Arsenal vs Newcastle United

The model projects 10.12 total corners for this fixture, sitting fractionally below the Premier League baseline of 10.6. That gap is narrow. The key tension here is between Arsenal's home-match average of 8.9, which pulls the projection down, and Newcastle's away-match average of 10.6, which matches the league baseline exactly and pulls in the opposite direction. Neither team's recent profile is dominant enough to break the stalemate, and that equilibrium is part of what keeps this in medium-confidence territory.

Newcastle's last five away corner totals read 9, 6, 9, 10, 14. The spread is wide. Arsenal's home five, 9, 11, 11, 15, 7, tell a similar story: one fixture produced 15 corners, another only 7. The distance between 15 and 7 in a five-match sample doesn't point to a settled pattern. Both sequences carry the kind of variance that widens the model's probability distribution and reduces confidence in a precise total, which is precisely what the medium tier captures.

Medium confidence here reflects genuine structural ambiguity. Where the projected total lands close to the league baseline and both teams' recent corner sequences are noisy, the model doesn't manufacture certainty that the data can't support. The medium tier carries more distributional width than high-confidence calls by definition. Apply medium-tier stake discipline. Don't size up on the strength of a near-baseline projection.

Spectral's medium-confidence record in the Premier League: 40 settled, 42.5% win rate.