---
title: Precision
description: 'Among predicted-positive cases, the proportion that are truly positive.'
canonical_url: 'https://darkfactory.dev/glossary/precision'
markdown_url: 'https://darkfactory.dev/glossary/precision.md'
collection: glossary
date_published: '2026-08-03T00:00:00-04:00'
date_modified: '2026-08-26T00:00:00-04:00'
---

# Precision


## Definition

Precision is the share of predicted-positive cases that are truly positive:

`precision = true positives / (true positives + false positives)`

It answers: when the system says "positive," how often is it correct? Precision depends on the decision threshold and the prevalence of the positive class in the evaluated population, so a value measured in one setting may not transfer unchanged to another.

For multi-class tasks, report whether precision is calculated per class and then macro-averaged, weighted, or micro-averaged. Those summaries weight common and rare classes differently.

## Distinguish it from nearby terms

Recall asks how many actual positives were found. Precision asks how trustworthy positive predictions are. Specificity concerns actual negatives, and calibration concerns predicted probabilities rather than only thresholded classes.

## Check your understanding

A security scanner has 99 percent precision but flags only one of every ten real vulnerabilities. Its alerts are trustworthy, yet its low recall makes it a weak coverage control.

## Related terms

- [Recall](https://darkfactory.dev/glossary/recall)
- [False positive](https://darkfactory.dev/glossary/false-positive)
- [F1 score](https://darkfactory.dev/glossary/f1-score)

## Evidence and further reading

- [NIST AI Resource Center Glossary](https://airc.nist.gov/glossary/)
- [Google Machine Learning Glossary](https://developers.google.com/machine-learning/glossary/)