Group A · 11 June 2026 · home venue — Mexico

Mexico v South Africa

Three views of the match

Mexico
Draw
South Africa

Modelbias-corrected

83%

16%

Sportsbook46 books, vig stripped

68%

21%

11%

Polymarket

69%

20%

10%

Market hedges more than the model

The model puts Mexico 15 pp higher than the market — 83% vs 68%.

The model assigns Mexico a 83% win probability while the sportsbook consensus stands at 68%—a 15-percentage-point gap favoring the model's confidence in a Mexico victory. The market simultaneously raises South Africa's chances to 11% versus the model's 1%, suggesting real-money flows are pricing in scenarios of upset or competitive balance that the model's Elo-based framework does not capture. This hedging could reflect recent lineup changes, injury concerns, or current form patterns unknown to the model at its last update. The dashboard flags this divergence as a meaningful signal: both sources agree Mexico are favored, but they disagree sharply on the margin of that advantage, indicating asymmetric information or risk assessment between the statistical forecast and live market pricing.

Preview

The model rates Mexico as a strong favorite, assigning an 83% win probability. South Africa's chances stand at less than 1%, leaving a wide gulf between the two sides in the statistical forecast.

The win-draw-loss spread reflects a heavily asymmetric contest. Mexico's dominant probability leaves little room for South Africa, though a draw at 16% is plausible as a secondary outcome. The match is modeled as decidedly uncompetitive in expectation, with the favorite's baseline advantage substantial.

The model and market diverge meaningfully here. The sportsbook consensus sits 15 percentage points lower on Mexico (68%) and substantially higher on a South Africa upset (11% versus under 1% from the model). Polymarket prices align closely with the books. The dashboard does not adjudicate between these sources; the gap suggests either the model overestimates Mexico's edge or the market prices in tail risk the Dixon-Coles framework does not capture.

Written by Claude from the figures above — it has no squad, injury, or form data. See the methodology.