Croatia vs. Ghana - Post-Event Review

2026 World Cup, Group L MD3 | Sat 27 June 2026 | Lincoln Financial Field, Philadelphia, PA (open-air)

⚠️ FINDING 24 PARTIALLY INVALIDATED | Predicted vs. Actual | Calibration: ~70%

TL;DR

Result: Croatia 2-1 Ghana (HT 1-0; Šučić 30' long-range · Luckassen 72' VAR equalizer · Vlašić 82' from Modrić corner). Croatia's old guard willed a must-win: Modrić — on his 201st cap — became the oldest WC assist-provider on record (40y 291d) to set the winner, Šučić won FIFA POTM, and a draw-suffices Ghana side actually attacked (HT double sub), invalidating the F24 low-event read.

Calibration Breakdown: Winner 25/25 · Scoreline 15/25 · Who Scores 10/20 · Findings 13/20 (F24 0, F6/Asare 3, F17 5, F12 5) · Model-breaking 9/10 · deductions 0 = 72/100 → ~70% (C band)


1-5. Summary

6. Match Stats (FINAL)

Stat Croatia Ghana
Possession 53% 47%
xG (FOX / GSN / xGscore) 0.53 / 0.42 / 0.57 0.61 / 0.74 / 0.67
Shots 8 6
SoT 4 1
Corners 3 2
Fouls 9 13
Yellows 1 1
Reds 0 0
Pens 0 0
GK saves 0 2 (Asare 29', 81')

(Sources: FOX Sports boxscore, GhanaSoccerNet, xGscore.io. CRO won despite Ghana out-xG-ing them on every feed.)

7. Key Learnings

  1. Finding 24 needs an "agency override": the draw-suffices side is not always a park-the-bus side. F24's "must-win vs draw-suffices → low-event" failed here because Ghana (pre-qualified, draw-suffices) chose to attack (HT double sub chasing the game). Refinement: before applying the F24 low-event tilt, check manager temperament (Queiroz's "play to win" rhetoric was real, not bluster) and whether the draw-suffices team has a counter-attacking/attacking identity. Park-the-bus ≠ draw-suffices. Propose adding to F24: "draw-suffices + attacking-identity manager → do NOT tilt to low-event."
  2. Finding 6 (hot-GK draw floor) can be overcome by a set-piece against a tiring block. Asare performed exactly as predicted (2 saves, MOTM-tier), the draw was live at 80', but a single corner (Modrić → Vlašić) broke it. The mechanism fired; CRO's set-piece quality was the unmodeled closer. Refinement: the hot-GK draw floor is fragile to late set-pieces — when the favorite has a corner/set-piece creator (Modrić) + a header target, the draw floor leaks at 80'+.
  3. Finding 17 (no elite #9) validated again on CRO — and the "clinical-on-low-xG" escape valve is real. CRO won 2-1 from 0.53 xG: Budimir blanked (F17 mechanism), but midfield goals (Šučić long-range, Vlašić header) compensated. Refinement: F17 caps the striker's goals, not the team's — a midfield-heavy side can beat its own xG via non-#9 sources. Don't over-shade the team win% down on F17 alone when creation is elite.
  4. Both-consensus-#1-missed is a recurring WHO-SCORES failure mode worth tracking. Budimir (CRO) and Semenyo (GHA) — the two #1 picks — both blanked; goals came from secondary picks (Šučić) and a defender (Luckassen). When the modal scoreline is low (1-0/1-1) and the favorite has no elite finisher, the #1 scorer pick is low-confidence and deep-list picks deserve more weight. Feeds Finding 19.
  5. Modrić age-decline (F16) suppresses scoring, NOT creating — re-confirmed. 0 goals but a match-winning assist + 22nd WC appearance. Demote age-declined stars in the scorer list, keep them elevated as creators. (Consistent with the CRO-PAN refinement.)

8. Group L after MD3 (final)

Pos Team P Pts GD
1 England 3 7 +4
2 Croatia 3 6 0
3 Ghana 3 4 0
4 Panama 3 0 -4

(GD grounded: ENG 4-2 / 0-0 / 2-0 = +4; CRO 2-4 / 1-0 / 2-1 = 0; GHA 0-0 / 1-0 / 1-2 = 0; PAN 0-2 / 0-1 / 0-1 = -4. All MD3 matches played; Panama eliminated. Ghana's 4pts qualifies as a best-3rd-place team.)


Post-event review for TASK (CRO-GHA). Category 0 PASSED: CRO 2-1 GHA confirmed from scout brief (≥8 sources: FOX, Guardian, Sky, ESPN, Reuters, USA Today, AS, FIFA). Sources: sources/matches/cro-vs-gha/post-event/ (7 files). FINAL stats only. Flags: ⚠️ Finding 24 partially invalidated (draw-suffices side attacked); VAR equalizer (75' monitor review, goal stood); Modrić oldest WC assist record (40y 291d).