Model Drift
Model drift refers to the gradual deterioration of a model’s performance as market behavior changes. Signals that once worked may weaken due to competition, regulation, or evolving investor dynamics.
Drift is inevitable. Markets adapt, and edges erode.
Quant frameworks monitor drift by tracking live performance, factor efficacy, and drawdown patterns. Persistent deviations from expectations trigger review and recalibration.
Managing drift involves refreshing data, adjusting parameters, introducing new signals, or retiring outdated models.
Model drift shifts quant investing from static design to continuous evolution. Sustainable strategies treat models as living systems, not finished products.
