Non-Stationary Time Series

Non-stationary time series are data sequences whose statistical properties, such as mean and variance, change over time. Most standard financial models assume stationarity, which simplifies the math but fails to capture the reality of markets.

In crypto, prices are notoriously non-stationary, making them difficult to model with traditional tools. When a model ignores this property, it is prone to producing inaccurate forecasts and failing to adapt to new trends.

Advanced techniques like cointegration and transformation are required to handle these series correctly. Recognizing non-stationarity is a prerequisite for building resilient quantitative models.