WC2026 Model Tracker
Bivariate-Poisson predictions vs actual results ยท a live, honest experiment

๐Ÿ‡ณ๐Ÿ‡ฑ Netherlands vs Japan  WINNER MISS

FIFA World Cup 2026, Group F, Matchday 1 ยท AT&T Stadium, Arlington TX (indoor, sea level, neutral site)
Kickoff Jun 14, 2026 ยท 20:00 UTC
Final ยท Netherlands 2โ€“2 Japan
Goals: 51' Virgil van Dijk (Netherlands); 57' Keito Nakamura (Japan); 64' Crysencio Summerville (Netherlands); 89' Daichi Kamada (Japan)
Actual stats: xG 0.78 โ€“ 0.59 ยท 9 corners ยท 3 cards (0 red)

Prediction vs result  2/8 markets hit

MarketModel calledActualOutcome
1X2JapanDraw (2-2)MISS
Score1-12-2MISS
O/U 2.5Under 2.5 (58.8%)4 goalsMISS
BTTSNo (51.2%)YesMISS
CornersOver 8.5 (50.2%)9 cornersHIT
CardsOver 3.5 (52.4%)3 cardsMISS
Red cardNo (89.1%)NoHIT
AHNetherlands +0.00PushPUSH

Model vs market

MarketModelBook oddsBook p*EV
Netherlands win32.2%2.0047.3%-35.5%
Draw29.8%3.5027.1%+4.4%
Japan win37.9%3.7025.6%+40.4%
Over 2.5 goals41.2%2.0047.4%-17.5%
Under 2.5 goals58.8%1.8052.6%+5.8%
Both teams score: yes48.8%1.8052.0%-12.1%
Both teams score: no51.2%1.9548.0%-0.2%
Over 8.5 corners (exp 8.69)50.2%โ€”โ€”โ€”
Under 8.5 corners49.8%โ€”โ€”โ€”
Over 3.5 cards (exp 3.79)52.4%โ€”โ€”โ€”
Under 3.5 cards47.6%โ€”โ€”โ€”
Any red card10.9%โ€”โ€”โ€”
Model-only estimates (no straight market equivalent): expected goals Netherlands 1.12 โ€“ 1.23 Japan ยท fair Asian handicap Netherlands +0.00
In the table above, Model = our probability and Book odds / Book p* = the bookmaker's (p* has the margin removed). EV = value per unit staked at the model's probability (positive = model sees value). โ€œโ€”โ€ = market not in the captured feed.
Model reasoning (chain of thought)
PREDICTION RATIONALE โ€” Netherlands vs Japan  (FIFA World Cup 2026, Group F, Matchday 1)
Venue: AT&T Stadium, Arlington TX (indoor, sea level, neutral site)

1. DATA. Last-10 internationals (Sofascore interception scraper).
   Netherlands: record DWWWDWWDLW, 2.20 GF / 0.80 GA,
   xG 1.99 / xGA 0.74.
   Japan: record WDLDWWWWWW, 1.40 GF / 0.60 GA,
   xG 1.33 / xGA 1.4.
   Opponent-strength tiers (configs.MASTER_TIERS) applied per match.

2. RATINGS. Netherlands atk 1.14 def 0.84; Japan atk 1.12 def 0.73.
   NEUTRAL VENUE: home boost / away penalty set to 1.0/1.0 (both teams
   travel to a host-nation site; no host advantage applies โ€” a key difference
   from the Korea/Canada/USA host-nation games). lambda Netherlands 1.12, Japan 1.23
   after 85% model weight + shrink to 1.3 baseline.
   Injuries: {'Netherlands': ['No major absences reported'], 'Japan': ['No major absences reported']}.
   Context: Group F opener. Netherlands deeper/favoured; Japan a strong, well-organised side (beat Germany & Spain in 2022). Neutral venue, both travelling.

3. MODEL. Bivariate Poisson (lambda3 0.12). 1X2 Netherlands 32.2% /
   draw 29.8% / Japan 37.9%. Winner Japan; score 1-1; fair AH Netherlands +0.00.
   Over 2.5 41.2%; BTTS 48.8%.

4. EVENTS. Referee Ismail Elfath (USA): 3.9 Y + 0.1 R/match
   blended (weight 0.47) with WC base 3.6027 -> expected
   cards 3.79, P(over 3.5) 52.4%; P(any red)
   10.9%. Corners expected 8.7 (scale
   1.0341) -> P(over 8.5) 50.2%.

5. LIMITATIONS. 10-match samples; hand-set tiers/priors; Poisson corners; lineups
   probable not confirmed; card base rate calibrated on only 4 WC matches.