Methodology

How Proof Systems generates and validates market intelligence.

Data collection

We maintain live WebSocket connections to 5 major exchanges (Coinbase, Kraken, Gemini, OKX, Crypto.com). Order book data is captured every second. All data is timestamped and stored for verification.

Signal detection

Our system detects price differentials across exchanges in real time. Strategies include cross-exchange arbitrage (XARB), triangular arbitrage (TRIG), mean reversion (MREV), and flash crash harvesting (FCR). Each detected opportunity is evaluated against execution costs including fees, spread, and latency.

Execute first, publish second

Every signal is acted on by our execution system before being published. This ensures we have skin in the game on every signal we share. Publication is a byproduct of execution, not the other way around.

Quality filtering

Most detected differentials are noise — stale data, thin liquidity, below-fee spreads. Only signals that pass our quality threshold are published. Every qualifying signal is published. No cherry-picking.

Outcome tracking

Every signal outcome is recorded and published. Wins and losses. You can verify any signal using its unique ID at /verify.