherpunks | Bitcoin

Graph View

Backlinks