Copula-Based Pairs Trading with Conditional Probability Thresholds

Summary

This strategy uses a fitted copula and marginal cumulative distribution functions to estimate conditional probabilities for two assets. During a formation period, the model is trained on historical prices. As new prices arrive, their marginal distributions are converted to probability values, which feed the copula’s conditional probability calculations. The method interprets opposing extreme probabilities as relative mispricing: it goes long the spread when the first asset appears low relative to the second, and short it under the reverse condition.

Positions are closed when the relevant probabilities cross exit thresholds around the midpoint. The implementation allows either an AND rule, requiring both probability crossings, or an OR rule, which may permit more exits. Its authors note that the AND condition can be too strict in some cases. The signal calculation is independent of current positions, and the code tracks open and closed trades and their associated timestamps. The document describes the mechanics but gives no backtest, transaction-cost analysis, asset-selection guidance, or evidence that the thresholds are profitable. Copula fit quality and changing dependence between the assets are important practical limitations.

Key ideas

• Fit a copula on a formation sample and transform each asset’s prices through its marginal cumulative distribution function.
• Use conditional probability extremes to signal long or short positions in the relative-value spread.
• Exit signals are based on probability crossings, with configurable AND or OR logic.
• The described implementation provides trade tracking but no performance evidence or cost analysis.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.