Out-of-Sample Data Feature Overview

Introduction:

The Out-of-Sample Data feature is designed to validate trading robots using independent market data that was not part of the original training or testing datasets. This helps prevent overfitting and ensures that strategies are reliable in real-world trading conditions.

Purpose:

The primary goal of this feature is to improve the robustness of trading strategies by using unseen data, increasing the likelihood of consistent performance in varying market conditions.

How It Works:

• Data Segmentation: The feature separates market data into training and out-of-sample datasets, ensuring that validation is performed on independent data.
• Performance Analysis: Robots are tested on out-of-sample data to assess their resilience and ability to adapt to new market conditions.

Best Practices:

• Always use the Out-of-Sample Data feature when optimizing strategies to ensure reliability in live trading.
• Compare performance metrics on both training and out-of-sample datasets to avoid overfitting.