Pairs / Statistical Arbitrage
Go long one asset and short a related one, profiting when their spread reverts.
How it works
Pairs and statistical arbitrage exploit temporary mispricings between historically related securities. In classic pairs trading, you find two co-moving stocks, and when their spread widens abnormally, you buy the laggard and short the leader, betting the gap closes. Statistical arbitrage scales this to hundreds of positions using quantitative models, staying roughly market-neutral so broad market moves cancel out and only relative mispricings drive returns. Profits are small per trade but frequent, harvested at high volume. The approach is data- and technology-intensive. Its core risk is that a relationship you assumed was temporary has permanently broken, so the spread keeps widening instead of reverting. Because the edge per trade is tiny, practitioners lever the book heavily to make it worthwhile — and that leverage converts a stuck spread into a catastrophic, forced-liquidation loss. This is the family that blew up Long-Term Capital Management in 1998: placid returns for years, then near-total ruin in a single tail event.
The trade-offs
✅ Strengths
- Market-neutral: can profit regardless of market direction
- Diversified across many small, uncorrelated bets
- Low net exposure dampens broad drawdowns
⚠️ Weaknesses
- Relationships can break permanently, causing spread losses
- Crowded quant trades can unwind violently (Aug 2007 'quant quake')
- Requires heavy data, tech, and low trading costs to work
Publicly associated with
Naming a practitioner is historical, educational context — never an endorsement.
Legends who play this way
Play the Pairs / Statistical Arbitrage style in Conviction League
Draft a critter that trades this way, train it on a simulated market, and backtest it on the leaderboard — free and fully simulated, so there's zero real-money risk.