A historical 90% rate is not a 90% guarantee

Nine successes in ten look like 90%, but the true uncertainty is much wider. If the same ten games qualify for five similar rules, that does not provide five independent confirmations.

Shrinkage and the lower bound

SUPER EDGE uses beta shrinkage and the Wilson lower bound: a small sample is pulled towards a cautious baseline probability. A strong historical pattern needs enough independent observations as well as a high percentage.

Historical frequency as supporting evidence

A historical rate becomes useful when context, period, team role and the live picture agree. It should not automatically outweigh a current change in pace, personnel or the score.

What the model checks

• Sample size N
• The Wilson lower bound
• How recent the games are
• Overlap between samples

Analysis describes probability and risk; it does not guarantee outcomes. This material is intended for readers aged 21+.