Underfitting

Underfitting happens when a model is too simple to capture the real pattern. Both training and validation error are high, and they are close together.

Examples: fitting a straight line to clearly curved data; a decision tree of depth 1 for a complex problem; too strong regularisation that forces all weights near zero; stopping training far too early.

Unlike overfitting, more data does not fix underfitting — the model cannot use it. You need more capacity.