Contents · 37 / 58
Analyse Pairs Trading and Cointegration
Updated 31 October 2026 · 3 min read · 14 Contents
Plain-language meaning
Pairs Trading Analytics scans listings in one sector on one exchange for pairs whose log prices move together in a stable, mean-reverting relationship, and tracks how far each pair's spread currently sits from its recent average.
Why it is useful #
It identifies pairs with a statistically stationary spread and gives each one a hedge ratio, a half-life and a current spread z-score, so relative-value research starts from tested evidence instead of a chart impression.
Where it appears and the workflow #
Open Pair Analytics (/Analytics/Pairs). Choose an exchange and a sector, then review the listed pairs with their hedge ratio, ADF statistic, half-life, current z-score and any divergence flag. Analysis requires the advanced-research entitlement.
Calculation or source #
For each candidate pair the hedge ratio comes from an ordinary least-squares regression of log price B on log price A (log B = a + b × log A) over the latest 250 shared sessions. The residual spread is tested with the Augmented Dickey-Fuller test against a 5% critical value of −3.37. The z-score uses a 60-session rolling window, and a pair is flagged as diverged at ±2.0.
Prerequisites, inputs, period and unit #
Prerequisites are synchronised closes for both listings with at least 200 overlapping sessions, and an absolute log-return correlation of at least 0.60. The z-score is unitless; half-life is in sessions.
Worked example #
Illustrative only: a pair with hedge ratio 1.45, ADF statistic −3.62 (beyond the −3.37 critical value), half-life 14 sessions and current z-score +2.15 has a spread 2.15 standard deviations above its 60-session mean and is flagged as diverged.
What high and low mean #
A z-score above +2.0 means the spread is high relative to its recent history and below −2.0 means it is low. It measures statistical distance, not a guaranteed return to the mean.
Positive, negative and zero #
A positive z-score means the spread is above its rolling mean, negative below and zero at the mean. A more negative ADF statistic is stronger evidence of a stationary spread.
Limitations and common mistakes #
Cointegration can break after mergers, regulation changes or results surprises. A test passed on past data says nothing certain about the future, and costs and slippage are not modelled.
Market-specific differences #
Pairs are formed only within a single exchange, so they never mix currencies, holiday calendars or settlement cycles.
Suggested next steps #
Choose one sector, open the pair with the most negative ADF statistic and record its half-life and z-score. Check both companies' recent filings for any event that could break the relationship.
Educational use only #
This guide explains descriptive research evidence. It is not investment advice, a price prediction, a recommendation, a claim of predictive accuracy, or an instruction to buy, sell, rebalance or place an order.