Understanding Statistical Arbitrage

A practical guide for traders.

The Trading Process

We take advantage of temporary price dislocations between carefully matched pairs of assets. When two historically correlated assets diverge beyond their typical range, we place opposing bets that they'll realign—profiting from the convergence regardless of overall market direction.

1. Identify Opportunity

   Our system continuously scans for pairs of historically correlated assets that have temporarily drifted apart. Don't just enter the trade, do your own research and make sure the opportunity is valid before entering the trade.

Example

• Coca-Cola (KO) and Pepsi (PEP) shares typically move together
• If KO drops 5% while PEP stays flat, our system flags this as a potential opportunity
2. Place Balanced Trade

   When an extreme divergence is detected, we:

   • Buy the underperforming asset (KO in our example)
   • Short the overperforming asset (PEP)
   • Equal dollar amounts on both sides to maintain market neutrality
3. Wait for Convergence
   • Most positions are held for 0-10 days
   • Exit when prices return to their normal relationship
   • Profits come from the convergence, regardless of overall market direction
4. Repeat
   • Our system continuously monitors thousands of potential pairs
   • Strict quality filters ensure we only trade the highest-probability opportunities

Risk Management

Core Risks & How We Handle Them

1. Market Risk - When Correlations Break

   Even strong historical relationships between assets can temporarily collapse due to news, earnings surprises, or macroeconomic shocks.

• Spread Limits - Positions close if the price gap exceeds safe historical levels
• Sector Caps - No more than 20% of capital in any single industry
2. Execution Risk - Slippage & Failed Trades

   Orders may fill at worse prices than expected (slippage), or fail entirely in fast-moving markets.

• Liquidity Checks - Only trades assets with sufficient daily volume
• Redundant Connectivity - Backup feeds to major exchanges
3. Model Risk - When History Doesn't Repeat

   Market behavior changes, making historical patterns less reliable.

• Regime Detection - Adjusts parameters when market conditions shift
• Human Oversight - Experienced traders monitor all automated decisions
4. Liquidity Risk - Getting Stuck in a Trade

   Thinly traded assets may be hard to exit without moving prices against you.

• Volume Filters - Excludes assets with less than $10M average daily volume
• Position Limits - Never holds more than 5% of an asset's typical volume

Next Steps

Here are some next steps to get started:

• Our pipelines and data processing
• Statistical Arbitrage Deep Dive
• Understanding Risk Management
• Frequently Asked Questions