Analisi personalizzata

Analisi personalizzata, introdotto in Build 131 Dev 4, enables you to extend the default capabilities of StrategyQuant X by running custom computations and filtering logic on individual strategies or entire databanks. This feature is implemented via code snippets and offers significant flexibility for advanced users who require custom metrics or integration with external analytical tools.

Panoramica

Custom Analysis enhances the existing evaluation and filtering mechanisms by allowing:

• Creation of new custom performance metrics
• Filtering strategies using user-defined logic
• Integration with external analysis tools (e.g., Python-based scripts)
• Strategy-level and portfolio-level post-processing

Types of Custom Analysis

There are two primary scopes where custom analysis can be applied:

1. Per Strategy Analysis

This analysis is executed after all backtests and cross-checks are complete, but before the strategy is saved into the databank.

Capabilities:

• Access and process all available backtest and cross-check data
• Compute custom metrics or indicators
• Optionally act as a filtro: if the custom analysis method returns false, the strategy can be excluded from the databank (if filtering is enabled)

Configurazione:

• Available in the Ranking tab of the Costruttore, Ritestatore, e Ottimizzatore
• Abilita il Filtro option to automatically exclude strategies that fail the custom analysis

Example Use Case: Reject strategies with high drawdown across all cross-checks or flag those with specific behavioral patterns in equity curves.

2. Per Databank Analysis

This type of analysis processes the entire databank after strategy generation or import, enabling broader comparisons and batch operations.

Capabilities:

• Access all strategies in the databank simultaneously
• Perform comparative analysis across strategies
• Compute aggregate statistics or apply batch filtering/removal
• Modify or restructure the databank content

Configurazione:

• Can only be used within a Project personalizzato, via the Attività di analisi personalizzata
• The task supports up to four sequential steps:
  1. Per Strategy Analysis
  2. Per Databank Analysis
  3. Per Strategy Analysis
  4. Per Databank Analysis

This allows layered processing, where individual strategy evaluations can be followed by portfolio-wide adjustments or vice versa.

Example Use Case: Rank strategies by a custom portfolio metric and remove the bottom-performing quartile.