Skip to content

FRM Part II · FRM Exam Part II · Credit Scoring and Rating

A validator compares a rating model's performance on the development sample with a later holdout sample and finds a much lower accuracy ratio on the holdout. Which issue is the most likely explanation?

Overfitting is the most likely explanation. When a model performs much worse on a holdout sample than on its development data, it has fitted noise specific to the development sample rather than stable relationships, so out-of-sample discriminatory power drops.

  1. AOverfitting to the development sampleCorrect
  2. BExcessive rating migration
  3. CLow default correlation
  4. DUse of a through-the-cycle philosophy

Explanation

A large deterioration from in-sample to out-of-sample performance is the classic sign of overfitting: the model captured noise specific to the development data. Migration, default correlation and rating philosophy affect stability or test interpretation but do not explain a gap between in-sample and holdout discrimination.

Did you get it right without looking?

One question tells you little. A timed set on Credit Scoring and Rating shows your real accuracy, how long you take and where you lose marks.

Practise Credit Scoring and RatingFree diagnostic mock

More Credit Scoring and Rating questions