AI predicted 1–1 — it finished 2–1 ✗

Unified Prediction: 1-1 Draw

The experts converge strongly on a low-scoring draw as the single most likely outcome, with four of five models selecting 1-1. GLM-5.2 (highest historical accuracy at 70% result rate and +7% ROI) leads with a 42% draw probability, emphasizing tactical neutralization and group-stage caution. Grok 4.3 (53% confidence) and GPT-5.4 (39%) reinforce this via Canada's qualification math and Koné's absence, while Claude Opus 4.6 stresses the asymmetric incentives despite low overall confidence. Kimi K2.6 diverges toward a narrow Swiss win, citing tournament pedigree, but receives lower weight due to weaker recent accuracy (47% result rate, negative ROI).

How the Match Unfolds

Switzerland control possession through Xhaka and Freuler but face a compact, motivated Canadian side unwilling to chase goals. Canada's high press is blunted by Swiss composure, while Canada's transitions (Davies, David) are contained by organized defending. The second half grows cagey as both prioritize advancement over risk. Koné's injury forces Canada into deeper midfield shape, reducing their ability to disrupt or counter effectively. Set pieces or a single moment of quality decide a tight contest rather than open play.

Resolution of Disagreements

Kimi's Swiss lean is discounted relative to GLM's stronger track record and the structural evidence from other models. The group-decider context (Canada +6 GD, needs only a point) overrides raw quality edges, pushing probability toward draw over the 32-35% Swiss win range some models imply. This yields a calibrated 48% draw probability — honest given base-model overconfidence and market efficiency.