West Ham United vs Arsenal

West Ham's home corners average of 10.8 sits above the league baseline of 10.5, but that figure flattens a distribution that's far from stable. Three of their last five home matches settled at exactly 7 corners. Two spiked: 16 and 14. The model reads the predicted total at 9.95, pulling below the baseline, and that output aligns with the low-end cluster in West Ham's recent home record. The problem is that their high-end outcomes aren't noise, and the spike frequency isn't trivial.

Arsenal away compounds the picture. Their match total average of 9.4 rests on a similarly compressed base: three of the last five away fixtures produced exactly 7 corners, one reached 13, and one came in at 4. That range, from 4 to 13 in a single team's recent sample, signals the same distributional instability the West Ham data shows. Two volatile series running in parallel don't produce a high-precision forecast.

Medium confidence reflects that instability. The model's own fixture-level confidence is high, but that rating reflects model certainty in the predicted total, not stability of the raw input data. When two teams combine corner totals from 4 to 16 across their recent fixtures, a single-number prediction carries inherent error bars. A high model tag on a volatile fixture is not a prompt to treat it as high-confidence. Discipline holds.

Spectral's medium-confidence record in the Premier League: 83 settled, 45.8% win rate.