AI predicted 2–1 — it finished 0–1 ✗

Unified Prediction Synthesis

The six expert forecasts cluster heavily on a Turkey win, with five models selecting home_win (scores of 2-1 or 1-0) and only GLM-5.2 diverging to a 1-1 draw. Historical accuracy metrics show Kimi, Grok, Gemini, and GPT clustered around 42-52% confidence on the home side with solid or positive ROI trends, while GLM's high accuracy is discounted due to its minuscule sample size. Claude's low 18% confidence on 1-0 further tempers over-optimism. The synthesis calibrates Turkey's edge at 48% probability after resolving disagreements through evidence strength rather than volume.

Core Agreement on Asymmetries

Experts converge on Turkey's superior underlying performance versus Australia (30 shots, 2+ xG, 72% possession) as a finishing outlier, not structural failure, versus Paraguay's genuine defensive collapse (4-1 loss, 3-0 at halftime to USA). Yildiz's fitness confirmation transforms Turkey's attack. Must-win stakes for both sides in Group D make cagey draws unlikely, favoring open, high-event play where Turkey's technical depth prevails.

Resolving the Draw Outlier

GLM's low-block caution argument ignores how Paraguay's compact system already failed under pressure and how Turkey's creators exploit transitions. Market odds (draw at 2.10 implying ~47%) contradict sharp line movement drifting outward and expert xG reads; the draw is structurally overpriced.