Lorentzian-Neural Compression Breakout
Family: hybrid · Regime: trending · Complexity: high · Asset classes: FX, Equities, Crypto · Timeframes: H1, M15
Thesis
The strategy operates on the hypothesis that price compression (Inside Bars) acts as a spring, and the direction of the ensuing expansion can be predicted by comparing current multi-dimensional indicators (Lorentzian) to historical successes and current structural positioning (NNind). By entering on a Williams %R reversal, we capture the 'break' of the spring precisely as momentum shifts, with ATR-based risk management ensuring that the position size is commensurate with current market volatility.
Components
- Lorentzian Classification (regime) — Uses non-Euclidean distance to identify if current market features (RSI, ADX, etc.) resemble historical bullish or bearish regimes.
- Neural Network Indicator (NNind) (direction) — Provides a secondary momentum direction filter based on Close Location Value (CLV) processed through a 3-layer MLP.
- Williams %R (entry) — Triggers the trade on a mean-reversion move (cross-above -80 for longs) after the regime and direction are established.
- Momentum Oscillator (exit) — Exits the trade when the rate of change slows down, indicating the impulse from the volatility expansion has peaked.
- TMA Risk Panel (risk) — Handles dynamic position sizing based on account equity and ATR-derived stop loss distances.
- Inside Bars (MTF Framework) (volatility_filter) — Serves as a volatility filter, ensuring the trade only occurs following a period of price compression (Inside Bar).
Known failure conditions
- Lorentzian KNN 'Neighbors Count' provides inconsistent regime labels in backtests (regime noise)
- Neural Network CLV output stays in a dead-zone (46-54) for extended periods during high volatility
- Market exhibits 'grinding' trends without the compression phases detected by Inside Bars.
Explore the full interactive blueprint with parameter ranges and evidence on WOBR StrategyVerse, or generate this strategy as an MT4/MT5 Expert Advisor with QuantMogul AI Engine (free download).