Germany vs. Paraguay — Consensus Ensemble
2026 World Cup, Round of 32 (Match #74) | Monday June 29, 2026 | Gillette Stadium, Boston
Kickoff: 20:30 GMT (22:30 CEST) | Referee: Jalal Jayed (Morocco)
TL;DR
Germany favored to advance (~74%) but Paraguay's low block + Gill hot-GK form cap the margin: GER 55% win in 90' / Draw 23% / PAR 22%. Modal scoreline GER 1-0 (18%). Undav (unanimous #1 scorer, ~22%) returns the favor after 3 goals off bench; Almirón (~10%) is PAR's lone counter threat. Musiala benched = 60' weapon. Not a blowout — F22 exclusion fires.
- 3-model spread: 13pt on GER win (54-67%) — wider than BRA-JPN. g5 most bullish (67%, 80% advance), mm3/g4 more cautious (54-55%, 70-72%). The disagreement: how much does PAR's low block + Gill hold? g5 says GER breaks through; mm3/g4 see a tighter, late-goal 1-0/1-1 archetype.
- Draw at 90' low (15-25%) — clear-favorite KO with non-veteran-GK block (calibrated down from 26% per BRA-JPN learning). PAR's path to a draw is Gill heroics + sterile GER possession.
- GER tactical shift (3-4-2-1, Musiala benched) is the biggest model-breaker — Nagelsmann reacting to low-block struggles. Undav starts (aerial #9), Musiala = 60' weapon.
- PAR creativity gutted (Diego Gomez suspended, Alderete bench) — Enciso sole creator. But single absence = -3 to -5pt (NOT F4a gutting). Almirón returns as counter talisman.
- Weather 28°C at threshold (open-air Gillette) — borderline F11, does NOT fully fire.
- Watch: Musiala 60' entry (swing moment), Gill MOTM form (draw-flipper), Tah-Rüdiger chemistry (zero WC minutes together).
Decision Tree (§7K Knockout Format)
| Phase | GER | Draw/ET | PAR |
|---|---|---|---|
| 90' Regulation | 55% | 23% | 22% |
| Extra Time (conditional on draw) | 50% | 30% (→pens) | 20% |
| Penalties (conditional) | 55% | — | 45% |
| Advance | 74% | — | 26% |
Modal Match Path
GER 1-0 in 90' (18%) — Undav or Wirtz breaks the low block; Gill keeps PAR close but can't force ET. If level at 90', GER favored in ET (Musiala/Goretzka bench depth).
WDM Anchor (shared starting point)
| Factor | Weight | GER share | Notes |
|---|---|---|---|
| Squad value | 3.0 | 0.89 (€775M vs €100M) | Dominant |
| FIFA rank | 1.0 | 0.52 (GER #11 vs PAR #41) | |
| Tournament form | 1.0 | 0.61 (GER 2.0 PPG vs PAR 1.33 PPG) | |
| Qualifying | 1.0 | 0.55 (GER 2.2 vs PAR 1.8 PPG) |
WDM anchor: GER 49% / Draw 26% / PAR 25% → writers adjust to ~55/23/22.
Per-Model Predictions
| Agent | GER 90' | Draw | PAR 90' | GER Advance | Modal Scoreline |
|---|---|---|---|---|---|
| mm3 | 54% | 25% | 21% | 72% | 1-1 → GER 2-1 ET |
| g4 | 55% | 24% | 21% | 70% | GER 1-0 |
| g5 | 67% | 19% | 14% | 80% | GER 2-0 |
| Consensus | 55% | 23% | 22% | 74% | GER 1-0 |
g5 is the outlier (+12pt on GER). The disagreement is the sterile-domination screen: g5 says GER breaks through with Undav+Musiala-60'; mm3/g4 see PAR's block holding longer. Flagged but not dropped — the spread signals genuine uncertainty about whether Nagelsmann's tactical fix works.
Who Scores
GER
| Rank | Player | Probability |
|---|---|---|
| 1 | Deniz Undav | ~22% |
| 2 | Florian Wirtz | ~18% |
| 3 | Kai Havertz | ~15% |
| 4 | Jamal Musiala (sub) | ~10% |
| 5 | Leroy Sané | ~8% |
PAR
| Rank | Player | Probability |
|---|---|---|
| 1 | Miguel Almirón | ~10% |
| 2 | Julio Enciso | ~9% |
| 3 | Gabriel Ávalos | ~7% |
| 4 | Antonio Sanabria (sub) | ~5% |
| 5 | Ramon Sosa (sub) | ~4% |
Referee & Cards
| Referee | Jalal Jayed (MAR, 39, CAF) |
| Career Y/gm | 3.84 → tournament ~2.8-3.2 |
| Expected yellows | 3-4 (PAR has 3+ players on cumulative yellows) |
| Reds | 0 likely (0.099 career) |
| Penalty | 25-35% band (leans GER if awarded) |
| Confederation | CAF for UEFA vs CONMEBOL = neutral |
Predicted Lineups (confirmed)
GER (3-4-2-1)
Neuer; Kimmich, Tah, Rüdiger; Sané, Pavlović, Nmecha, N.Brown; Wirtz, Havertz; Undav
Bench: Musiala, Goretzka, Raum, Leweling, Stiller, Anton, Thiaw, Woltemade, Nübel
PAR (4-4-2)
Gil; Cáceres, Gómez, Alonso, Canale; Cubas, Bobadilla, Almirón, Galarza; Enciso, Ávalos
Bench: Gatito Fernández, Sanabria, Sosa, Arce, Alderete, Pitta, Velázquez
Consensus ensemble — 3 independent agents (mm3, g4, g5). Spread: 13pt (g5 outlier flagged). WDM anchor applied. GER favored but PAR's low block + Gill form keep this honest.