Skip to main content
Loading accuracy data...

Historical Evaluation

Loading historical evaluation…

Forecast Calibration

Probability Calibration

A well-calibrated model assigns probabilities that match observed frequencies. Bins close to the dashed reference line indicate good calibration.

Perfect calibration (y = x)Observed bins (size = sample count)

Calibration Error

0.0603

Lower is better

Forecast Sharpness

1.2369

Lower is better

vs Uniform Baseline

Baseline not provided

Sample size: 2250 observations

Note: Only Round 4 of 2026 has actual results; no multi-season historical (predicted, observed) pairs are available in-repo. Isotonic calibration requires materially more data (target: 2023+2024+2025 backfilled) before calibration.applied can be set to true. Until then, exported probabilities are raw Plackett-Luce Monte Carlo outputs.

Generated: 5/27/2026, 11:20:13 AM