Pair is strongly reverting towards the mean with accelerating velocity and stable volatility.
Regime: STRONG_REVERSION (high confidence)
Correlation: 0.71 · Cointegrated: yes
Z-score: 1.47 entry / 1.67 rolling
Half-life 1.0h · Hurst 0.87 · Hedge ratio 0.69
Pair volatility: 25.69%
Backtest: 57.14% win · Sharpe -3.49 · -0.89% return · 1.20% max drawdown
Pair Analysis: Long SMH / Short ASML
This setup is a classic semiconductor sector mean-reversion trade, capitalizing on the relative performance divergence between the broader semiconductor ETF (SMH) and the specialized lithography leader (ASML).
Statistical Diagnostics
- Roll Z-Score: +1.67 (1h) / +1.58 (4h). A positive z-score indicates that ASML is currently "rich" relative to the SMH basket. Your direction (Long SMH / Short ASML) is quant-optimal, betting on the spread to mean-revert as ASML cools off or SMH catches up.
- Correlation: 0.71 (1h) to 0.80 (4h). The pair maintains a strong, stable relationship, which is critical for a mean-reversion strategy.
- Half-Life: ~8 hours (4h timeframe). This suggests a relatively fast expected reversion, making this a tactical, short-to-medium-term trade rather than a long-term hold.
- Cointegration: Confirmed on 1h and 4h timeframes, supporting the statistical validity of the pair.
Market Context & Sentiment
- Sector Sentiment: Bullish. The broader chip/AI sector is experiencing strong tailwinds from data center demand and structural infrastructure buildouts.
- Trade Logic: While the sector is bullish, the positive z-score suggests ASML has outperformed the broader basket (SMH) recently. This trade is not a bet against semiconductors, but a bet that ASML’s recent relative strength is overextended and will normalize toward the sector average.
Execution Notes
- Venue: Both assets are available on Hyperliquid (xyz deployer).
- Sizing: Use 4h hedge weights for sizing (approx. 42% SMH / 58% ASML).
- Risk: The primary risk is a sector-wide breakout where ASML continues to lead the rally, pushing the z-score further into extreme territory. Monitor the 1h z-score closely; if it pushes above +2.0, the mean-reversion thesis may be invalidated by a regime shift.
SMH
ASML