Model Drift Detection

Definition

Model drift detection represents the systematic identification of performance degradation in quantitative trading models when the statistical properties of input data deviate from the training distribution. Within cryptocurrency and derivatives markets, this process is essential because high-frequency shifts in volatility, liquidity, and correlation regimes often invalidate previously optimized pricing engines. Traders utilize these monitoring frameworks to trigger automated re-calibration cycles, ensuring that hedging strategies and option Greeks remain anchored to current market realities.