Model vs market
our call = 59% model + 41% market hereThe raw Elo+Poisson model can't see squad quality and is mis-scaled across confederations, so on lopsided matchups it disagrees sharply with a deep multi-book market. We publish a blend, and trust the model less the more it disagrees — so the model's share above is lower for the games where it strays furthest. The divergence tracker shows the raw gaps.
| 1X2 | South Africa | Draw | South Korea |
|---|---|---|---|
| Raw model | 22.9% | 27.6% | 49.6% |
| Market (de-vig) | 24.2% | 28.2% | 47.5% |
| Our call | 23.4% | 27.8% | 48.7% |
Asian markets
model probabilities · derived from the score matrixHandicaps from South Africa's side (−0.5 = must win outright; +0.5 = covers unless they lose). Half-lines can't push; whole lines can. No AH odds sourced yet, so these are probabilities, not edges.
| Handicap | South Africa | Push | South Korea |
|---|---|---|---|
| -1.5 | 8.5% | — | 91.5% |
| -1 | 8.5% | 14.4% | 77.1% |
| -0.5 | 22.9% | — | 77.1% |
| 0 | 22.9% | 27.6% | 49.6% |
| +0.5 | 50.4% | — | 49.6% |
| +1 | 50.4% | 23.3% | 26.2% |
| +1.5 | 73.8% | — | 26.2% |
| Total | Over | Under |
|---|---|---|
| 1.5 | 73.2% | 26.9% |
| 2.5 | 46.7% | 53.3% |
| 3.5 | 25.1% | 74.9% |
More markets
model probabilities · derived from the score matrixDouble chance, Draw-No-Bet and the likeliest exact scores, summed from the same Dixon-Coles scoreline distribution as the 1X2 numbers. No odds sourced for these, so they're probabilities, not edges.
| Double chance | Prob |
|---|---|
| South Africa or Draw | 50.4% |
| Draw or South Korea | 77.1% |
| South Africa or South Korea (no draw) | 72.5% |
| DNB — South Africa | 31.6% |
| DNB — South Korea | 68.4% |
| Likeliest score | Prob |
|---|---|
| 1–1 | 13.1% |
| 0–1 | 11.2% |
| 0–2 | 9.6% |
| 1–2 | 9.4% |
| 0–0 | 8.9% |
| 1–0 | 6.7% |
Bookmaker odds
line-shopping · 28 books · as ofAll 28 bookmakersline-shopping · best price marked ★
| Bookmaker | South Africa | Draw | South Korea | Over 2.5 | Under 2.5 |
|---|---|---|---|---|---|
| Best available | 4.10 (best price) | 3.45 (best price) | 2.08 (best price) | 2.20 (best price) | 1.70 (best price) |
| 888sport | 3.80 | 3.10 | 1.95 | — | — |
| Betclic (FR) | 3.78 | 3.30 | 1.91 | — | — |
| Betfred (UK) | 3.75 | 3.40 | 2.00 | — | — |
| BetOnline.ag | 4.00 | 3.42 | 2.03 | — | — |
| BoyleSports | 3.80 | 3.25 | 1.95 | — | — |
| Casumo | 3.50 | 3.25 | 2.08 (best price) | 2.17 | 1.61 |
| Coral | 3.75 | 3.40 | 2.00 | — | — |
| Everygame | 4.10 (best price) | 3.45 (best price) | 1.95 | — | — |
| Grosvenor | 3.50 | 3.25 | 2.08 (best price) | 2.20 (best price) | 1.62 |
| GTbets | 3.84 | 3.35 | 2.03 | 2.11 | 1.70 (best price) |
| Ladbrokes | 3.75 | 3.40 | 2.00 | — | — |
| LeoVegas | 3.50 | 3.25 | 2.08 (best price) | 2.17 | 1.61 |
| LeoVegas (SE) | 3.50 | 3.25 | 2.08 (best price) | — | — |
| LiveScore Bet | 3.50 | 3.25 | 2.08 (best price) | — | — |
| Marathon Bet | 3.84 | 3.35 | 1.95 | — | — |
| MyBookie.ag | 3.80 | 3.32 | 1.94 | 2.08 | 1.67 |
| Paddy Power | 3.80 | 3.30 | 2.00 | — | — |
| Pinnacle | 4.02 | 3.38 | 1.91 | — | — |
| PMU (FR) | 3.50 | 3.25 | 2.08 (best price) | 2.08 | 1.55 |
| Sky Bet | 3.75 | 3.30 | 2.00 | — | — |
| Smarkets | 3.95 | 3.10 | 1.88 | — | — |
| Unibet (NL) | 3.50 | 3.25 | 2.08 (best price) | 2.17 | 1.61 |
| Unibet (SE) | 3.50 | 3.25 | 2.08 (best price) | 2.17 | 1.61 |
| Unibet (UK) | 3.80 | 3.30 | 1.95 | — | — |
| Virgin Bet | 3.50 | 3.25 | 2.08 (best price) | — | — |
| William Hill | 3.80 | 3.10 | 1.95 | 2.10 | 1.67 |
| Winamax (DE) | 3.45 | 3.20 | 1.84 | — | — |
| Winamax (FR) | 3.30 | 3.05 | 1.80 | — | — |
Analysis
Group A finale with real qualification stakes — model rates Korea clearly, and the price agrees just enough to back it.
Group A, matchday 3 at Estadio BBVA in Monterrey. A final-round fixture that for both sides could be a play-in for the knockout lanes, with Mexico expected to have the group’s top tickets already.
South Africa (Elo 1604) are the lowest-rated side in the group and host here as underdogs. Athletic and quick in transition, but the finishing has been the recurring problem, and the model gives them only 18%.
South Korea (Elo 1716) are the class side on paper — Son-led, well-drilled, and 112 Elo points clear. The model trusts the gap: 59% to win, xG 1.72–0.83. If they need three points, they should be equipped to take them.
Model view. Korea 18% / 24% / 59% (home/draw/away), xG 0.83–1.72, BTTS 46%, Under 2.5 53%. A firm Korea lean.
Betting angle. The data flags Korea as the value side: 2.00 implies 50% but the model gives them 59% — an 8.7% edge. The price hasn’t fully respected the rating gap.
- South Korea ML at 2.00 — model 59%, fair ~1.69; the standout edge in this fixture
- Korea -0.5 AH = the straight win, 59%, same fair ~1.69
- Under 2.5 at 1.61 — model 53%, fair ~1.89, supported if South Africa sit deep
Avoid: South Africa ML on home-crowd logic (only 18%), over 2.5 at the implied price, BTTS yes short (model 46%).
This is a Korea-back spot: the model rates them well above the host and the 2.00 price still leaves an edge. Confirm Korea actually need the win — a side playing for a draw can blunt the ML value.
Verify before betting
- Both teams' points going in — this could be a straight knockout for second/third
- South Korea's Son and whether Korea need a win or can manage a draw
- Monterrey heat for an afternoon-ish kickoff
- South Africa's attacking personnel if they must chase the game
Team news
South Africa · South KoreaNo recent news mentioning South Africa or South Korea.
Bets on this match
No bets placed.