Accuracy
Predicted vs Actual Ks
153 pitcher-starts evaluated
MAE · Placed Bets
1.84
Ks · n=300
MAE · All Predictions
1.79
Ks · n=14,689
Bias · Placed Bets
+0.06
neutral
Bias · All Predictions
+0.04
neutral
Error Distribution
How far off are predictions, in Ks?
Green bar = perfect predictions · symmetric histogram = no bias · long tails = the model occasionally misses badly
| Placed Bets n=306 | All Predictions n=14,689 | |
|---|---|---|
| Hit-rate bands · how often is the rounded prediction within X K? | ||
| Exact (±0 K) | 14.4% | 17.8% |
| Within ±1 K | 50.0% | 49.6% |
| Within ±2 K | 75.5% | 73.8% |
| Within ±3 K | 86.9% | 88.7% |
| > 3 K off (catastrophic) | 13.1% | 11.3% |
| CI coverage · % of actual outcomes inside the predicted interval | ||
| 80% CItarget 80% | 86.6% | 87.9% |
| 90% CItarget 90% | 92.2% | 94.4% |
| 95% CItarget 95% | 98.0% | 97.5% |
| 99% CItarget 99% | 99.7% | 99.8% |
Conformally calibrated intervals
The model is slightly conservative — nominal NB 80% catches 87.9% of actuals due to discrete-quantile overshoot. Pitcher detail charts use tighter calibrated percentiles that hit each target empirically.
80% CI
[10%, 90%]
[15.1%, 84.9%]
90% CI
[5%, 95%]
[8.4%, 91.6%]
95% CI
[2.5%, 97.5%]
[4.6%, 95.4%]
99% CI
[0.5%, 99.5%]
[1.3%, 98.7%]
MAE Over Time
Drift sentinel
Flat = consistent model · rising = retrain alarm
Scatter
One dot per pitcher-start
Dashed diagonal = perfect prediction. Hover any point to inspect.