Correlation & Cointegration Analysis

Correlation says whether two assets move together, and for a pair trade it is the wrong question: two prices can be strongly correlated and drift apart forever. Cointegration asks whether they come back, which is the property a spread actually needs, and both are measured here on the same screen.

• Correlation across many assets at once, over rolling windows rather than one static figure
• Cointegration tests that isolate the pairs which really mean revert
• Stability of each relationship through time, so a number computed on one regime is not read as today