HP-Kalman Adaptive Regime Scalper

Family: hybrid · Regime: trending · Complexity: high · Asset classes: Crypto-assets (ETH/BTC, BTC/USD), Equities (Indices/Pairs) · Timeframes: 1H, 4H, 1D

Thesis

Asset pairs (like ETH/BTC) exhibit mean-reverting behavior around a dynamic mean that can be tracked via a Kalman Filter. However, these mean-reversions are most profitable when they occur in the direction of the underlying macro-cyclical trend (identified by the HP Filter) and the immediate adaptive trend (MultiKAMA). By sizing risk based on the relative performance of the asset versus its peers, we capture alpha from temporary dislocations while staying aligned with broader market momentum.

Components

Hodrick-Prescott Filter (HP) (regime) — Defines the 'Macro Regime' by filtering out high-frequency noise to ensure trades align with the lower-frequency cyclical trend.
MultiKAMA (TyphooN) (direction) — Acts as a volatility-adaptive trend filter; entry only occurs when price is on the correct side of the adaptive mean.
AI Pair Trading System v1.0 (entry) — The core trigger; identifies statistical deviations (Z-score) in a spread or ratio using a Kalman filter for dynamic mean tracking.
MACD & OSMA Cloud Hybrid (exit) — Detects momentum exhaustion via OsMA color changes to exit positions before a full trend reversal.
Performance Table (risk) — Uses relative weekly performance to calibrate stop-loss width and position sizing based on cross-asset volatility.
Fisher-based Scalping Indicator (confirmation) — Confirms the entry by ensuring the price is recovering from a cyclical extreme via the Fisher transform.

Known failure conditions

Correlation breakdown between the assets in the AI Pair Trading System (e.g., ETH/BTC decoupling permanently).
Extended periods of 'stair-stepping' in the MultiKAMA, indicating a total lack of efficiency ratio utility.
Divergence between the HP Filter Trend and the MultiKAMA for more than 50 bars, suggesting conflicting regime signals.

Explore the full interactive blueprint, parameter ranges and evidence on WOBR StrategyVerse, or generate this strategy as an MT4/MT5 Expert Advisor with QuantMogul AI Engine.