Cointegrated Breakout Mean Reversion (CBMR)
Family: hybrid · Regime: ranging · Complexity: high · Asset classes: Equities (Pairs), Crypto (Pairs), Forex (Crosses) · Timeframes: H1, H4
Thesis
In a pair of assets that are statistically cointegrated, the price spread between them is stationary. When the spread deviates beyond 2 standard deviations (Z-score), it is likely to return to the mean. However, mean reversion timing is difficult; by requiring a momentum shift (RSIOMA) and a breakout of the previous day's range (Daily High/Low), we enter only when the market structure confirms the reversal has begun.
Components
- Cointegration (COINTEGRATION) (regime) — Acts as the regime filter to ensure that the two assets have a statistically significant mean-reverting relationship before applying arbitrage logic.
- RSIOMA (RSI of Moving Average) (direction) — Provides a smoothed momentum signal to ensure the 'snap-back' toward the mean has actually commenced.
- Daily High Low (entry) — Provides the final mechanical trigger. A breakout of the previous day's range in the direction of the mean reversion serves as a price-action confirmation.
- Chandelier Exit (Heiken Ashi) (exit) — An aggressive trailing exit based on Heiken Ashi smoothing to capture the bulk of the mean-reversion move while exiting quickly on reversals.
- ATR SL Finder (risk) — Provides a volatility-adjusted initial stop loss based on the current candle's range.
- AI Pair Trading System v1.0 (volatility_filter) — Uses a Kalman Filter and Z-score to identify when the spread between the assets has reached an extreme deviation (volatility/momentum stretch).
Known failure conditions
- Cointegration ADF statistic stays above -3.4 for extended periods, indicating a structural decoupling of the assets (divergence).
- Z-score continues to trend beyond 4.0 without returning, suggesting a 'New Normal' in the spread.
- High correlation between the assets drops significantly, rendering the Engle-Granger residuals meaningless.
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.