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

πŸ‡ΊπŸ‡Έ USA vs Australia πŸ‡¦πŸ‡Ί  WINNER HIT

FIFA World Cup 2026, Group D, Matchday 3 Β· USA (HOST β€” home advantage applies)
Kickoff Jun 19, 2026 Β· 19:00 UTC
Final Β· USA 2–0 Australia
Goals: 11' Cameron Burgess (USA); 43' Alexander Freeman (USA)
Actual stats: xG 1.08 – 0.35 Β· 11 corners Β· 7 cards (0 red)

Prediction vs result  3/8 markets hit

MarketModel calledActualOutcome
1X2USAUSA (2-0)HIT
Score1-12-0MISS
O/U 2.5Over 2.5 (50.8%)2 goalsMISS
BTTSYes (53.9%)NoMISS
CornersUnder 8.5 (53.6%)11 cornersMISS
CardsUnder 3.5 (51.8%)7 cardsMISS
Red cardNo (87.8%)NoHIT
AHUSA -0.50Full WinHIT

Model vs market

MarketModelBook oddsBook p*EV
USA win52.0%1.6059.2%-16.9%
Draw25.4%4.3321.9%+10.2%
Australia win22.6%5.0018.9%+13.1%
Over 2.5 goals50.8%1.9150.0%-3.0%
Under 2.5 goals49.2%1.9150.0%-6.1%
Both teams score: yes53.9%1.9548.0%+5.1%
Both teams score: no46.1%1.8052.0%-17.0%
Over 8.5 corners (exp 8.41)46.4%β€”β€”β€”
Under 8.5 corners53.6%β€”β€”β€”
Over 3.5 cards (exp 3.59)48.2%2.2541.1%+8.5%
Under 3.5 cards51.8%1.5758.9%-18.6%
Any red card12.2%β€”β€”β€”
Model-only estimates (no straight market equivalent): expected goals USA 1.68 – 1.05 Australia Β· fair Asian handicap USA -0.50
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). β€œβ€”β€ = market not in the captured feed.
Model reasoning (chain of thought)
PREDICTION RATIONALE β€” USA vs Australia  (FIFA World Cup 2026, Group D, Matchday 3)
Venue: USA (HOST β€” home advantage applies)

1. DATA. Last-10 internationals (Sofascore interception scraper).
   USA: record WDWWWLLWLW, 2.20 GF / 1.60 GA,
   xG 1.7 / xGA 1.26.
   Australia: record WWLLLWWLDW, 1.40 GF / 1.00 GA,
   xG 0.97 / xGA 1.21.
   Opponent-strength tiers (configs.MASTER_TIERS) applied per match.

2. RATINGS. USA atk 1.46 def 0.94; Australia atk 0.87 def 0.90.
   NEUTRAL VENUE: home boost / away penalty set to 1.1219/0.8648 (both teams
   travel to a host-nation site; no host advantage applies β€” a key difference
   from the Korea/Canada/USA host-nation games). lambda USA 1.68, Australia 1.06
   after 85% model weight + shrink to 1.3 baseline.
   Injuries: {'USA': ['No major absences reported'], 'Australia': ['No major absences reported']}.
   Context: Group D MD3. USA are co-HOSTS β€” home advantage applies. MD3 caveat: possible rotation/dead-rubber if standings are set (model doesn't see group context).

3. MODEL. Bivariate Poisson (lambda3 0.12). 1X2 USA 52.0% /
   draw 25.4% / Australia 22.6%. Winner USA; score 1-1; fair AH USA -0.50.
   Over 2.5 50.8%; BTTS 53.9%.

4. EVENTS. Referee Felix Zwayer (Germany): 4.63 Y + 0.13 R/match
   blended (weight 0.31) with WC base 3.0543 -> expected
   cards 3.59, P(over 3.5) 48.2%; P(any red)
   12.2%. Corners expected 8.4 (scale
   0.9834) -> P(over 8.5) 46.4%.

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