Brighton and Hove Albion vs Manchester United

The model lands at 10.55 predicted corners, a fraction above the Premier League baseline of 10.5. Proximity to baseline is telling. When a projection sits this close to the league average, there is little structural reason to depart from prior, and that absence of differentiation drives a medium-confidence assessment here. The model registers high confidence on the predicted figure itself; that confidence applies to the prediction, not to how much it departs from the league baseline.

Both teams carry a 9.5 match-total average. Variance is significant. Brighton's last five corner totals ran 3, 8, 11, 12, 7; Manchester United's read 3, 13, 8, 14, 6. Both distributions contain a 3-corner floor; peaks reach 12 for Brighton and 14 for Manchester United. The model's projected 10.55 sits above both teams' match-total averages of 9.5, leaning on the league baseline to justify the uplift rather than strong team-level signal from either side. That gap between team averages and projected total introduces genuine uncertainty: the numbers don't rule the projection out, but they don't confirm it either.

Stake discipline is appropriate. The structural case is modest, and nothing in the last five fixtures from either side shifts that assessment materially. Medium confidence reflects a wider distribution of plausible outcomes, not an ambiguous directional signal, and it's the appropriate classification when a projection lands this close to baseline.

Spectral's Premier League medium-confidence corners record: 134 settled, 45.5% win rate.