Model drift: why stable scores can hide worsening predictions
A lender can see the same mix of credit scores and still experience more defaults than predicted. Model drift explains why data stability, ranking and accurate risk estimates are different questions.