# Base Rate

*noun*

say it: **“BAYS RAYT”** **/ˈbeɪs ˈreɪt/**

How often something happens in a whole group, before any other information is taken into account.

## What is a **base rate**?

A **base rate** is how common the positive answer really is in a group, before any model is involved, such as the
share of a district's applicants who truly qualify. When groups' base rates differ, the fairness tests can't all be
met at once: equal selection rates would then mean unequal true and false positive rates.

```python
qualified = {"North": 40, "South": 15}
applied = {"North": 100, "South": 60}
for district in qualified:
    print(district, qualified[district] / applied[district])
# North 0.4
# South 0.25
```

## In a sentence

- “Because the **base rate** differed between districts, equal selection rates couldn't also give equal recall.”
- “North's **base rate** of qualified applicants was 0.4, South's 0.25.”

Level: Journeyman · [More AI Basics terms at this level](https://www.ask-codey.com/glossary?category=ai&level=3)

Updated: November 19, 2026

Source: https://www.ask-codey.com/glossary/base-rate
