Statistical Arbitrage Mean Reversion

Entry Logic

• Long Entry: A stock in a predefined basket of stocks underperforms the basket's average performance by a statistically significant amount.
• Short Entry: A stock in a predefined basket of stocks outperforms the basket's average performance by a statistically significant amount.
• Confirmation: The stock's performance begins to revert to the basket's mean.
• Timeframe: 1-minute to 5-minute.
• Location: Extreme relative performance within a basket.
• Market Condition: Any.

Exit Logic

• Profit Target: The stock's performance converges with the basket's mean.
• Scaling Out: No.
• Trailing Stop: No.
• Signal Failure: Exit if the divergence continues.
• Opposite Signal: Not applicable.
• Time Expiration: Exit by the end of the day.
• Momentum Loss: Exit if the reversion stalls.

Stop Loss Structure

• Hard Stop: A pre-defined maximum divergence.
• Soft Stop: If the stock does not start to revert quickly.
• Max Dollar Loss: Varies.
• Max Percent Loss: 2%.
• Structural Stop: Not applicable.

Risk Management Framework

• Risk Per Trade: 0.5%.
• Daily Limit: Many trades per day.
• Weekly Limit: 5%.
• Max Drawdown: 15%.
• R:R Requirement: 1.5:1.

Position Sizing Model

• Sizing Approach: Long the underperforming stock and short the basket (or a representative ETF) in a dollar-neutral way.
• Volatility Adjustment: Yes, based on the volatility of the stock and the basket.
• Conviction Sizing: No.
• Scaling In: No.
• Scaling Out: No.

Trade Filtering

• Market Conditions: Requires a high-speed connection and sophisticated software.
• Setups: Based on complex statistical models.
• Instruments: Baskets of highly correlated stocks (e.g., all the stocks in a specific industry group).
• Time Restrictions: None.
• Chop/News Avoidance: The model should account for news.

Context Framework

• Trend Direction: Not applicable.
• VWAP Relationship: Not applicable.
• MA Relationship: Not applicable.
• Range Location: At an extreme of relative performance.
• Higher TF Alignment: Not applicable.

Trade Management Rules

• Breakeven: No.
• Scale Out: No.
• Add Size: No.
• Fast vs Slow Moves: This is a high-frequency trading strategy.

Time Rules

• Optimal Window: Any.
• Times to Avoid: None.
• Session Notes: The domain of quantitative hedge funds.

Setup Classification

• A+ Setup: A large, statistically significant divergence.
• A Setup: A moderate divergence.
• B Setup: A minor divergence.
• C Setup: No signal.

Market Selection Criteria

• Instruments: Liquid stocks.
• Volume: Very high.
• Volatility: Any.

Statistical Edge Metrics

• Win Rate: 50-60%.
• Avg Win: 1.1R.
• Avg Loss: 1R.
• Profit Factor: 1.1.
• Expectancy: +0.05R (but traded at very high frequency).

Failure Conditions

• Market Conditions: A breakdown in the statistical relationships that the model is based on.
• Specific Scenarios: A flash crash, a major market-wide event.

Psychological Rules

• Discipline: This is a purely automated strategy. Human emotions are a liability.

Advanced Components

• Regime Detection: The model itself is a form of regime detection.
• Filters: The model uses many filters to identify trading opportunities.
• Correlation: The strategy is based on correlation.
• MTF Alignment: Not applicable.

Location

• Strongest: In electronic markets with low transaction costs.
• Weakest: In illiquid, high-cost markets.