Building Trading Strategies

Building a Day Trading System

Day Trading System Setup

To build a day trading system in GSB Pro:

1. Load intraday data: Use 1-minute, 5-minute, or 15-minute bars depending on your trading style
2. Set entry/exit rules: Configure whether to allow same-bar exits, end-of-day exits, or position holding overnight
3. Define fitness criteria: Select optimization targets (net profit, Sharpe ratio, profit factor, etc.)
4. Start the build process: GSB Pro will automatically discover indicators and build systems

Building a Swing Trading System

Swing Trading System Characteristics

Swing trading systems typically:

• Hold positions for multiple days or weeks
• Use daily or 30-60 minute bars
• Focus on trend-following or mean-reversion
• Require fewer trades than day trading systems

The Build Process

Step 1: Configure Build Settings

Before starting the build, configure your system parameters in the GSB GUI settings:

Configure strategy settings, data sources, and genetic algorithm parameters

• Strategy Settings: Number of indicators (3 shown), secondary filter mode, and data paths
• Optimization Split: Training % (40), Test % (60), Validation % (0 or adjust as needed)
• Genetic Algorithm: Generations (1000), Population (200), Restarts (10)
• Data Configuration: Public Data Directory path and Price Data symbols

Set capital, costs, exit rules, and general strategy parameters

• Capital Quantity: Starting capital (e.g., 10000) and Quantity Mode (CapitalInvestment)
• Costs: Commission and slippage settings (TradePerSide mode)
• Exits: Time-based exits (Minutes, Stop Loss, Profit Target)
• General: Fitness Criteria (NetProfit * AvgTrade), Long/Short Trading Allowed

Step 2: Define Fitness Function and Performance Filters

The fitness function determines what makes a "good" strategy. You should also configure performance filters to eliminate poorly performing systems:

Set minimum performance thresholds for each testing period

• Min. Profit Factor: Training (1.2), Test (1.2), Validation (1.8) - filters out unprofitable systems
• Min. Trades Count: Minimum 100 trades required for statistical significance
• Min. Pearson: Correlation coefficient (0.95) ensuring equity curve consistency
• Min. Net Profit: Minimum profit threshold (1) to eliminate marginally profitable systems

Step 3: Start the Build

Once configured, click the green Play button to start the genetic algorithm build process:

Click the green play button to begin strategy discovery

GSB Pro will:

1. Generate an initial random population of strategies
2. Test each strategy on your data
3. Rank strategies by fitness
4. Breed the best strategies to create new variations
5. Repeat the process for the specified number of generations
6. Present the top-performing systems for evaluation

Monitor machine resources and optimization progress during the build

During the build, GSB displays real-time resource usage (CPU, RAM) and optimization status. The process completes when all generations finish or you manually stop it.

Step 4: Review Results

After the build completes, GSB Pro displays:

• List of discovered systems ranked by fitness
• Performance metrics for each system
• Equity curves and trade-by-trade results
• Strategy rules and indicator combinations used

Understanding the Build Results

In-Sample vs Out-of-Sample

GSB Pro automatically divides your data:

• In-Sample: Training data used to discover the system (typically 70%)
• Out-of-Sample: Reserved validation data never seen during discovery (typically 30%)

What Makes a Good System?

Look for systems that show:

• Similar performance in-sample and out-of-sample
• Reasonable number of trades (not too few, not over-fitted)
• Smooth equity curve with controlled drawdowns
• Logical indicator combinations

Next Steps

After building trading systems, the critical next step is validation and robustness testing to ensure your systems will perform in live markets.