Detecting Correlation Breaks with Spread Z-Scores
Summary
This indicator measures whether two instruments that usually move together are beginning to diverge. It calculates rolling Pearson correlation between their returns and a rolling z-score for the log-price spread. A signal appears only when correlation falls below a chosen threshold while the spread reaches an extreme, combining evidence of decoupling with relative price displacement.
The guide interprets extreme negative spread readings as possible catch-up opportunities for the current symbol, and positive readings as possible pullback setups. It also recommends using the signal to review risk in hedged positions and to seek confirmation from price action or momentum. The example pairs EURUSD with GBPUSD, and the defaults use 50-bar windows and a z-score threshold of 2. No performance testing or evidence of predictive accuracy is supplied. Correlation breakdowns may mark a lasting change in the relationship rather than a temporary divergence, so the indicator cannot identify which instrument is correctly priced.
Key ideas
- The indicator compares rolling return correlation with a z-score of the log-price spread.
- A divergence marker requires both correlation below its threshold and an extreme spread reading.
- Extreme spread readings can suggest relative-value mean reversion, but may also signal a new regime.
- Price action or momentum confirmation can help assess a divergence before trading it.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.