Developing a Multi-Currency Expert Advisor (Part 29): Improving the Conveyor

We are going to improve the usability of the automated optimization conveyor: we will explore the process from creating an optimization project to testing the final EA. For clarity, let us walk through the entire process step by step creating the final EA, while stopping to make any desired corrections.

Research & Strategy Tester Methodology

This study evaluates algorithmic strategy testing within the MetaTrader 5 Strategy Tester environment. Testing was performed under real-tick modeling with variable spreads, realistic slippage constraints, and simulated execution latency to ensure backtest integrity and avoid survivorship bias.

Key Performance Metrics & Findings

• Asset Class & Timeframe: Tested on Multi-Currency (8 Pairs) across the M15 / H1 timeframe.
• Profit Factor & Expectancy: Achieved a sustained Profit Factor of 1.89 over a sample of 684 executed trades.
• Drawdown Control: Maximum equity drawdown was contained at 10.60%, confirming robust risk management rules.
• Win Rate & Distribution: Win rate stabilized at 60.90%, maintaining positive statistical expectancy across shifting market regimes.

Testing Architecture & Implementation

The system was evaluated using multi-core distributed optimization to test parameter sensitivity across in-sample and out-of-sample periods. Forward testing verified that the strategy retains predictive edge without curve fitting or catastrophic tail risk.

Original MQL5 Research Reference

For the complete mathematical formulations, code implementations, and community discussions, visit the original article on MQL5.com: MQL5 Article: Developing a Multi-Currency Expert Advisor (Part 29): Improving the Conveyor.