This is the recommended optimization and forward-testing workflow used by our team to refine Expert Advisors (EAs) like BananaEA and others. It ensures you get the most robust setup using a blend of statistical filtering and real market simulation..
🔁 Step-by-Step Optimization Process
Step 1: Define Parameter Ranges
- Identify which parameters you want to optimize.
- Example:
ATR Trigger— Start at 1, Step 1, End at 10.
- Example:
- Set the ranges accordingly in the EA’s input settings.
Step 2: Run Initial Optimization (Fast Mode)
- Right-click the Strategy Tester and select “Expert Properties” → “Inputs”.
- Enable the checkbox beside each parameter to optimize.
- Run the optimization using “Fast Genetic-Based Algorithm” or “Control Points” (for speed).
- This gives a broad sweep across the parameter space to locate high-potential setups quickly.
Step 3: Select Top 3 Setups Based on Criteria
- Use the Strategy Tester’s results tab to sort and filter based on:
- Lowest Drawdown (DD)
- Highest Net Profit
- Best Profit Factor (PF)
- Save each selected setup using the “Save” button at the bottom of the inputs tab.
Step 4: Retest with “Every Tick” Model
- Load each of the saved
.setfiles. - Change the model to “Every Tick”, which is the most accurate backtest method.
- Run a full simulation again to validate robustness under realistic tick conditions.
Step 5: Save Final Sets for Forward Testing
- Save the
.setfile again (overwrite if preferred). - These files are now ready for forward testing in live or demo environments.
✅ Why This Method Works
- Step 2 quickly identifies promising configurations without burning CPU cycles.
- Step 4 ensures those promising setups aren’t curve-fitted by validating them under high-precision data.
- By using diverse performance filters (Profit, DD, PF), you hedge against overfitting to a single metric.
- The final
.setfiles are battle-tested for robustness and edge.