Edge Ensemble

Edge Ensemble is a professional trading advisor built on an ensemble of 10 independent neural networks of three different architectures: LSTM, TCN and Transformer. Each network in the ensemble acts as a standalone trading strategy, while the final trading signal is formed based on the collective decision of the entire model set.

Instead of searching for a single universal model trained using the classical 80/20 scheme, Edge Ensemble employs an ensemble specifically designed to work with non-stationary financial time series

Core idea: ensemble + different neural network types

In the classical approach, a trader relies on a single model with a single view of the market. Even with proper data splitting, such a model inevitably adapts to a specific historical period. When market structure changes, accuracy declines.

The ensemble approach addresses this problem — and Edge Ensemble takes it one step further by combining two fundamentally different neural network architectures.

How it works? The advisor uses 10 models of three different types: LSTM, TCN (Temporal Convolutional Networks) and Transformer.

Each model: 1) is trained independently on its own segment of historical data, 2) analyzes the market from a different temporal and structural perspective, and 3) has its own unique strengths and weaknesses. LSTM networks are better at capturing long-term dependencies and context; TCN networks excel at detecting regime shifts, local patterns, and impulse dynamics; Transformer networks handle global relationships and non-linear dependencies across the entire time horizon.

The final trading signal is an aggregated decision of three different classes of models.

Why ensembles may be more robust for time series than 80/20 training

Financial markets are not a stationary problem. What worked yesterday may stop working tomorrow.

An ensemble improves robustness by design:

• an error in one model does not determine the final outcome,
• local overfitting effects are smoothed out,
• different market phases are recognized by different networks,
• reactions to change become smoother and more stable.

Simply put: one model can fail, while twelve models of different types — much less often. 

Edge Ensemble is not just “more models.” It is a combination of different neural network approaches within a single ensemble. LSTM handles memory and context. TCN handles structure and local dynamics. Transformer handles global dependencies and relationships across the full data horizon. And the ensemble together provides robustness and risk control. This approach allows the advisor to maintain its edge even as market behavior changes. 

Primary timeframe: H6 — offering an optimal balance between market noise reduction and signal responsiveness.

Market conditions are continuously reassessed as new data becomes available.

The EA operates in multi-currency mode, simultaneously trading six currency pairs.

Input parameters

Range between positions (points) – Minimum distance between additional positions.
Risk per trade (%) – Risk per trade as a percentage of the deposit (automatic lot calculation).
Spread max – Maximum allowed spread for market entry.
Suffix for symbols – Broker symbol suffix (e.g., .m, _i).

=== Standard Settings ===
Comment — Order comment
Magic number — Unique trade identifier
Take profit (points) — Fixed TP
Stop loss (points) — Fixed SL
Trailing stop (points) — Trailing stop (0 — disabled)

=== Misc ===
Draw font color / size — Visual settings
Log messages (min) — How often to print messages to the event log (minutes).
Multi mode — Multi-currency mode (recommended true)

=== Your Symbols ===
Enable or disable trading for each of the 6 available symbols — you can keep only the markets you consider suitable.