NeuroTrend Fiber Pro
- Experts
- Version: 1.2
- Updated: 11 May 2026
- Activations: 20
NeuroTrend Fiber Pro is a high-tech trading robot built on deep learning. The EA is based on a multilayer perceptron (MLP) integrated via the modern ONNX standard. Unlike classic indicator-based bots, this EA doesn't simply follow rules; it analyzes the market context through 35 complex features, adapting to current volatility.
Key advantages:
ONNX technology: Using a neural network model trained in Python using the Scikit-learn and RobustScaler libraries for maximum forecast accuracy. Multi-period analysis (H1 + H4 + D1): The Expert Advisor synchronizes data from two timeframes using the iBarShift algorithm, capturing both local micro-movements and the global trend.
Out-of-Sample: The model is trained on historical data from 2018-2025.
The exceptional growth in profits in 2025-2026 on data “unfamiliar” to the model proves its real predictive ability, and not simple overfitting.
Signal accuracy: ~95% (due to filtering transactions based on the MinConfidence neural network confidence level).
Basic Parameters
- **LotSize** - Fixed Lot Size
- **MagicNumber** - Unique Identifier
- **StopLoss** - Stop Loss in pips
- **TakeProfit** - Take Profit in pips
- **MaxTrades** - Maximum Simultaneous
Positions Added an optional trailing stop (UseTrailingStop = false)
Dynamic Lot
- **UseDynamicLot** - Enable automatic calculation
- **RiskPercent** - Risk Percentage of Deposit ML Parameters
ML parameters - **MinConfidence** - Minimum Confidence for Entry
- **ExitConfidence** - Position Closing Sensitivity. Determines the algorithm's confidence threshold for closing or reversing a trade. The higher the value (e.g., 0.70), the stronger the exit signal required. The lower (0.30), the faster the robot will close trades at the slightest change in trend.
- **TimeFrame1** (H1) - Main timeframe
- **TimeFrame2** (H4) - Context timeframe
- **TimeFrame3** (D1) - Level timeframe (yesterday's candlestick)
⚠️ ATTENTION: The timeframe parameters (H1, H4, D1) are constants for the current version of the ONNX neural network model. Changing these settings will cause the algorithm to operate incorrectly. Please use the default values.
Filters
- **FilterWeekends** - Weekend filter
- **UseTimeFilter** - Time filter
- **StartHour** - Start of trading
- **EndHour** - End of trading
- **MaxSpreadPips - Maximum spread
"Intelligent Position Management"
The Expert Advisor is based on a neural network model with in-depth context analysis (H4, D1) and the main timeframe (H1). The algorithm's unique feature is its dynamic exit of trades based on the ExitConfidence parameter. The system doesn't simply wait for a stop-loss, but closes or reverses the position immediately as soon as the signal confidence drops below a critical threshold. This is confirmed by the highest Sharpe ratio (9.12) in forward tests, making the strategy resilient to sudden market changes.
RECOMMENDATIONS
**Currency pair**: EURUSD
**Timeframe**: H1 (recommended)
**Minimum deposit**: $100
**Broker type**: ECN/STP with low spread
IMPORTANT for proper operation:
The Expert Advisor uses deep context analysis on higher timeframes (H4 and D1). To ensure the neural network receives accurate data from the first tick, complete the following steps before launching or testing:
1. Open the EURUSD currency pair charts on the H4 and D1 timeframes.
2. Wait a few seconds for the quotes to fully load (scroll back a bit).
3. Now you can launch the Expert Advisor on the H1 timeframe.
This ensures that the ONNX model immediately sees the full market picture and volatility levels.
ADVANTAGES
✓ Objective, emotion-free decisions
✓ Instant analysis of multiple indicators
✓ Adaptation to market conditions
✓ Transparent logic
✓ Regular updates
The Expert Advisor is fully universal and works correctly on all trading account types: Cent, Standard, ECN, and Pro. It is fully compatible with both Netting and Hedging systems. The algorithm automatically adjusts to your broker's spread and freeze levels.
I am looking for high-quality user feedback, so I am implementing a progressive pricing system:
- 1 - 10 copies: $249 (Current price) — only a few spots left!
- 11 - 20 copies: $399
- 21 - 50 copies: $599
- Final price: $799
