Cointegration
Cointegration is a statistical property of a collection of time series variables which indicates that a linear combination of them is stationary. Even if individual series are non-stationary, they may move together over the long run, with their deviations from a common trend being temporary.
This concept is the mathematical foundation for pair trading and statistical arbitrage. In cryptocurrency, two tokens that are historically linked, such as a native token and a wrapped version or two assets in the same ecosystem, may exhibit cointegration.
By identifying these relationships, a trader can profit when the price spread between the two assets deviates from its historical norm, expecting it to revert. Cointegration is a much stronger relationship than correlation, as it implies a stable, long-term equilibrium.
It is a powerful tool for building mean-reversion strategies that are statistically grounded. However, cointegration can break down if the underlying relationship between the assets changes.
Monitoring for the stability of this relationship is essential for the success of any cointegration-based trading strategy. It is a cornerstone of quantitative strategies that rely on relative value rather than absolute price movement.