Argentina vs. Cape Verde - Post-Event Review

2026 World Cup, Round of 32 | Sat July 4, 2026, 00:00 CEST | Hard Rock Stadium, Miami FL (OPEN-AIR, 33°C)

R32 (knockout) | Predicted vs. Actual | Calibration: ~70%

TL;DR

Result: Argentina 3-2 Cape Verde (AET — ARG advance to R16 vs Egypt). HT 1-1; 90' 1-1; Messi 29', L.Martínez 92', Borges 111' OG / D.Duarte 59', Cabral 103'. ARG were the heavy favorite who won — but via the non-modal path (draw at 90' → ET → own goal), pushed to the limit by a block+hot-GK template that came within 8 minutes of a penalties upset.

• Winner (advance): ✅✅ ARG (advance ~75%, the strict §7K modal). The advance bet paid — but it took ET and a 111' own goal; CPV led 1-2 at 103'. The clean 90' modal (ARG win 58%) did NOT fire.
• Scoreline: ❌ modal ARG 2-0 wrong; ✅ non-modal draw shape hit — modal ARG 2-0 (90'); actual 90' was 1-1 draw (the 23% draw band, named as "0-0/1-1 → ET"); AET 3-2 was unmodeled.
• Who Scores: ✅✅ #1 Messi hit (29', ~30% pick). Extends WC all-time record to 19. L.Martínez (#2, ~22%) also hit (92'). CPV #1 Pina ❌ missed — surprise scorers Duarte + Cabral.
• Finding 11 (heat): ✅ — 33°C open-air degraded the ARG press; match went to ET in the heat exactly as the elevated draw band warned.
• Finding 22 (blowout exclusion): ✅ fired hardest-possible — 4+ margins capped 5-8%; actual was 3-2 AET, the closest possible margin. The partial-exclusion call was correct.
• Finding 6 / hot-GK Vozinha: ✅ — 8 saves, confirmed-form breakout held (×0.5-0.6 validated); the block+GK compound is why this went to ET.
• Finding 12 (conservation = 0%): ✅ — Scaloni's side chased to ET; no conservation.

Calibration Breakdown: Winner 25/25 · Scoreline 5/25 · Who Scores 15/20 · Findings 20/20 · Model-breaking 10/10 · −5 (Finding 19: CPV surprise scorers Duarte + Cabral materially affected the 3-2) = 70/100 → ~70% (C band)

1-5. Summary

• Decisive narrative: ARG won the advance bet but lost the clean-path bet. The §7K modal was "ARG win in 90' (58%)"; instead the 23% draw band fired (1-1 at 90'), and ARG needed a Lautaro equalizer-frame goal (92'), then a Borges own goal (111') to survive. This is a "right winner, wrong path" outcome.
• Messi milestone rule holds (5/5+ in R32): Messi opened the scoring (29'), now 19 WC goals — the unanimous #1 pick delivered as the milestone rule predicted.
• CPV are not a minnow (Finding 5 re-validated): Having already drawn ESP 0-0 and URU 2-2 at this venue, CPV pushed ARG to ET and led 1-2 at 103'. The defensive-block-specialist calibration is now triply validated.
• Clinical finishing inverted the xG story: ARG 2.26 xG → 3 goals (incl. an OG); CPV 0.47 xG → 2 goals. CPV scored 2 from 0.47 — extreme counter-attack/set-piece efficiency that is the signature of the block-and-break template.
• Referee Fischer as profiled: 0 in-game pens (pen band 25-30% → 0), tournament-lenient, 25 total fouls, Messi unbooked. No shootout was needed (ARG won in ET).
• No milestone letdown: ARG (defending champion, milestone Messi) showed F12 = 0% conservation — they exerted maximum effort; the draw came from CPV's quality, not ARG apathy.

6. Match Stats (FINAL)

Shots/SoT/Yellows cells marked "—" were not surfaced as FINAL in the scout brief; per project rules they are left blank rather than fabricated. Factual cells that ARE confirmed (score, scorers, possession, xG, fouls, pens, GK saves) come from ESPN Matchcenter gameId 760500.

7. Key Learnings

1. Finding 6 (low-block + hot-GK) re-validated at the extreme — and the 23% draw band was CORRECT to be elevated. CPV's block + Vozinha (8 saves) held ARG to 1-1 at 90' and led at 103'. This is the third CPV result at Hard Rock (ESP 0-0, URU 2-2, ARG 1-1→3-2 AET). Refinement: for a team with ≥2 prior validated block+GK results at the SAME venue, the draw floor should sit at the top of the 20-25% range, not the middle — 23% was right but the model still leaned on the 58% clean-win modal as if it were the likeliest path. The "ARG win in 90'" modal was over-confident given CPV's venue precedent.
2. Finding 22 partial blowout exclusion — fully validated as the correct tier. 4+ margins were capped at 5-8%; actual was 3-2 AET (closest possible). The block + veteran-GK template prevented the blowout completely. This confirms the partial-exclusion (vs full-exclusion) bucket is well-calibrated for "deep block + capable-but-not-elite veteran GK."
3. Finding 11 heat × Tier-1 modulator — fires but does not eliminate the penalty. ARG (Tier 1, heat-acclimatizing quality) still saw their press degrade and the match stretch to ET in 33°C. The modulator reduced but did not cancel the draw lift. Refinement: when heat (>28°C) co-occurs with a block+hot-GK underdog, the compound should be treated multiplicatively — it pushed the draw from a 15% floor to a 23% realized outcome.
4. Finding 19 (unpredicted scorers) — CPV's goals came from non-#1 routes. Both CPV scorers (Duarte 59', Cabral 103') were outside the top-3 (Pina/Da Costa/Borges predicted). For defensive-block underdogs, the scoring route is counter/set-piece by committee — the #1 pick (often the set-piece specialist) is a weak anchor. Proposed nuance: for block-and-break underdogs vs elite favorites, weight "any-of-3 CB/counter runners" over a single #1 pick; the #1 hit rate is structurally low.
5. Right answer, wrong path — flagged honestly. The advance modal (ARG ~75%) won, but via the draw+ET+OG path, not the 58% clean-90'-win path. No calibration inflation applied: the Winner component credits the advance bet (which paid), while the Scoreline component is scored low (modal ARG 2-0 wrong). This is the NED-MAR lesson in action — separate "did the bet pay?" from "did the modal path fire?"

8. Round of 16 — ARG's Path

ARG advance to face Egypt in the R16. Egypt's R32 result is not part of this review's source set and is not stated here.

Post-event for TASK-[R32 ARG-CPV]. Category 0 PASSED: score, scorers, possession, xG confirmed from ESPN Matchcenter (gameId 760500) — single primary source cross-checked against the scout brief. Sources: sources/matches/arg-vs-cpv/post-event/. FINAL stats only. STRICT rubric applied (modal-pick scoring, no "in-top-3" inflation). ⚠️ Shots/SoT/Yellows cells left blank where the brief did not surface FINAL values — not fabricated.