France vs. Senegal - Post-Event Review

2026 World Cup, Group I, Match 17 | June 16, 2026 | MetLife Stadium, NJ

Predicted vs. Actual | Calibration: ~75%

TL;DR

Result: France 3-1 Senegal (HT 0-0). France won as predicted (all 5 versions). Mbappé was the unanimous #1 scorer pick and scored TWICE — the strongest Who Scores hit of the tournament. But Finding 12 (Conservation) was the real story: France were 0-0 at half-time, sluggish, conserving, and only turned it on after 60'. The draw never materialized — quality told late — but the conservation pattern was textbook.

• Winner: ✅ France — unanimous across all 5 versions (consensus ~56%)
• Who Scores: ✅✅ Mbappé — unanimous #1 (~25%), scored 66' + 90+6'. Best scorer hit yet
• Scoreline: ⚠️ 3-1 — modal pick was FRA 2-1; 3-1 was a secondary scenario in human versions (8-9%), not in machine top-3
• Senegal scorer: ❌ — models split Jackson vs Mané; NEITHER scored. It was Ibrahim Mbaye (18yo sub)
• GK: ✅ Mendy started — UK sources right, L'Equipe (Diaw) wrong
• Cards: ✅ ZERO yellows — extreme referee leniency, Finding 3 validated again
• Weather (Finding 11): ✅ VALIDATED — B-version corrected the A-version's error (A said 30-32°C peak heat; actual NWS was 26°C). Match conditions confirmed comfortable. The weather correction was the right call — the A-version's heat thesis was wrong.
• Possession (Finding 8): ✅ — France 54% possession. Expected slight majority, no flip.
• xG: France 1.92 xG, Senegal 0.50 xG — France's 3 goals were fully deserved (xG overperformance but not absurd). Senegal's 1 goal from 0.50 xG = slightly lucky/efficient.
• Finding 12 (Conservation): ✅ — see detailed analysis in §3
• Finding 12: ✅ VALIDATED — 0-0 at HT, France sluggish, conserved for the second half

Watch (validated): Senegal's GK. Mendy started and made several good saves. The model's GK ambiguity flag was the right call to raise.

1. Calibration Scorecard

Overall: ~75% — Winner ✅, Who Scores ✅✅, Cards ✅, but Scoreline misses top-3 and Senegal scorer missed.

2. Version-by-Version Comparison

The human versions (A and B) outperformed the machines on scoreline — both had FRA 3-1 at 8-9% in their top 5, while no machine model had 3-1 in its top picks. This is the first match where the human analysis beat the machine ensemble on scoreline calibration. The machines over-weighted the draw (mm3 28%, g4 28%) which never materialized.

3. Finding 12 (Conservation) — VALIDATED

This is the most important post-event finding. France were 0-0 at half-time, sluggish, and conserving energy. From the Guardian:

"In classic Didier Deschamps fashion, France struggled early but recovered to overwhelm Senegal."

• France had the deeper attack and the quality edge, but played the first half at ~60% intensity
• Senegal had the better first-half chances (Sarr missed a big one before HT)
• France only "turned it on" after 60' — Mbappé's opener was the release valve
• This is the textbook Finding 12 pattern: Tier 1 contender conserves on MD1, doesn't go full-throttle, plays within themselves until the quality gap forces it open

Key nuance: Conservation delayed the win, it didn't prevent it. France conserved for 60 minutes, then their quality told. The draw probability (~25% consensus) was too high in retrospect — conservation made the first half tighter, but France's ceiling was always higher. The lesson: conservation shifts the TIMING of goals (later), not necessarily the result.

If Finding 12 had been applied pre-match (which it wasn't — it was developed AFTER this match), the overlay would have:

• Pushed FRA 1-0 and 0-0 HT higher (sluggish starts)
• Pushed "late goals" scorelines (FRA 2-1 with a 70'+ winner) higher
• Still kept France favored — conservation doesn't flip the winner, just the shape

4. Who Scores — The Best Hit of the Tournament

Mbappé was the unanimous #1 pick across all 5 versions (~22-25%). He scored twice (66', 90+6'). This is the strongest Who Scores validation we've had — no model disagreed, and the pick was emphatically correct.

• Mbappé became France's all-time top scorer (58 intl goals) and top WC scorer (14)
• Both goals came from Olise creativity — validating star-as-creator (Finding: Olise was predicted as #2-3 scorer but instead was POTM as creator)

The Senegal scorer miss: Models split Jackson (human) vs Mané (machine). Neither scored. The actual scorer was Ibrahim Mbaye, an 18-year-old PSG sub who came on at 76'.

Partial credit: Jackson had a goal ruled out for offside (68') — the model's instinct on Jackson as Senegal's main threat was directionally right, but he couldn't beat the flag. Neither Jackson nor Mané found the net from open play.

5. Referee & Penalty

Zero yellow cards. Faghani called one of the most lenient matches of the tournament. The 20-25% tournament discount (Finding 3) continues to be validated — every single tournament referee has been more lenient than their domestic career rate.

Penalty drama: At 59', Mbappé was fouled by Mané in the box. VAR sent Faghani to the monitor. He said NO PENALTY — a controversial decision that pundits and former referees strongly disagreed with. The penalty MODEL was directionally right (a pen-worthy incident occurred, ~15-20% chance), but the call went unrewarded. This is irreducible referee variance.

6. Key Learnings

1. Finding 12 validated (Conservation): France (Tier 1 contender) were 0-0 at HT on MD1. But conservation shifted goal TIMING (later), not the result. France's quality told in the second half. New calibration note: conservation makes the first half tighter and pushes goals later — model "late winner" and "second-half scoring" scenarios higher. It doesn't necessarily increase the draw as much as the base rule implies when the favorite's quality edge is large.

2. Human > machine on scoreline (first time): The human versions (A, B) had FRA 3-1 at 8-9% in their top 5; no machine did. The machines over-weighted the draw. The human's intuition for "quality gap opens late" outperformed the machine's draw-heavy calibration on this match.

3. Mbappé unanimous #1 = best Who Scores hit yet. When all models agree on a clear #1 scorer, the confidence should be high. No dissent = strong signal.

4. Star-as-creator confirmed again: Olise was predicted as a scorer (#2-3) but was POTM as creator (assisted Mbappé's opener, changed the game). Model creators as goal-scorers AND assist-providers. Olise's xA impact was the match's decisive factor.

5. GK ambiguity resolved correctly: The pre-match flag on Mendy vs Diaw was the right call. Mendy started and made several key saves (kept it 0-0 at HT). The model correctly identified this as a model-breaking variable.

6. Sub scorers (again): Both Barcola (FRA, 82') and Mbaye (SEN, 90+5') scored as substitutes. Substitution risk modeling continues to be validated — late subs are live scoring threats.

7. Draw epidemic did NOT strike here. 8 of the first 14 matches drew, but France broke the pattern. When the favorite's quality edge is large AND they have a generational scorer (Mbappé), conservation produces a slow start — not a draw. The draw epidemic applies most to matches where the underdog has a defensive structure (low block + hot GK). Senegal played an open 4-3-3 with a high line — exploitable.

7. Updated Cumulative Stats (15 matches)

Post-event review for TASK-42. Sources archived: sources/matches/fra-vs-sen/post-event/01-guardian-report.md. No betting odds used. FINAL stats only.