KI Power signale

Manus Machiene Learning Beast – Indicator Description

Overview

Manus Machiene Learning Beast is an advanced TradingView indicator that combines Machine Learning (Lorentzian Classification) with trend, volatility, and market regime filters to generate high-quality long and short trade signals.

The indicator is designed for rule-based, disciplined trading and works especially well for set-and-forget, semi-automated, or fully automated execution workflows.

⸻

Core Concept

At its core, the indicator uses a machine-learning model based on a modified K-Nearest Neighbors (KNN) approach.
Instead of standard Euclidean distance, it applies Lorentzian distance, which:
• Reduces the impact of outliers
• Accounts for market distortions caused by volatility spikes and major events
• Produces more robust predictions in real market conditions

The model does not attempt to predict exact tops or bottoms.
Instead, it estimates the probable price direction over the next 4 bars.

⸻

Signal Logic

Long Signals

A long signal is generated when:
• The ML model predicts a positive directional bias
• All enabled filters are satisfied
• A new directional change is detected (non-repainting)
• Optional trend filters (EMA / SMA) confirm the direction
• Optional kernel regression confirms bullish momentum

📍 Displayed as a green label below the bar

Short Signals

A short signal is generated when:
• The ML model predicts a negative directional bias
• Filters confirm bearish conditions
• A new directional change occurs
• Trend and kernel filters align

📍 Displayed as a red label above the bar

⸻

Filters & Components

All filters are modular and can be enabled or disabled individually.

1. Volatility Filter
• Avoids trading during extremely low or chaotic volatility conditions

2. Regime Filter (Trend vs Range)
• Attempts to filter out sideways markets
• Especially important for ML-based systems

3. ADX Filter (Optional)
• Trades only when sufficient trend strength is present

4. EMA / SMA Trend Filters
• Classic trend confirmation (e.g., 200 EMA / 200 SMA)
• Ensures trades are aligned with the higher-timeframe trend

5. Kernel Regression (Nadaraya-Watson)
• Smooths price behavior
• Acts as a momentum and trend confirmation filter
• Can be used in standard or smoothed mode

⸻

Moving Average Overlays

For visual market context, the indicator includes optional overlays:
• ✅ SMA 200
• ✅ HMA 200

Both can be toggled via checkboxes and are visual aids only, unless explicitly enabled as filters.

⸻

Exit Logic

Two exit methods are available:

1. Fixed Exit
• Trades close after 4 bars
• Matches the ML model’s training horizon

2. Dynamic Exit
• Uses kernel regression and signal changes
• Designed to let profits run in strong trends

⚠️ Recommended only when no additional trend filters are active.

⸻

Backtesting & Trade Statistics

The indicator includes an on-chart statistics panel showing:
• Win rate
• Total trades
• Win/Loss ratio
• Early signal flips (useful for identifying choppy markets)

⚠️ This is intended for calibration and optimization only, not as a replacement for full strategy backtesting.

⸻

Typical Use Cases
• Swing trading (M15 – H4)
• Rule-based discretionary trading
• Set-and-forget trading
• TradingView alerts → MT4/MT5 → EA execution
• Prop-firm trading (e.g. FTMO), with proper risk management

⸻

Important Disclaimer

This indicator:
• ❌ does not guarantee profits
• ❌ is not a “holy grail”
• ✅ is a decision-support and structure tool

It performs best when:
• Combined with strict risk management (e.g. ATR-based stops)
• Used in trending or expanding markets
• Executed with discipline and consistency