What we announced, put against the result — the denied ones included.
The gap between the stated probability and the observed hit rate is the honest measure of calibration. A hit rate above the stated probability on a small sample is luck, not skill.
No analysis in this selection
Widen the filters to see the full resolved sample.
Every row is an immutable snapshot taken before kick-off. Analyses are never rewritten after the fact, and a disproved analysis is kept in the record exactly like a confirmed one.