Nice 2-1 Lille
Model accuracy
Retrospective model estimate · generated after the match, not in the track recordThere was no estimate locked before kick-off for this match. The model was run after the match using only data known before kick-off (the result was not used); it is not counted in the track record or accuracy stats.
The 3 clearest pre-match estimates and how they turned out. How are these chosen?
Statistical model estimate; for information and entertainment only.
Match stats
Standouts: Y. Diouf 8.3 · M. Amoura 8.3 · A. Ueda 7.7
Match analysis
5 sections12:02Nice beat Lille 2-1 in the Ligue 1 matchday 5 on 20 September 2026.
Read more ›Close
Result and story of the match
It was 1-0 at half-time. Lille scored first, in the 22nd minute.
Goals and key moments
Goals: 22' N. Ngoy (own goal) → 0-1; 67' M. Amoura, assisted by G. Hein → 1-1; 79' A. Ueda, assisted by M. Kjaergaard → 1-2.
Yellow cards: Nice 3, Lille 1.
Statistical comparison
Expected goals (xG) Nice 1.50, Lille 1.58; shots 10-21 (4-8 on target); possession 41%-59%; corners 3-7; pass accuracy 86%-87%.
Goals by half: 1-0 in the first, 1-1 in the second.
Standout players
M. Amoura was Nice's highest-rated player at 8.3 in 77 minutes: 1 goal, 1 shot on target, 2 key passes, 4 tackles, won 5 of 11 duels.
A. Ueda was Lille's highest-rated player at 7.7 in 90 minutes: 1 goal, 4 shots on target, 1 key pass, 1 tackle, won 3 of 4 duels.
How the model did
There was no model estimate locked before kick-off for this match, so it is not part of the track record.
Retrospective model estimate (generated after the match from pre-match data only; not part of the track record): Nice 25%, draw 28%, Lille 47%.
Statistical model estimate; for information and entertainment only.
Match summary
Nice beat Lille 2-1 in the Ligue 1 matchday 5 on 20 September 2026. Goals: N. Ngoy 22' (o.g.), M. Amoura 67', A. Ueda 79'. Match stats: possession Nice 41%, Lille 59%; shots on target 4-8; expected goals (xG) 1.50-1.58.
Updated: · Sources: Methodology, Track record
Points are only for rankings and badges; they have no other value.The model's score is now the “predicted score”: a pick that agrees with the match outcome and follows the real score distribution over a season (previously the single most likely score). Past points are unchanged.