← All terms
DRBL systems
DRBL validation
DRBL was stress-tested on held-out and later seasons to check whether the estimates are useful, without claiming it beats DARKO or other public models.
Why it matters
- Fans deserve to know these numbers were tested, not invented for the UI.
- It also matters what the tests did not show, so outside comparisons aren't read as claims.
How to interpret it
- Reserved testing holds out data the model did not tune against.
- Out-of-time testing checks whether earlier seasons help predict later ones.
- Those tests shaped the published method. They are not settings on the public boards.
How DRBL uses it
- Published DRBL/100 uses the small-sample adjustment that held up in testing.
- Product boards do not claim DRBL beats DARKO, BPM, or other externals.