Portugal vs. Uzbekistan - Consensus Ensemble
2026 World Cup, Group K MD2 | June 23, 2026 | NRG Stadium, Houston (roof CLOSED)
⚠️ UZB MUST-WIN (0pts, eliminated with loss)
TL;DR
Portugal ~75% / Draw ~19% / Uzbekistan ~5%. Modal scoreline: POR 2-0 (consensus #1, ~17%). Ronaldo #1 scorer (~22%, drought + milestone + pen duty). The headline is the REVERSED GK dynamic: UZB benched undisputed #1 Yusupov (44 caps) → 3rd-choice Nematov (14 caps, WC debut) = Finding 14 model-breaking, FLIPS the POR-DRC draw mechanism (hot underdog GK → weak underdog GK). Finding 17 (sterile-domination) fires as a RISK but mitigated by Nematov weakness + Bernardo→Félix mobility swap. Houston roof closed → Finding 11 neutralized. Jayed lenient (0Y in GER-CUR), 30-40% pen band.
- Win%: POR 75% / DRAW 19% / UZB 5% — tight (POR 75-76%).
- Scoreline: POR 2-0 modal (mm3 14%, g4 22%, g5 16%). Runner-up: POR 2-1 / POR 3-0 (Finding 22 blowout fires per g4 ~17% — Nematov has no veteran-GK exclusion).
- Who Scores: ✅ Ronaldo #1 (~22%) — drought + age-discount (Finding 16) but pen duty + milestone (6th separate WC) keep him #1. Bruno Fernandes (~13%) #2.
- Finding 14 (GK change): ✅✅ MODEL-BREAKING — Yusupov (44 caps #1) → Nematov (3rd-choice, 14 caps, WC debut). +3-5pt to POR.
- Finding 17 (sterile-domination): ⚠️ risk, mitigated — POR MD1 was 0.64 xG vs low block; UZB 5-4-1 similar. BUT Nematov backup GK < DRC's Mpasi → draw held to ~19%, not 25%+.
- Finding 11: ✅ NEUTRALIZED — Houston roof CLOSED.
- Finding 12: ~3-5% POR conservation (Martínez pragmatic, MD3 vs COL matters) BUT pressure to respond after poor MD1 offsets.
Model Comparison
| Model | POR | Draw | UZB | Modal Score | #1 Scorer | F22 Blowout |
|---|---|---|---|---|---|---|
| mm3 | 76% | 19% | 5% | POR 2-0 (14%) | Ronaldo 22% | — |
| g4 | 76% | 20% | 4% | POR 2-0 (22%) | Ronaldo ~22% | fires ~17% (no veteran-GK exclusion) |
| g5 | 75% | 18% | 7% | POR 2-0 (16%) | Ronaldo 22% | moderate |
| Consensus | 75% | 19% | 5% | POR 2-0 | Ronaldo ~22% | partial (g4 bullish) |
Spread: 1pt (75-76% POR). g4 most bullish on blowout (Finding 22 fires — Nematov has no veteran-GK exclusion), g5/g4 agree Ronaldo #1 at 22%.
Decision Graph (doc-5) Resolution
- Group context: COL 3, POR 1, DRC 1, UZB 0. UZB MUST-WIN (eliminated with loss). POR win = near-qualified.
- Conservation (F12): POR mild ~3-5% (Martínez pragmatic, MD3 vs COL matters) but pressure to respond after poor MD1 (1 SoT) offsets. UZB cannot conserve.
- Weather × style: Houston roof CLOSED → Finding 11 NEUTRALIZED.
- Weather × defense: N/A (roof closed).
- Temperament: Martínez possession/patient/trusts Ronaldo. Cannavaro 5-4-1 low block, counters via Fayzullaev/Shomurodov. UZB GK swap = defensive question mark.
- Synthesis: POR base ~75-80%. UZB GK swap (+3-5pt POR, Finding 14 model-breaking). POR MD1 poor form (-2-3pt, offset by pressure). UZB must-win opens up → favors POR scoring. Finding 17 risk mitigated by Nematov weakness. Net POR 75%.
Key Risks & Watch
- 🔴 UZB GK CHANGE (Nematov in) — biggest single factor. 3rd-choice, 14 caps, WC debut. If he has a "Mendy heroics" game, draw live; if he's shaky, POR clears the draw floor and Finding 22 blowout becomes possible.
- Ronaldo drought vs milestone — 10-game major-tournament drought (Finding 16 age-decline) BUT milestone motivation (could score at 6th separate WC) + Jayed's 30-40% pen band + designated taker. Plays 90, sub-risk-free.
- Finding 22 blowout — g4 fires (~17% on 4+ margin) because Nematov has NO veteran-GK exclusion (unlike ECU-CUR/URU-CPV). mm3/g5 more conservative (weather-roof-closed caps margin). If POR scores early, 3-0+ live.
- UZB counter-threat — Fayzullaev (scored MD1, first-ever UZB WC goal) + Shomurodov. UZB MUST-WIN desperation opens them up → favors POR but also gives UZB counter chances.
- COL-DRC MD2 (same day) — Group K final positions depend on this too. Marked
(plays Jun 23), no fabrication.
Watch
Nematov's first 20 minutes. If the 3rd-choice GK looks shaky on an early set-piece/cross, POR clears the draw floor and 2-0+ is live. If he makes an early wonder-save, the 1-1 repeat of POR-DRC is genuinely in play despite the downgrade.
Consensus ensemble for POR-UZB. 3 models (mm3/g4/g5) from full gathering. Full post-event after kickoff. Last updated Jun 23, 2026.