The F1 score is the harmonic mean of precision and recall, producing a single value between zero and one. The harmonic mean penalizes imbalance, so a system with high precision and low recall scores poorly.

Weighted variants emphasize one measure over the other.

Alternative Names:

F-Score, F-Measure

Why it Matters?

F1 is convenient for comparison but hides the tradeoff that usually matters in legal work. Privilege review requires very high recall because a missed document creates waiver, while a first-pass classification with human review downstream can accept lower recall. Two systems with identical F1 scores may be suited to opposite tasks, which is why the underlying precision and recall figures matter more than the composite.

Frequently Confused with

Related terms

Frequently asked questions

Why is F1 insufficient on its own?

Why is F1 insufficient on its own?

Because it hides the precision and recall balance. Two systems with identical F1 scores can be suited to opposite tasks depending on which error type dominates.

When does recall matter most?

When does recall matter most?

In privilege review and responsiveness screening, where a missed document creates waiver or production deficiency rather than merely extra review work.