Project Brief: MQL5 Machine Learning Trading Robot

Develop an MQL5 Expert Advisor that trades entirely via Machine Learning or Deep Learning.
The EA must open only one trade at a time and provide only one configurable input for the user: RiskPercentPerTrade .
All other trading logic and parameters (signal generation, SL/TP, exit rules, timeframes, filters, etc.) are defined by the developer through the ML/DL model.

Requirements

• Platform: MetaTrader 5

• Execution: Maximum one open position at a time (no hedging or pyramiding)

• Single user input: RiskPercentPerTrade (e.g., 0.5% – 2%)

• Autonomous decision-making: Model defines when to enter, where to place stop loss and take profit, and how to exit trades

• Adaptivity: Must adjust to changing market conditions (walk-forward, retraining, or adaptive pipeline)

• Risk Management:

  • Position sizing calculated dynamically from RiskPercentPerTrade and model-defined stop loss

  • No martingale, grid, or averaging techniques

• Backtesting & Validation:

  • Walk-forward backtests across multiple market periods

  • Metrics required: Profit Factor, Sharpe ratio, Max Drawdown, CAGR, and stability across sub-periods

  • Developer must provide detailed backtest results with equity curves

• Deliverables:

  1. Fully functional MQL5 EA (compiled and source code)

  2. Backtest report with walk-forward validation and performance metrics

  3. Technical documentation explaining model approach, features used, and risk management logic

  4. User guide (1 page): installation, single input usage, and reading logs

Acceptance Criteria

• Only one active position at any time

• Only one input parameter: RiskPercentPerTrade

• Reproducible backtests without overfitting

• Backtest performance with controlled drawdown and consistent stability across periods

• Clean and documented code

Candidate Profile

• Proven experience developing MQL5 Expert Advisors

• Demonstrated work in Machine Learning / Deep Learning applied to trading

• Ability to deliver robust backtests and explain technical design clearly

• Preferably with prior experience in adaptive or self-learning trading systems