Hamburger SV vs Hoffenheim

Variance defines this matchup. Hoffenheim's last five away corner totals, 13, 8, 4, 7, and 13, make any projection less reliable than the team average alone suggests, because that spread is too wide to compress into a single trend. HSV's home figures tell a similar story: three of their last five matches produced 9, 9, and 11 corners, while two hit 13. The Bundesliga baseline sits at 9.6 per match, and the model's predicted total of 8.63 falls below it.

That gap is the key tension. HSV's home match average is 9.8; Hoffenheim's away figure is 10.3. Both sit above the predicted total. The model runs on more than rolling averages, and the medium confidence rating reflects genuine disagreement between those inputs and the broader feature set. We don't paper over that disagreement.

What keeps this from a high-confidence read is the distribution width. When a team's last five away totals include both 4 and 13, the range accommodates almost any outcome. It's a data signal. Medium confidence exists precisely for matchups where the model outputs a signal but the noise floor remains high enough to matter.

Size it accordingly. Proportionate sizing means treating this read as what the model flags it: medium confidence, with variance wide enough that outlier outcomes remain plausible across the full range of recent team data. Spectral's medium-confidence tier has settled 39 picks at a 53.8% win rate in the Bundesliga.