Brighton and Hove Albion vs Crystal Palace

Brighton's recent home data provides genuine support for an elevated projection. Their last four home fixtures produced match totals of 13, 11, 13, 7, giving a four-match average of 11.0 that sits above both the Premier League baseline of 10.4 and Crystal Palace's recent away figure of 8.0. The model outputs a predicted total of 11.75. That's not trivial. Medium confidence signals real tension in the underlying data, however, and that framing matters as much as the projection itself.

Crystal Palace are the source of that tension. Their four-match away average of 8.0 corners per game sits well below both Brighton's figure and the league baseline, and the distribution sharpens the concern: the individual match totals read 13, 7, 8, 4. That spread is wide. A Palace side producing 13 corners in one fixture and 4 in another introduces distributional variance that prevents a tight projection, and the medium confidence tier reflects that directly.

Position sizing should track the confidence tier, not the magnitude of the predicted total. Size down. The staking framework is built to absorb exactly this kind of asymmetry: one team carrying broad distributional spread across recent away fixtures, the other producing consistent high counts at home, with the model outputting a predicted total above baseline but not at the level of certainty that justifies full sizing.

Across 150 settled medium-confidence picks in the Premier League, Spectral's win rate stands at 47.3%.