← Lessons

quiz vs the machine

Gold1400

Machine Learning

Bias, variance & overfitting

The single most important tradeoff in supervised learning.

6 min read · core · beat Gold to climb

The tradeoff

Every model error decomposes into bias (wrong assumptions — underfitting) and variance (sensitivity to the training sample — overfitting).

How you spot it

  • Underfitting: high training error and high validation error.
  • Overfitting: low training error but high validation error — the gap is the tell.

Levers

Increase capacity or features to cut bias. Add regularization, dropout, or more data to cut variance. The art is finding the capacity where validation error bottoms out.

Key idea

You never optimize training error — you optimize generalization. A model that memorizes the training set has learned nothing useful.

Check yourself

Answer to earn rating on the learn ladder.

1. Low training error but high validation error indicates…

2. Which most directly reduces variance?