How to Create Strategies in StrategyQuant X

StrategyQuant X provides multiple approaches for generating algorithmic trading strategies. The two primary build modes are Random Generation and Genetic Evolution, with an additional option to create Custom Projects for more advanced, automated workflows.

To fully understand the strategy generation logic, it is recommended to first review the section How StrategyQuant Works.

Build Modes Overview

You can choose from the following strategy build modes:

1. Random Generation

In Random Generation mode, StrategyQuant X continuously creates and evaluates randomly constructed strategies based on the selected configuration.

The top-performing strategies (as determined by your filtering and validation criteria) are saved in the Databank for review and further analysis.

Advantages:

• Faster and easier to set up than Genetic Evolution
• Can run indefinitely, potentially evaluating millions of strategies over time
• Less prone to overfitting, as no further optimization is applied post-generation

Limitations:

• Strategies are not automatically refined or optimized after creation
• To enhance these strategies, you can use them later as the initial population for a Genetic Evolution run

2. Genetic Evolution

Genetic Evolution builds upon Random Generation by first generating an initial population of strategies, then evolving them across multiple generations using genetic programming techniques.

• Strategies are selected and combined based on performance criteria to produce new, potentially superior candidates.
• The process continues until a specified number of generations is completed or no further improvement is detected.

Advantages:

• Can produce more refined and higher-performing strategies compared to Random Generation
• Allows continuous improvement of promising candidates
• Efficiently searches large strategy spaces using evolutionary algorithms

Limitations:

• Generally slower than Random Generation
• May reach suboptimal "dead ends" if not properly configured
• Strategy results are more susceptible to overfitting due to the optimization-like nature of the process
• Population size limits the diversity of the output

3. Custom Projects

While not a distinct build mode, Custom Projects enable the creation of complex workflows combining multiple strategy generation and analysis tasks.

• For example, you can configure a project to:
  • Run a series of random builds
  • Use the resulting strategies as the starting population for Genetic Evolution
  • Automatically validate, filter, and export final strategies

This flexible framework is ideal for advanced users seeking automation and full control over the strategy development process.

For more details, refer to the Introduction to Custom Projects section.