Lorentzian-HP Ghost Persistence Strategy

Family: hybrid · Regime: trending · Complexity: high · Asset classes: FX, Equities, Crypto · Timeframes: H1, H4, D1

Thesis

Market edge exists when ML-detected regimes align with causal cycles and statistical persistence. By entering on structural 'urgency' signals from linear regression channels only when autocorrelation confirms momentum persistence, we filter out the noise of mean-reverting phases. Lorentzian distance provides a superior classification metric because it is less sensitive to the 'fat tails' often found in financial time-series compared to standard Euclidean-based KNN.

Components

• Lorentzian Classification (regime) — Identifies the current market regime by comparing multidimensional feature sets to historical winners via K-Nearest Neighbors.
• Hodrick-Prescott Filter (HP) (direction) — Provides the macro directional bias by decomposing price into a causal trend line, ensuring we trade with the underlying cycle.
• Ghost Channels [mph1nance] (entry) — Detects structural breakouts and momentum acceleration through parallel regression rails and the proprietary urgency score.
• Auto Fibonacci (exit) — Calculates potential profit-taking targets based on standard retracement and extension levels derived from the current swing.
• Nadaraya-Watson Envelope (EMA Proxy) (risk) — Provides a volatility-adjusted baseline to scale position sizes and assess if price is overextended relative to the trend.
• Autocorrelation Function (ACF) (confirmation) — Confirms that the current price movement is persistent (trending) rather than mean-reverting (ranging).
• ATR SL Finder (volatility_filter) — Establishes a dynamic, volatility-based stop loss level anchored to recent high/low ranges.

Known failure conditions

• Lorentzian classifier feature selection (RSI/CCI) becomes noise-heavy in low-volatility sideways markets.
• ACF flips frequently between positive and negative values, indicating a loss of serial correlation.
• Price oscillates around the HP trend line, leading to rapid directional signal flipping (whipsaw).

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).