Pair Trading System

A market-neutral statistical-arbitrage engine for identifying and trading cointegrated pairs.

What it does

• Screening — scans a universe for cointegrated pairs (Engle–Granger / correlation filters)
• Signal generation — computes the spread z-score and triggers entry/exit at configurable thresholds
• Backtest — evaluates the market-neutral strategy (PnL, hit rate, max drawdown) on historical data

Why it matters

Pairs trading isolates relative mispricing while hedging out broad market direction — a core building block of quantitative and systematic equity strategies.

Stack

TypeScript · time-series statistics · backtesting harness

Research/educational project. Not investment advice.