Mansfield Town vs Oxford United

The model projects 9.58 total corners at Field Mill, placing the estimate below the League One baseline of 10.7. The data point one way. Mansfield's home matches average 8.8, with their recent games producing 8, 8, 11, and 8 corners, a sequence that's tight and consistently below the league mean. Oxford's road fixtures average 10.2, but their recent away games produced 9, 15, 5, and 12 corners, a distribution wide enough that the central estimate carries genuine uncertainty regardless of its direction.

That spread in Oxford's recent away record is the primary reason confidence sits at medium. The 15-corner fixture drags their road average upward, while the 5-corner game sits well below it. Precision suffers. When the away team's data is this dispersed, the model's output retains directional coherence but loses precision, and it's that precision shortfall that keeps this away from a higher confidence tier. Mansfield's home sequence is far more stable, which is useful context, but it doesn't resolve the noise in Oxford's numbers.

Stake discipline applies here. Medium confidence reflects a directional signal without the data coherence that would warrant high-confidence tier treatment, and the discipline to honour that distinction consistently is where the process holds together over time. The model isn't working from a clean data set on both sides.

Spectral's medium-confidence record in League One: 48.5% from 101 settled picks.