Model vs market
our call = 9% model + 91% 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 | Turkey | Draw | USA |
|---|---|---|---|
| Raw model | 67.1% | 21.5% | 11.4% |
| Market (de-vig) | 26.5% | 23.1% | 50.4% |
| Our call | 30.1% | 22.9% | 46.9% |
Asian markets
model probabilities · derived from the score matrixHandicaps from Turkey'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 | Turkey | Push | USA |
|---|---|---|---|
| -1.5 | 42.8% | — | 57.2% |
| -1 | 42.8% | 24.3% | 32.9% |
| -0.5 | 67.1% | — | 32.9% |
| 0 | 67.1% | 21.5% | 11.4% |
| +0.5 | 88.6% | — | 11.4% |
| +1 | 88.6% | 8.1% | 3.3% |
| +1.5 | 96.7% | — | 3.3% |
| Total | Over | Under |
|---|---|---|
| 1.5 | 76.6% | 23.4% |
| 2.5 | 51.5% | 48.5% |
| 3.5 | 29.4% | 70.6% |
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 |
|---|---|
| Turkey or Draw | 88.6% |
| Draw or USA | 32.9% |
| Turkey or USA (no draw) | 78.5% |
| DNB — Turkey | 85.5% |
| DNB — USA | 14.5% |
| Likeliest score | Prob |
|---|---|
| 2–0 | 13.2% |
| 1–0 | 12.3% |
| 1–1 | 10.2% |
| 2–1 | 9.5% |
| 3–0 | 8.9% |
| 0–0 | 7.3% |
Bookmaker odds
line-shopping · 40 books · as ofAll 40 bookmakersline-shopping · best price marked ★
| Bookmaker | Turkey | Draw | USA | Over 2.5 | Under 2.5 |
|---|---|---|---|---|---|
| Best available | 3.75 (best price) | 4.35 (best price) | 2.06 (best price) | 1.76 (best price) | 2.18 (best price) |
| 1xBet | 3.53 | 4.21 | 2.02 | 1.76 (best price) | 2.13 |
| 888sport | 3.30 | 3.80 | 1.85 | — | — |
| Bet Victor | 3.60 | 3.60 | 1.95 | — | — |
| BetAnything | 3.40 | 4.00 | 1.96 | — | — |
| Betclic (FR) | 3.45 | 4.05 | 1.90 | — | — |
| Betfair | 3.65 | 4.20 | 2.00 | — | — |
| Betfair | 3.65 | 4.20 | 2.00 | — | — |
| Betfred (UK) | 3.75 (best price) | 3.75 | 1.91 | — | — |
| BetOnline.ag | 3.56 | 4.35 (best price) | 1.91 | — | — |
| Betsson | 3.70 | 3.95 | 1.95 | 1.70 | 2.18 (best price) |
| Betway | 3.30 | 4.00 | 1.91 | — | — |
| BoyleSports | 3.60 | 3.75 | 1.87 | — | — |
| Casumo | 3.45 | 4.10 | 1.92 | 1.68 | 2.12 |
| Codere (IT) | 3.60 | 3.95 | 1.94 | 1.66 | 2.05 |
| Coolbet | 3.56 | 4.12 | 1.97 | 1.75 | 2.16 |
| Coral | 3.60 | 3.80 | 1.91 | — | — |
| Everygame | 3.40 | 4.10 | 1.95 | — | — |
| Grosvenor | 3.40 | 4.00 | 1.89 | 1.68 | 2.12 |
| GTbets | 3.48 | 3.98 | 1.99 | 1.73 | 2.15 |
| Ladbrokes | 3.60 | 3.80 | 1.91 | — | — |
| LeoVegas | 3.45 | 4.10 | 1.92 | 1.68 | 2.12 |
| LeoVegas (SE) | 3.45 | 4.10 | 1.92 | — | — |
| LiveScore Bet | 3.40 | 4.00 | 1.89 | — | — |
| Matchbook | 3.65 | 4.20 | 2.00 | — | — |
| MyBookie.ag | 3.55 | 3.90 | 1.91 | 1.68 | 2.07 |
| Nordic Bet | 3.65 | 3.90 | 1.98 | 1.75 | 2.10 |
| Paddy Power | 3.40 | 3.80 | 1.91 | — | — |
| Pinnacle | 3.44 | 4.08 | 1.98 | — | — |
| PMU (FR) | 3.40 | 4.00 | 1.88 | 1.65 | 2.08 |
| Sky Bet | 3.40 | 3.90 | 1.91 | — | — |
| Smarkets | 3.70 | 3.95 | 2.06 (best price) | — | — |
| Tipico | 3.60 | 4.00 | 1.90 | — | — |
| Unibet (FR) | 3.45 | 3.80 | 1.89 | — | — |
| Unibet (NL) | 3.55 | 4.20 | 1.94 | 1.68 | 2.12 |
| Unibet (SE) | 3.55 | 4.20 | 1.94 | 1.71 | 2.16 |
| Unibet (UK) | 3.50 | 3.80 | 1.91 | — | — |
| Virgin Bet | 3.40 | 4.00 | 1.89 | — | — |
| William Hill | 3.60 | 3.90 | 1.83 | 1.65 | 2.15 |
| Winamax (DE) | 3.30 | 4.20 | 1.90 | — | — |
| Winamax (FR) | 3.25 | 4.10 | 1.88 | — | — |
Analysis
Group D finale with top-spot stakes — model edges Turkey and the price leaves a thin edge on the home win.
Group D, matchday 3 at SoFi Stadium in Inglewood. Two of the group’s stronger sides meet in a fixture that could settle top spot and the knockout seeding.
Turkey (Elo 1691) are rated marginally above the USA and the model makes them the favourite at 44%. The attack is the strength; if it clicks, they’re the side more likely to take the game by the scruff.
USA (Elo 1654) are 37 Elo points back and may be content to manage the game depending on standings. The model gives them 30% to win and a real share of the draw equity.
Model view. Turkey 44% / 26% / 30%, xG 1.42–1.13, BTTS 51%, Under 2.5 53%. Turkey-leaning but competitive.
Model vs market. The data flags a modest edge on Turkey: 2.70 implies 37% but the model gives 44% — a 6.7% edge. Real but not large.
- Turkey ML at 2.70 — model 44%, fair ~2.27; the flagged value
- Draw at 3.30 — model 26%, fair ~3.85, live if the USA play for a point
- BTTS yes ~1.85 — model 51%, two functional attacks
Avoid: USA ML at a short price if they only need a draw, over 2.5 against a 53% under read, Turkey -1 chasing a margin.
Turkey are the model’s pick and the 2.70 price leaves a thin edge — but standings will dictate the USA’s approach. If the USA are playing for a draw, lean the draw and BTTS rather than the Turkey ML.
Worth checking before you lean on this
- Group D standings — this could decide the group winner
- USA's qualification scenario and whether a draw suffices
- Turkey's attacking shape vs a US side that may sit deeper
- Inglewood environment and any rotation if already qualified