The Complete Guide to AI-Hybrid Trader: Mastering Your Intelligent Trading System

The Complete Guide to AI-Hybrid Trader v6.02

Mastering Your Self-Evolving Intelligent Trading System

Welcome to the AI-Hybrid Trader v6.02 Family! 🎉

Congratulations! You now own one of the most advanced self-evolving Expert Advisors available on the MQL5 marketplace. Whether you're upgrading from a previous version or starting fresh, this guide will walk you through everything you need to know — from your very first setup to mastering the powerful AI and risk management modules.

🧠 1. Core Philosophy: Adaptation is the Key to Success

The greatest strength of the AI-Hybrid Trader lies in a single word: adaptation. We designed it not as a rigid set of "if-then" rules, but as an entity that learns, adapts, and evolves alongside the market. This is achieved through four core pillars:

1. The Pillar of Sophistication (The AI Brain) 🧠

   At the heart of this EA is a two-layer AI system working in synergy:

   • Predictive Neural Network Ensemble: Think of this as your market analysis team. Multiple neural networks work together to recognize complex patterns in price movements that are often missed by conventional analysis, providing a "predictive" outlook on where the price is likely to go.
   • Reinforcement Learning (RL) Decision Engine: If the Neural Network is the analyst, the RL engine is the strategic portfolio manager. It doesn't care about predictions — it cares about results. After each trade is closed, the RL analyzes the outcome and learns. Over time, it gets better at answering the crucial question: "Given the current conditions, will taking this trade maximize my long-term profitability?"
     
     

2. The Pillar of Strength (The Adaptive Engine) ⚡

   This is what makes the EA truly alive. The Adaptive Engine constantly monitors the market's "personality." It knows when the market is in a strong trend, when it's ranging (sideways), and when it's volatile (wild). Based on this detection, it automatically adjusts the EA's internal parameters — risk tolerance, profit targets, and AI weighting.

3. The Pillar of Security (Your Safety Net) 🛡️

   Long-term profitability is impossible without robust capital protection. We've built in multiple layers of security, including the Circuit Breaker (automatic loss limits), a News Filter, and solid Risk Management.

4. 🆕 The Pillar of Self-Evolution (v6.0+) 🔬

   The EA now evaluates its own performance in real-time. It tracks every prediction it makes, grades its accuracy per market regime, and can automatically halt trading if it detects that its own model is underperforming. It also exports structured data so you can train even more powerful AI models externally. The EA is no longer just intelligent — it's self-aware.

---

⚡ 2. Quick Start Guide — Get Trading in 5 Minutes

METHOD 1: Via Strategy Tester (Recommended — Faster)

1. Open MT5 Strategy Tester (press Ctrl+R)

2. In the "Settings" tab:

   • Select the EA: AI-Hybrid-Trader v6.02
   • Select your Symbol (e.g., XAUUSD) and Timeframe (e.g., M15 or H1)
   • Set a custom period (e.g., 2024.01.01 to 2024.12.31)

3. In the "Inputs" tab:

   • Set the Operating Mode to TRAIN_AND_SAVE
   • Note the Magic Number — you will need the exact same number later!

4. Click "Start" — the backtest will run very quickly and then stop. Check the "Journal" tab to see: "Model saved successfully."

5. Now attach the EA to a live/demo chart of the same symbol:

   • Set the Operating Mode to MODE_TRADING
   • Use the SAME Magic Number as in the Strategy Tester!

6. Enable AutoTrading (the colored button in the MT5 toolbar) — the EA is now live! 🚀

METHOD 2: Via Info Panel (On a Live Chart)

1. Attach the EA to your chart with Operating Mode = MODE_TRADING
2. Enable the Info Panel in the Inputs
3. Click the "Train Model" button on the panel and watch the Visual Progress Bar fill to 100%
4. Click "Save Model" to save the trained AI brain
5. Click "Load Model" to activate the AI
6. The EA is now ready for trading! ✅

---

📋 3. Our Promise: No "History Reading"

Before we dive into the details, it's important for us to address one thing: honesty in backtesting.

Many traders have experienced the disappointment of an EA that shows incredible backtest results but fails completely in a live account. Often, this is because the EA "cheats" by "reading history" or peeking at future data during the backtest — something impossible in the live market.

AI-Hybrid Trader is designed to be 100% honest.

This EA never peeks into the future. Our Train & Save methodology is proof of this. It allows you to train the AI model on one time period (e.g., 2023) and then test it on a completely different, unseen time period (e.g., 2024). This is a true forward-testing methodology.

The backtest results you see are an honest and reliable representation of how the strategy will behave in a market it has never seen before.

🆕 Since v6.0: The AI Model Evaluator adds another layer of honesty. After deployment, you can monitor the AI's actual prediction accuracy in real-time on your dashboard. No more guessing whether the AI is working — you can see the numbers.

---

⚙️ 4. Complete Settings Guide

Let's break down each setting group so you can tailor the EA perfectly to your trading style.

📌 Group: Operating Mode

• Operating Mode: (Advanced Feature for Professional Testing!) This gives the EA a specific "task."

  • MODE_TRADING: The standard mode. The EA will trade normally, loading an existing model if available.
  • TRAIN_AND_SAVE: A special mode. The EA will not trade. Its only job is to learn from all the historical data, create an AI model, save it to a file, and stop.

• Magic Number: A unique identifier for this EA's trades. Critical: Use the same Magic Number in both training and live modes!

• Trade Comment: A label attached to each order for your reference.

• Show Info Panel: Show or hide the on-chart dashboard with real-time AI status.

📌 Group: AI Core Settings

This is the control center for the EA's "brain." The default settings are already optimized, but advanced users can fine-tune the AI's learning behavior and the balance between its two intelligence layers.

• AI Learning Speed: Controls how quickly the Reinforcement Learning engine adapts from new trade outcomes. Lower = more stable and conservative. Higher = faster adaptation to recent market conditions.

• Long-Term vs. Short-Term Focus: Determines the AI's time horizon. High values prioritize long-term cumulative gains; lower values focus on immediate, shorter-term profit opportunities.

• Hybrid Intelligence Weighting: Sets the balance between the Predictive Neural Network and the Reinforcement Learning engine. The default calibration delivers the "Hybrid" intelligence the EA is named after.

• Dynamic Adaptation: The master switch for the Adaptive Engine. Highly recommended to keep this enabled.

📌 Group: Genetic Optimizer

This feature is exclusively for use in the Strategy Tester's Optimization mode. It uses a genetic algorithm to automatically discover the best combination of AI parameters for your chosen symbol and timeframe. Enable this only when optimizing — disable it for live trading.

📌 Group: Risk Management

• Lot Sizing Mode: Choose your position sizing method:

  • Fixed Lot: Uses a constant lot size that you define manually.
  • Volatility-Adaptive Lot: (Recommended) Dynamic position sizing based on your risk percentage and current market volatility conditions.

• Risk Percentage: The percentage of your equity you risk per trade (e.g., 1.0 = 1% of equity).

• Max / Min Lot Limits: Safety boundaries to keep lot sizes within a reasonable range.

• Max Spread Filter: Protects you from poor market conditions. The EA will not open a trade when the spread is excessively wide.

📌 Group: Safety Features (Circuit Breaker)

This is your automatic loss protection system.

• Circuit Breaker: Master switch for automatic trading halt.

• Daily Loss Limit: Maximum percentage of account balance you're willing to lose in a single day. If reached, the EA stops all trading until the next day.

• Weekly Loss Limit: Same as above, but for weekly limits.

• Max Consecutive Losses: If this many consecutive losses occur, trading is paused for protection.

📌 Group: Stop Loss & Take Profit

• SL / TP Mode: Choose how Stop Loss and Take Profit distances are calculated:

  • Volatility-Adaptive: (Recommended) Dynamic SL/TP that automatically adapts to market volatility — wider in volatile conditions, tighter in calm ones.
  • Risk-Reward Ratio: Places TP based on a defined profit-to-loss ratio relative to the Stop Loss.

• Risk-Reward Ratio: Your desired profit-to-loss ratio (e.g., 2.0 means the profit target is 2× the stop loss distance).

• Volatility Multipliers for SL / TP: Fine-tune how aggressively the SL and TP are set relative to current market volatility conditions.

📌 Group: Trade Management (Exit Strategy)

• BreakEven: Secures your trade by moving the Stop Loss to break-even once price has moved sufficiently in your favor.

  🆕 New in v6.02: BreakEven now supports a Volatility-Adaptive Mode — trigger and lock distances scale automatically with current market conditions. This is now the default. You can still switch back to Fixed Distance mode if preferred.

• Trailing Stop Mode: Choose a trailing stop method to lock in profits as the price moves in your favor. Options include Fixed Distance, Volatility-Adaptive, and Step trailing modes.

📌 Group: Time & Day Filter

Gives you full control over the EA's trading schedule.

• Time Filter: Master switch for time-based filtering.

• Trading Hours: Define the start and end of the allowed trading window, based on your broker's server time.

• Day Selection: Simply enable or disable trading on specific days of the week.

📌 Group: News Filter (Your Automated Risk Manager) 📰

This is one of our most advanced safety features. It automatically protects you from the extreme volatility, slippage, and spread widening that occurs during high-impact news events.

News Filter Settings:

• Enable News Filter: The master switch for this feature.

• Pre-News Pause: The "quiet" time in minutes before a news release. (Default: 30 minutes)

• Post-News Pause: The "cool-down" period in minutes after a news release. (Default: 30 minutes)

• High-Impact Only: Highly recommended to keep enabled. Filters only high-impact (🔴 red) news events.

How the News Filter Works:

• ✅ Automatic Currency Relevance: The EA is smart enough to know which news matters for your chart:
  • If you trade EURUSD, it automatically monitors news for EUR and USD
  • If you trade GOLD (XAUUSD), it knows that USD news is the most influential
  • If you trade GBPJPY, it monitors news for GBP and JPY
• ✅ Fully Backtest-Safe: The EA automatically detects when it's running in the Strategy Tester and safely skips the calendar check — no errors, no interference with backtesting.

---

🆕 5. Self-Evolving AI Modules — Complete Guide

These powerful modules transform the EA from simply intelligent to self-evolving.

📊 5.1 — Structured Telemetry Logger

What it does: Records every trade with full transparency in structured CSV files. Every entry, exit, and model prediction is captured and stored for analysis.

Setting: Enable or disable the telemetry master switch. Recommended: always ON.

What Gets Logged:

• Entry Logs (AI_TradeLog_Entry_*.csv)
  • Timestamp, symbol, timeframe
  • All AI signal scores and confidence levels
  • Market regime at entry (Trending / Ranging / Volatile)
  • Lot size, SL/TP prices, and trade entry reason
• Exit Logs (AI_TradeLog_Exit_*.csv)
  • Trade duration and outcome (profit/loss)
  • Exit reason (SL hit, TP hit, or other)
  • Whether the AI prediction was correct
• Model Performance Logs (AI_ModelPerf_*.csv)
  • Rolling accuracy statistics (updated every 10 trades)
  • Performance breakdown by market regime

Where to Find the Files:

All CSV files are saved in your MetaTrader 5 data folder:

    MQL5\Files\ — look for files starting with AI_TradeLog_ or AI_ModelPerf_

(In MT5: File → Open Data Folder → MQL5 → Files)

🧠 5.2 — AI Model Evaluator

What it does: Continuously grades the AI's prediction accuracy. Think of it as a live report card for your AI.

Setting: Enable or disable the evaluator (Master switch). Recommended: always ON.

Key Features:

• 📈 Rolling Accuracy Tracking: Continuously measures how often the AI was right over the last N trades
• 🎯 Per-Regime Analysis: Shows accuracy separately for Trending, Ranging, and Volatile markets — so you know exactly where the AI excels and where it struggles
• ⚠️ Degradation Detection: Alerts you in the Journal when accuracy is declining significantly
• 📺 Live Dashboard: Model accuracy is shown in real-time on the Info Panel

How to interpret the Info Panel display:

• Green — Accuracy ≥ 50%: The AI is performing well ✅
• Red — Accuracy < 50%: The model may need retraining ⚠️
• "Collecting data..." — Not enough trades yet; let it run longer

🛡️ 5.3 — Risk Intelligence Engine

What it does: Takes risk management to the next level by making position sizing adaptive based on AI signal confidence — and can automatically halt trading when the model's accuracy falls too low.

Key Settings:

• Risk Intelligence Switch: Master switch for confidence-based position sizing.

• Minimum Signal Confidence: The minimum combined AI confidence required to open a trade. Trades below this threshold are blocked entirely. (Range: 0.0 – 1.0)

• Confidence-Based Scaling: Controls how aggressively lot size scales with confidence. At the midpoint setting, a high-confidence signal will result in a meaningfully larger position than a low-confidence one.

• Maximum Simultaneous Positions: Caps the number of concurrent open positions to prevent overexposure.

• Model Circuit Breaker: If enabled, automatically halts all trading when model accuracy drops below the defined minimum threshold.

• Minimum Accuracy Threshold: The percentage below which the Model Circuit Breaker activates and halts trading until the AI recovers.

How Confidence-Based Sizing Works:

1. The EA calculates a combined confidence score from both AI and rule-based signals
2. If the score is below the Minimum Signal Confidence, the trade is blocked entirely
3. If the score passes the minimum threshold, the lot size is scaled proportionally to the confidence level
4. The Confidence-Based Scaling parameter controls how aggressive this scaling is

Illustrative Example:

• Very low confidence signal → ❌ Trade BLOCKED (below minimum threshold)
• Moderate confidence signal → ✅ Trade opens with standard lot size
• High confidence signal → ✅ Trade opens with larger lot size

📡 5.4 — Training Data Pipeline

What it does: Exports structured trade data and feature snapshots for use in external AI/ML frameworks like Python, TensorFlow, or PyTorch. This is for advanced users who want to train custom models using the EA's real trading data.

Setting: Enable or disable the export pipeline. Enable when you want to collect data for external model training.

How to Use:

1. Enable the pipeline in settings
2. Let the EA trade for a period of time (days or weeks)
3. Collect the exported CSV files from the MQL5\Files\ directory
4. Use the data in your Python / Jupyter Notebook / Colab environment for custom model training
5. Future (roadmap): Hot-reload trained models back into the EA automatically via standard model formats

What Gets Exported:

Each row contains the complete market and signal snapshot at trade entry, plus the trade outcome. It includes price data, all AI signal values, confidence levels, market regime, and a binary outcome label (profitable / loss) — ready for supervised learning.

🌐 5.5 — Portfolio Orchestration

What it does: Enables multiple AI-Hybrid Trader instances (running on different symbols) to share performance data through MetaTrader's Global Variables system. An external orchestrator script can then coordinate position sizing across all instances based on real-time performance.

Key Settings:

• Portfolio Reporting: Each EA instance publishes its performance metrics (win rate, model accuracy, P&L, open positions) to shared Global Variables.
• Report Interval: How often (in seconds) the EA updates its performance report.
• External Lot Override: If enabled, the EA reads its lot size from an external source (set by a management script) instead of calculating it internally — enabling true cross-symbol position control.

Use Case Example:

You run AI-Hybrid Trader on XAUUSD, EURUSD, and GBPUSD simultaneously. Enable Portfolio Reporting on all three. Each instance publishes its win rate, model accuracy, and current P&L. An external management script reads this data and can:

• Increase position sizes for the best-performing EA instance
• Reduce exposure for underperforming instances
• Halt specific instances if their model accuracy is too low

---

🎯 6. Professional Workflow: Train → Test → Deploy

Use the Train & Save methodology to conduct honest and reliable tests. The process is simple: Train the model on one time period, then Test it on a subsequent period it has never seen.

Scenario: Train a model on GOLD H1 data from 2023 and then test it on 2024.

Step 1: Create Your AI Model (Training Phase) 🏋️

In this step, we use the Strategy Tester as a "model factory."

1. Open the Strategy Tester (Ctrl+R)
2. In the "Settings" tab: Select the EA, Symbol (XAUUSD), and Timeframe (H1). Set the custom period from 2023.01.01 to 2023.12.31
3. In the "Inputs" tab: Set Operating Mode to TRAIN_AND_SAVE
4. Click "Start"
5. The backtest will run very quickly and stop. Check the "Journal" tab for: "Model saved successfully." ✅

Step 2: Test Your Model (Validation Phase) 📈

Now, let's see how your newly created AI brain performs on a completely new market it has never seen:

1. Stay in the Strategy Tester
2. Change the custom period to 2024.01.01 to 2024.12.31
3. Change the Operating Mode to MODE_TRADING. Do not change any other parameters!
4. Click "Start"
5. Review the results — the equity curve now shows the AI's true forward-test performance

Step 3: Deploy to Live/Demo 🚀

1. Open a chart with the same symbol and timeframe
2. Drag the EA onto the chart
3. Set Operating Mode = MODE_TRADING
4. Use the SAME Magic Number as during training
5. Configure your Risk Management settings conservatively
6. Ensure the "Algo Trading" button in your MT5 toolbar is active (green)
7. 💡 Tip: Keep Structured Logging and Model Evaluation ON so you can monitor the AI's real-world performance from day one

---

📊 7. Working with Telemetry Data

One of the most powerful aspects of the EA is the structured telemetry system. Here's how to get the most out of it.

📂 Finding Your Log Files

Log files are saved in your MetaTrader 5 MQL5\Files\ directory. To find this folder:

1. In MT5, go to File → Open Data Folder
2. Navigate to the MQL5\Files\ subfolder
3. Look for files named:
   • AI_TradeLog_Entry_[SYMBOL]_[TIMEFRAME]_[MAGIC].csv
   • AI_TradeLog_Exit_[SYMBOL]_[TIMEFRAME]_[MAGIC].csv
   • AI_ModelPerf_[SYMBOL]_[TIMEFRAME]_[MAGIC].csv

📈 Analyzing Your Data in Excel or Google Sheets

1. Open the CSV file in Excel or Google Sheets
2. The first row contains column headers
3. Use filters to sort by market regime, profit, or date
4. Create pivot tables to analyze win rate per market regime
5. Build charts to visualize accuracy over time

🐍 Analyzing with Python

For advanced users, you can load the data directly into Python for deeper analysis — merging entry and exit logs, calculating per-regime win rates, visualizing equity curves, and feeding the data into your own ML pipeline.

---

🛡️ 8. Risk Intelligence Setup Guide

Follow this step-by-step guide to safely enable the Risk Intelligence module.

Phase 1: Observation (Week 1-2) 👁️

1. Keep Risk Intelligence: OFF
2. Keep Structured Logging: ON and Model Evaluation: ON
3. Let the EA trade normally and collect telemetry data
4. Monitor model accuracy on the Info Panel

Phase 2: Conservative Activation (Week 3) 🟢

1. Enable Risk Intelligence
2. Set a low Minimum Signal Confidence — this blocks only the very weakest signals
3. Set a gentle Confidence Scaling — small effect on lot size
4. Monitor which trades get blocked in the Journal

Phase 3: Full Configuration (Week 4+) 🎯

1. Gradually increase the Minimum Signal Confidence threshold as you gain confidence in the system
2. Increase the Confidence Scaling for more aggressive lot adjustment
3. Enable the Model Circuit Breaker for maximum capital protection
4. Set the Minimum Accuracy Threshold — trading halts if the AI falls below this level

---

🌐 9. Multi-EA Portfolio Setup

Here's how to run multiple AI-Hybrid Trader instances across different symbols as a coordinated portfolio.

Step 1: Set Up Individual Instances

1. Train a separate AI model for each symbol (e.g., XAUUSD, EURUSD, GBPUSD)
2. Each symbol needs its own unique Magic Number
3. Attach the EA to each chart with its trained model

Step 2: Enable Portfolio Reporting

On each instance, enable Portfolio Reporting in the settings. Each EA will begin publishing its live performance metrics to MetaTrader's Global Variables.

Step 3: Monitor Cross-Instance Performance

You can view portfolio data in MT5's Global Variables window (Tools → Global Variables). Each EA instance publishes its current win rate, model accuracy, P&L, and open position count in real-time.

Step 4 (Advanced): External Lot Override

For advanced users who write custom management scripts: enable External Lot Override on each EA instance. Your external script can then write a target lot size value which the EA reads and applies — instead of its internal calculation. This enables true centralized portfolio-level position management.

---

❓ 10. Troubleshooting & FAQ

Q: The EA is not trading at all. What should I check?

• ✅ Is AutoTrading enabled? (The button in the MT5 toolbar should be green)
• ✅ Is a trained model available? (Check the Journal for "Model loaded successfully")
• ✅ Is the spread within your configured limit?
• ✅ Is the News Filter currently blocking trades? (Check the Info Panel)
• ✅ Is the Circuit Breaker active due to daily loss limits?
• ✅ Is the Risk Intelligence module blocking low-confidence signals? (Check Journal)
• ✅ Is the Model Circuit Breaker halting trades due to low accuracy? (Check Journal)

Q: I upgraded from v6.01. Do I need to retrain my model?

Recommended, yes. While v6.01 model files are technically compatible, retraining after upgrading to v6.02 ensures the AI is calibrated to the improved trend reference engine introduced in this version. Training only takes a few minutes per symbol.

Q: Where are the telemetry CSV files?

In your MT5 data folder: MQL5\Files\. Go to File → Open Data Folder → MQL5 → Files in MetaTrader 5.

Q: The Info Panel shows "Collecting data..." for Model Accuracy

This is normal. The Model Evaluator needs at least a few completed trades to calculate accuracy. Let the EA run longer.

Q: My model accuracy is below 50%. Should I worry?

Not necessarily. A few things to consider:

• If you just started, give it more time — at least 10+ trades for reliable accuracy readings
• Check the per-regime analysis in the performance logs — the AI may perform well in trending markets but struggle in ranging ones. This is normal!
• Consider retraining the model on more recent data
• If accuracy is consistently below 40%, enable the Model Circuit Breaker for additional protection

Q: Does the News Filter work during backtesting?

No — and this is by design. The EA automatically detects the Strategy Tester environment and safely disables the News Filter during backtests. This ensures clean, error-free backtesting while keeping the full news protection active in live trading.

Q: The EA entered a trade during a high-impact news event on backtest. Is this a problem?

No. As explained above, the News Filter is intentionally inactive during backtests. In live trading, the filter will correctly block entries around news events when enabled.

---

💡 11. Recommendations & Best Practices

📈 Recommended Settings

Setting	Recommendation
Minimum Balance	$500, ideally $1,000+
Broker	An ECN broker with low spreads
Leverage	1:100 or higher
Optimal Timeframes	M1, M5, M15, H1 (all timeframes benefit from v6.02's unified trend engine)
Best Symbols	XAUUSD (Gold), EURUSD, GBPUSD, USDJPY, and most major pairs
Account Type	Hedge or Netting
VPS	Strongly recommended for uninterrupted 24/7 AI learning

✅ Best Practices Checklist

1. ✅ Always start on a demo account for at least 2 weeks
2. ✅ Train the AI model first before going live — never skip this step
3. ✅ Keep Structured Logging ON — transparency is your best friend
4. ✅ Keep Model Evaluation ON — so you always know your AI's performance
5. ✅ Keep the News Filter enabled — it protects against extreme volatility events
6. ✅ Keep the Circuit Breaker enabled — it protects your capital during losing streaks
7. ✅ Use Volatility-Adaptive Lot Sizing for intelligent position management
8. ✅ Retrain periodically (monthly or quarterly) for optimal performance
9. ✅ Review telemetry data weekly to understand the AI's behavior patterns
10. ✅ Use VPS for uninterrupted operation and continuous AI learning

⚠️ Common Mistakes to Avoid

• ❌ DON'T use a different Magic Number between training and live trading
• ❌ DON'T skip the training step — the AI needs training data to function
• ❌ DON'T enable Risk Intelligence immediately — observe first, then enable gradually
• ❌ DON'T ignore the Model Evaluator accuracy readings — they tell you the truth
• ❌ DON'T disable all safety features — they exist to protect your capital
• ❌ DON'T use on timeframes higher than H4 — the AI is optimized for intraday analysis

---

🚀 12. What's New in v6.02

If you're upgrading from v6.01, here's a concise summary of what changed:

---

🙏 Thank You

Thank you for choosing AI-Hybrid Trader v6.02. We have poured our passion for AI and trading into building this system, and we are confident it will be an invaluable tool in your trading journey.

If you have questions, suggestions, or feedback, please don't hesitate to reach out via the MQL5 marketplace messaging system. We are here to help you succeed.

Happy Trading! 🚀📈

⚖️ Risk Disclaimer

Trading foreign exchange, commodities, and other financial instruments on margin carries a high level of risk and may not be suitable for all investors. The high degree of leverage can work against you as well as for you. Before deciding to invest, you should carefully consider your investment objectives, level of experience, and risk appetite. Past performance is not indicative of future results. The possibility exists that you could sustain a loss of some or all of your initial investment, and therefore you should not invest money that you cannot afford to lose. Use this software responsibly.