Evidence-appraisal glossary
Brier Score
The Brier score measures how close predicted probabilities land to what actually happened, rewarding predictions that are both confident and correct. Lower is better, with zero being perfect.
For each person it takes the squared difference between the predicted probability and the actual outcome coded as one or zero, then averages across everyone. Because it captures both discrimination and calibration in a single number, it is called an overall or proper accuracy score. A model that hedges every prediction near the base rate can still score reasonably, so the Brier score is best read alongside separate calibration and discrimination measures.
Read the full Reading the Evidence blog.
This is a plain-language methodology definition for reading research. It is general education, not medical advice.