Portugal vs. Spain — Post-Event Review

2026 World Cup, Round of 16 | Sun July 6, 2026 | AT&T Stadium, Arlington TX (roof CLOSED)

TL;DR

Result: POR 0-1 ESP (HT 0-0; Merino 90'+1', sub). Spain's 0 GA survived its first Tier 1 test — Simón's clean sheet and a stoppage-time sub winner settled a genuine coin-flip that a tight, low-event shape predicted perfectly but the modal specifics missed cleanly.

Winner: ⚠️ ESP advance (consensus 50/50 co-modal; ESP won in 90' on a genuine 33% band). The model refused a false favorite — honest coin-flip. g4 was the lone model to lean ESP (52% advance) and got it right.
Scoreline: ⚠️ modal wrong, shape nailed — consensus modal was 1-1; actual POR 0-1 ESP (tight 1-goal shape + exact margin right). g4's #1 scoreline ESP 1-0 (15%) hit EXACT.
Who Scores: ❌ both #1 picks blanked — Ronaldo (POR #1, ~22%) 0 goals; Oyarzabal (ESP #1, ~21%) 0 goals; Merino (unranked sub, 90'+1') decided it.
Finding 11: ✅ — roof CLOSED → F11 neutralized, exactly as predicted.
Finding 12: ✅ — 0% conservation both sides (R16, must-win, full effort to the final kick).
ESP 0 GA flag (g5): ❌ — flagged as opponent-inflated; it HELD vs a Tier 1 attack (POR xG 0.60, 0 goals). The defense is genuinely elite, not a soft-group mirage.
Ronaldo farewell (R32-5g): ❌ — record-chasing→goals predicted; farewell→0 goals (Finding 16 age-decline won at 41).

Calibration Breakdown: Winner 15/25 · Scoreline 15/25 · Who Scores 0/20 · Findings 10/20 · Model-breaking 7/10 [−5 Merino unpredicted sub scorer, Finding 19] = 42/100 → ~43% (F)

The strict rubric scores on the modal recommendation, not the shape. The model's structural reads were sound (coin-flip, tight, low-scoring, full effort) — but every modal pick missed (1-1 scoreline, Ronaldo/Oyarzabal #1 scorers) and the 0 GA defense held against the g5 flag. An honest coin-flip that landed on a real band is good calibration discipline, not prescience — the ~43% reflects "right shape, wrong specifics," consistent with the POR-CRO precedent (F19 sub winner, F16 Ronaldo blank).

1-5. Summary

Decisive narrative: A goalless, chance-starved Iberian derby was settled in the 91st minute by substitute Mikel Merino, breaking a POR side that had held firm for 90'. ESP's elite defensive spine (Cubarsí–Laporte–Rodri in front of Simón) never cracked — POR generated just 0.60 xG and zero goals. The 0 GA record survived its first genuine Tier 1 examination.
ESP were the better side on the ball: 56% possession, 1.77 xG (vs POR 0.60) — the form edge the WDM compressed was, in hindsight, the real separator. ESP created ~2 big chances, converted one (Merino).
Diogo Costa kept POR alive: 5 saves; without him the margin is wider. But POR's attack (Ronaldo, Bruno, Leão/Félix) never found rhythm against ESP's control.
Ronaldo's farewell ended goalless: full effort (F12 = 0%), but at 41 the milestone translated to minutes, not goals — the third WC 2026 validation of the age-decline rule (after POR-DRC, GHA-Ayew).
No weather, no pens, no controversies: roof closed neutralized F11; Anthony Taylor's low-pen profile held (0 pens, within the 25-35% band); no GK surprises (Costa/Simón both as confirmed).
Lineup as projected: Williams/Pino OUT for ESP (assessed −3-5pt, correctly NOT gutting); the Yamal–Oyarzabal–Pedri–Rodri spine started intact.
Coin-flip validation: a 1-0 decided by a 90'+1' sub is precisely the outcome a 50/50 advance prediction allows. The model did not embarrass itself with false confidence — it simply was not prescient on the specifics.

6. Match Stats (FINAL)

Stat	Portugal	Spain
Possession	44%	56%
xG	0.60	1.77
Big chances	1 (missed)	~2 (1 converted, 1 missed)
GK saves	Simón — (clean sheet)	Costa — 5
Fouls	9	13
Pens	0	0

Shots / Shots-on-target / Yellow / Red card counts were not captured in the scout brief and are left blank rather than estimated. ESP SoT ≥ 5 is inferable from Costa's 5 saves. Category-0 result confirmed from ESPN Matchcenter box score (gameId 760506); high-profile R16 Iberian derby, widely corroborated.

12-Category Validation Checklist

#	Category	Prediction	Actual	Verdict
1	Winner	50/50 advance; 90' POR 37% / Draw 30% / ESP 33%	ESP won 1-0 in 90', advance to QF	⚠️ Co-modal (ESP 50% advance) advanced; 90' ESP win was a genuine 33% band. g4 (52% ESP) the lone leaner — correct.
2	Scoreline	Modal 1-1 (unanimous: mm3 14% / g4 12%* / g5 16%)	POR 0-1 ESP	⚠️ Modal 1-1 wrong; tight 1-goal shape + margin exactly right; g4's #1 ESP 1-0 (15%) hit exact.
3	Who Scores	POR #1 Ronaldo ~22%; ESP #1 Oyarzabal ~21%	Merino 90'+1' (sub, unranked)	❌ Both #1 picks blanked; all top-5 missed; winner from a sub.
4	Weather	Roof CLOSED → F11 neutralized	Roof CLOSED; F11 neutral	✅ Validated exactly.
5	Possession (FINAL)	ESP to dominate territory	POR 44% / ESP 56%	✅ Form-favorite had the ball — not a possession flip.
6	xG	ESP form edge (~0 GA spine)	POR 0.60 / ESP 1.77	✅ ESP clearly superior; 0 GA held AND attack was better.
7	Cards / Pens	Taylor low-pen (25-35% band), 3-4Y, not high-pen archetype	0 pens; fouls POR 9 / ESP 13	✅ Pen band held (0 pens within band); ref profile accurate.
8	Lineup	Williams/Pino OUT (assessed, not gutting); spine intact	As projected	✅ No XI surprises; the −3-5pt single-absence read correct.
9	GK	Costa (POR) / Simón (ESP), both confirmed	Costa 5 saves; Simón clean sheet	✅ Both GKs as confirmed — no model-breaking change.
10	Substitutions	Sub risk noted (F12 interaction); no specific sub-scorer weighted	Merino (sub) 90'+1' winner	❌ Late-sub-winner (F19) recurred; not weighted in Who Scores.
11	Finding 12 (conservation)	0% both (R16 knockout, must-win)	0% both — full effort to 90'+	✅ Validated; ESP chasing a late winner proves zero conservation.
12	Key learnings	—	0 GA real (not inflated); F16 > R32-5g for 41yo farewell; F19 sub pattern	→ §7

* The consensus reconciliation listed g4's modal as 1-1 (12%), but g4's own scoreline distribution puts ESP 1-0 first at 15% and 1-1 second at 14% — i.e., g4's true #1 scoreline was ESP 1-0, the exact result. g4 was the consensus's best-performing model this match.

7. Key Learnings

REFINES the R32-5g vs Finding 16 boundary (Ronaldo farewell, 3rd validation). The pre-match applied "R32-5g record-chasing → goals" to make Ronaldo POR #1 (22%). Reality: 0 goals. Finding 16 (validated POR-DRC, GHA-Ayew) wins for a 41-year-old FAREWELL — milestone motivation buys minutes, not finishing. The clean rule emerging from three cases: **R32-5g (record-chasing → goals) applies to ACTIVE record pursuit (Haaland 12-game streak, Mbappé chasing Klose, Kane → Lineker) where the star is ≤32; it does NOT apply to a farewell/last-tournament 38+ star, where Finding 16 (age-decline, demote #1 → top-3) governs.** Ronaldo already HOLDS the 6-WC scoring record — he was farewelling, not chasing. Proposed edit to doc-1 Finding 16 / R32-5g: "farewell ≠ record-chasing; apply age-decline to farewell stars regardless of milestone framing."
ESP's 0 GA is REAL — the g5 opponent-inflated flag was wrong (refines hot-GK system rule). g5 (best-in-class model-breaker detection) flagged ESP's 0 GA as opponent-inflated (CPV/KSA/AUT only) via the ENG-MEX analogy (MEX's 0 GA broke vs ENG). It held: POR (Tier 1, xG 0.60) scored zero. The distinguishing factor: a 0 GA backed by an elite SPINE (Cubarsí–Laporte–Rodri–Simón) is genuine; a 0 GA backed by a low block + untested GK is the suspect kind. This extends the R32-5f "hot-GK = system product" rule: when grading a 0 GA record, weight the defensive SPINE quality, not just opponent strength. Don't apply the ENG-MEX warning reflexively to a side with a tier-1 club-spread back line.
Finding 19 (unpredicted late sub winner) is now the signature R16 scoring miss — weight subs higher in Who Scores. Merino 90'+1' is the third R16/R32 match decided by a sub scorer (Martinelli BRA-JPN 90+6', G.Ramos POR-CRO 90', now Merino). All were unranked in the Who Scores top-5. Calibration rule: in any KO match predicted as tight (draw band ≥25% or margin ≤1 goal), reserve a ≥8-10% slice for "impact sub after 75'" in the scorer distribution — currently under-weighted. The −5 deduction applied here (consistent with POR-CRO precedent).
VALIDATES the coin-flip honesty discipline (Finding 9). A €1000M vs €861M squad gap and a 37%-vs-33% 90' lean is NOT an edge — the ensemble correctly refused to tip a favorite (advance 50/50, spread 2pt). The match (1-0, 90'+1') genuinely could have gone either way. Learning: when fav_share is 0.49-0.52, the honest call is "no edge, trust the draw band" — not a soft lean. The cost of being honest is you can't claim a "winner hit" when the coin lands; the benefit is you never embarrass the framework with a false favorite. This is correct behavior, even though it scores low on a modal-pick rubric.
g4 (glm-5-turbo) was the match-winner model — and the consensus reconciliation under-credited it. g4 was the ONLY model to lean ESP advance (52%) AND nailed ESP 1-0 as its #1 scoreline (15%). The consensus table reported g4 as "advance POR 48%" without flagging that g4's directional lean (ESP) was uniquely correct. Process note: when one agent's modal scoreline is an EXACT hit and its advance lean is uniquely correct, the reconciliation should elevate that agent's reasoning in the "Watch" line, not just average it away. g4's favorite/form-edge sensitivity delivered outside its usual "clear-favorite" comfort zone.

8. Knockout Bracket — QF (after POR-ESP)

Stage	Matchup	Status
QF	Spain vs [ARG-EGY winner]	✅ ESP through (1-0 vs POR); opponent = ARG-EGY R16 winner (plays, pending)
QF	France vs Morocco	✅ Confirmed (FRA + MAR through)
QF	Norway vs England	✅ Confirmed (NOR + ENG through)
QF	[USA-BEL winner] vs [remaining R16 winner]	⚠️ TBD (R16 ties pending)

Eliminated: 🇵🇹 Portugal (R16). Through: 🇪🇸 Spain → QF vs ARG-EGY winner. Other R16 ties (ARG-EGY, USA-BEL, and the slot feeding QF4) are unplayed/pending and are NOT filled here — no fabrication of unplayed results.

Post-event for TASK-POR-ESP (R16). Category 0 PASSED: result Spain 1-0 Portugal confirmed from ESPN Matchcenter box score (gameId 760506); high-profile R16 Iberian derby, widely corroborated. Scout brief: sources/matches/por-vs-esp/post-event/00-post-event-brief.md (1 file). FINAL stats only. ⚠️ Full shot/SoT/card counts not captured in scout brief (left blank, not estimated). ⚠️ g4's true #1 scoreline was ESP 1-0 (15%) — the consensus reconciliation understated it as a 1-1 modal. Strict rubric applied (revised Jun 30): scored on modal picks; co-modal advance + tight-shape credit only (no "in top-3" inflation). F19 −5 deduction applied per POR-CRO precedent.