Which statement about the Brier score is correct?

• A. It Measures The Accuracy Of Probabilistic Predictions; Lower Scores Indicate Better Calibration
• B. It Measures Model Size
• C. It Measures Training Speed
• D. It Measures Data Diversity

Answer: (A)

Evaluating probabilistic forecasts with a proper scoring rule is the idea here. The Brier score measures how close your predicted probability is to what actually happened by taking (predicted probability − observed outcome) squared for each forecast and averaging those values. Since the outcome is 1 if the event occurs and 0 otherwise, smaller values mean your probabilities align better with reality. A perfect forecast gives a Brier score of 0, while worse forecasts push the score higher, up to 1. This score specifically captures how well-calibrated and refined your probabilistic predictions are, not how large the model is, how fast it trains, or how diverse the data is. That’s why this statement is the correct one.