Decoder. plain-English AI glossary

F1 Score

● Core

The harmonic mean of precision and recall—a single number that balances both when you can't choose.

Think of it like

Like the batting average in baseball: combines on-base percentage and slugging.

Example

Precision is 90%, recall is 80%. F1 score is approximately 85%.

How it actually works

F1 is useful when you care equally about precision and recall and want a single metric. It penalizes imbalanced tradeoffs (90% precision, 10% recall gets a low F1). Downside: it's less intuitive than its components. Most teams track all three: precision, recall, and F1.

For product teams

F1 is good for general-purpose classification. If you care more about one metric, optimize that instead.

For engineers

Report F1 as a headline metric, but plot precision-recall curves to see the full story.

Precision — Component.
Recall — Component.

Read anything AI without the jargon

Look up any term in plain English, or save terms as you read with the free Chrome extension.

Open Decoder Add to Chrome