When a model is too simple or insufficiently trained, it fails to capture the underlying trend in the data. In trading, an underfit model might miss even obvious market signals. For example, a linear model trying to predict highly non-linear forex movements could underfit. Underfitting leads to consistently poor performance on both training and test data, indicating the model needs more complexity or better features. The key in model development is to balance underfitting and overfitting (the bias–variance tradeoff) for optimal performance.

Back to Robots Glossary Index