Model Performance
Per-round Brier score, log-loss, RPS, and calibration error — tracking how the Match Engine improves over time.
0.1675
Avg Brier Score
lower = better (0.222 = random)
60.4%
Avg Accuracy
correct outcome picked
0.2089
Avg ECE
calibration error (↓ = better)
Model Performance by Round
Brier score and log-loss across rounds — lower is better.
Model Calibration
Expected Calibration Error (ECE) and Ranked Probability Score (RPS) — lower is better.
Detailed Metrics by Round
| Round | N | Brier ↓ | Log-loss ↓ | RPS ↓ | ECE ↓ | Accuracy ↑ | Version |
|---|---|---|---|---|---|---|---|
| Group Stage - 2 | 12 | 0.1418 | 0.7568 | 0.1490 | 0.2097 | 66.7% | meta_v1_n34 |
| Group Stage - 1 | 24 | 0.1932 | 0.9673 | 0.1574 | 0.2081 | 54.2% | ensemble_v1_poisson+elo |