Wednesday closes the group stage, and the model and the board are at war over one number: Bosnia. The market has Bosnia at 72¢ to beat a near-eliminated Qatar — but the 10,000-run model reads a 41% coin flip, the widest model-vs-board gap on the card. That split is the theme of the whole slate: across Groups A, B and C, the board is paying up for favorites in dead rubbers, and the draws are the value. The one favorite the model actually wants to back is Mexico at the Azteca. Here's the full card.
The Wednesday card
| Game (ET) | The position | The value angle |
|---|---|---|
| Switzerland vs. Canada · 3:00 PM | Pass / slight Canada lean | Model 40/28/32 ≈ board 40/32/30 — roughly fair |
| Bosnia vs. Qatar · 3:00 PM | Fade Bosnia — Qatar or the draw | THE EDGE: model 41% vs the board's 72¢ |
| Scotland vs. Brazil · 6:00 PM | Brazil, but the draw's live | Showcase; model 71% vs 76¢, draw 22% vs 18¢ |
| Morocco vs. Haiti · 6:00 PM | Take the draw / don't lay Morocco | Model 72% vs an inflated 84¢ on the favorite |
| Czechia vs. Mexico · 9:00 PM | Mexico to win | Underpriced: model 65% vs 53¢ at the Azteca |
| South Africa vs. Korea Republic · 9:00 PM | The draw | Model draw 31% vs 25¢; Korea rich at 61¢ |
Where the market is wrong
The disagreement this Wednesday isn't one game — it's a pattern. Five of the six finals feature a side that's already qualified or already eliminated, and the board keeps pricing the bigger name like the stakes are normal. The model keeps saying they aren't. Bosnia–Qatar is the centerpiece — a 72¢ favorite the model makes a 41% coin flip — but the same logic flags Morocco (72% model against an 84¢ price) and Korea (51% against 61¢) as overpriced chalk where the draw is the value the board won't pay for. The exception, and the one place the model wants the favorite, is Mexico: 65% in the simulation against a soft 53¢, because the Azteca is worth a goal the market is discounting. Switzerland–Canada is the one game where model and board essentially agree. Read the slate as a whole: back Mexico, take the draws, and fade the names in the games nobody has to win.
Switzerland vs. Canada — the one game priced right
The closest match to fair on the board. Both sides arrive on four points with a place in the round of 32 all but secured; this decides who tops Group B, and Canada hold the tiebreaker — a +6 goal difference, anchored by a 6-0 of Qatar, to Switzerland's +3, so Canada finish first with a win or a draw at home in Vancouver. The model and the market shake hands here: the simulation reads Switzerland 40%, the draw 28% and Canada 32%, while the board sits 40¢ / 32¢ / 30¢. The only nudge is the draw — the model has it three points under the market, so if you're playing the game, the two-way (Switzerland or Canada to win) is the marginally better side of a near-fair number. There's no standout here; the edges are elsewhere.
Model: Switzerland 40% / draw 28% / Canada 32% · Board (Kalshi): Switzerland 40¢ / draw 32¢ / Canada 30¢.
Bosnia vs. Qatar — fade the name in a dead rubber
The marquee read, because it's where the model and the board split hardest all week. Bosnia and Qatar are both all but out — Bosnia at -3 goal difference after a 4-1 loss to Switzerland, Qatar at -6 after a 6-0 from Canada — so this is a pride game with nothing riding on it, and that's exactly the context the board misreads. The market prices Bosnia like the better roster always wins: 72¢, with Qatar a throwaway 12¢ and the draw 19¢. The model, stripped of the reputation and looking at a flat, low-stakes game, reads a near three-way coin flip: Bosnia 41% / draw 30% / Qatar 29%. That's a 31-point gap on the favorite — the widest on the card. The position is to fade the 72¢: the cleanest expression is Qatar or the draw (double chance), which the board implies near 31¢ against a model close to 59%, and the draw alone at 19¢ is the single standout, a 30% model outcome trading at well under its number. This isn't a play on Qatar's quality; it's a play on a price that assumes a desperation that isn't there.
Model: Bosnia 41% / draw 30% / Qatar 29% · Board (Kalshi): Bosnia 72¢ / draw 19¢ / Qatar 12¢.
Scotland vs. Brazil — the showcase, and a live draw
The marquee name on the slate and a survival game for the underdog. Scotland sit on three points and need a result against Brazil to be sure of advancing, while Brazil — four points, a 1-1 with Morocco and a 3-0 of Haiti behind a Matheus Cunha brace — need only a draw to top Group C. The model and board mostly agree on the favorite: Brazil 71% in the simulation against 76¢ on the board, so there's no edge backing or laying the Brazil number — if anything it's a touch rich. The small disagreement is the draw: the model has it 22% against 18¢, nudged up by a Brazil side that may be content to manage a result it's already qualified with and a Scotland team that has to chase. The position is Brazil to win if you want the side, but the live value is the draw at 18¢, and the cleanest non-result play is the goals — a desperate Scotland opening the game is the script Brazil's front line punishes, so Vinícius Júnior or Cunha anytime is the focal scorer.
Model: Scotland 7% / draw 22% / Brazil 71% · Board (Kalshi): Scotland 10¢ / draw 18¢ / Brazil 76¢.
Morocco vs. Haiti — don't lay the inflated chalk
The most lopsided game on the slate, and the model says the board has overcooked it. Morocco are through-adjacent on four points — a 1-1 with Brazil and a 1-0 of Scotland — and host Haiti, already eliminated as the tournament's first side out. The market lays the obvious chalk: Morocco 84¢. But the simulation only makes them 72%, with the draw a live 23% against a 13¢ price — the model's read on a qualified favorite that may rotate and an eliminated opponent with nothing to lose and no pressure to chase. That's a 12-point gap on the favorite, and it kills the case for laying Morocco at -1.5 or a short moneyline. The position is the draw at 13¢ as the model's value, or simply a pass on the Morocco number; if you want Morocco's goals, take them as a scorer prop (Youssef En-Nesyri anytime), not a steep line the model won't back.
Model: Morocco 72% / draw 23% / Haiti 6% · Board (Kalshi): Morocco 84¢ / draw 13¢ / Haiti 6¢.
Czechia vs. Mexico — back the favorite the board is discounting
The one game where the model wants the chalk, because the board is selling it short. Mexico are already through as Group A winners after a 2-0 of South Africa and a 1-0 of South Korea, and they close at the Estadio Azteca, where the altitude and the crowd are worth a goal the market keeps discounting. Czechia sit on a single point and must win to have any chance — a tall order in Mexico City. The board prices Mexico at a soft 53¢; the simulation reads 65%, with Czechia a generous 25¢ against an 11% model. That's a 12-point gap on the favorite and a 14-point gap fading Czechia — the biggest two-way edge on the slate after the Bosnia number. The position is Mexico to win, and if you want a price off it, Santiago Giménez or Raúl Jiménez anytime is the focal scorer in front of a home crowd. The one caveat keeps the draw honest — a qualified Mexico could rotate — but the model has already priced that in and still lands at 65%.
Model: Czechia 11% / draw 24% / Mexico 65% · Board (Kalshi): Czechia 25¢ / draw 24¢ / Mexico 53¢.
South Africa vs. Korea Republic — the draw is the value
The other Group A finale, and another favorite the model trims. Korea Republic sit on three points after beating Czechia and a narrow loss to Mexico, and a draw very likely sends them through; South Africa are on one point and need a win and help. The board makes Korea a 61¢ favorite, but the simulation reads 51% — a more talented side that may be content to manage the point that qualifies it, against a Bafana attack that has produced one goal in two games. The model's value is the draw at 25¢ against a 31% read, the cleanest expression of a low-event game both sides can live with. There's no edge laying Korea at 61¢; if you want a side, South Africa or the draw (double chance) is the model-backed way onto the underdog without taking a near-must-win at face value.
Model: South Africa 17% / draw 31% / Korea 51% · Board (Kalshi): South Africa 16¢ / draw 25¢ / Korea 61¢.
Sizing the card
Same discipline The 7 Oracles ran all week: the Bosnia number is the one to lead with — a 31-point model-vs-board gap doesn't show up often — so let Kelly size the Qatar-or-draw ahead of the Mexico read and the Morocco draw. The through-line is simple: the board is paying full freight for favorites in games nobody has to win, and the model keeps pricing the draws and the dogs higher than the market does. Mexico is the one exception worth a favorite's price. If you string any of it, the Combo Builder shows the real combined number before you commit — and remember these are pride-game and dead-rubber prices that move fast once lineups drop, so confirm the live board before you take a position. Size to the edge, lean to the value.
Track the live model and Kalshi reference prices on the World Cup hub, the Sports · World Cup 2026 hub, the group standings, and the odds page. Today's headline position lives on Picks Today. Model figures are 10,000-run simulation outputs; market prices move — confirm the live board before taking a position. PredictionMarketsPicks publishes market analysis, not trading advice. Trade responsibly.