Premier League 2019–2025 · 2.464 Matchs · 7.274 Buts
Intelligence Buts
Six seasons, 7,274 goals, one counterintuitive discovery: early goals create the most volatile matches, not the most secure ones. A first goal before 30' leaves a 32% chance of the other team winning outright. A first goal after 60' locks in a draw 63% of the time. This is the Goal Cascade framework.
Buts marqués 76-90'
22.6%
1 but sur 4,4 dans les 15 dernières min
Marquer en premier → taux victoire
38.2%
Mais 28,9% des matchs voient des remontées
Conversion grosses occasions
61.4%
Taux global — 1 grosse occasion sur 1,6 convertie
Buts moyens / match
3.12
2.329 matchs avec buts enregistrés
🌊 Risque en Cascade des Buts — Trois Régimes de Match
The timing of the first goal determines which of three distinct match regimes takes hold. Counterintuitively, early goals produce the most volatile environments — not the most secure ones.
🔴 Jeu Ouvert · 0–30'
32.1%
risque en cascade
First goal before 30': the trailing team wins outright 1 in 3 times. The match stays fully open — the opening goal is a signal, not a sentence.
40.5% win · 27.3% draw · N=881
🟡 Jeu Contesté · 31–60'
24.8%
risque en cascade
First goal 31–60': moderate volatility. Draw becomes the most likely outcome (40%). Either side can still win — this is where live markets are least efficient.
35.2% win · 40.0% draw · N=335
🩵 Jeu Fermé · 61–90'
11.9%
risque en cascade
First goal after 60': the match locks in. Two thirds end as a draw; comebacks become rare events. Scoring late doesn't win games — it negotiates draws.
25.0% win · 63.1% draw · N=84
⚡ L'insight structurel
Early leads are rented, not owned. An opening goal before the 30th minute still leaves a 32% window for the other team to win. But a goal after the 75th minute is nearly irreversible — 7.7% comeback rate, 76.9% draw. The conventional wisdom — 'the earlier you score, the better' — inverts when you look at match regime: early goals lock in volatility, late goals lock in the result.
📊 Implication de Marché Risk-First
Markets systematically overprice the '1-0 hold' probability after an early goal. An Open Game still carries 32% cascade risk — yet pre-match odds on the opener rarely reflect this. The highest-value caution zone: a market favourite scores first in the 15–25' window. The lead looks decisive. The regime says otherwise. Contested Games (31–60') are where live market pricing is most uncertain — 40% draw probability with nearly equal win/comeback splits creates genuine market inefficiency.
⏱️ Timing des Buts — Distribution par 5 Minutes
7,274 goals across 2,329 matches (seasons 2019–20 through 2024–25). Each bar = proportion of all goals scored in that 5-minute window.
1-5'
284
6-10'
311
11-15'
337
16-20'
337
21-25'
343
26-30'
355
31-35'
391
36-40'
345
41-45'
463
46-50'
457
51-55'
433
56-60'
397
61-65'
426
66-70'
386
71-75'
368
76-80'
378
81-85'
394
86-90'
869
🔴 Rouge = fenêtre de pointe · 🟡 Ambre = au-dessus de la moyenne · 🟢 Vert = moyenne · 🩵 Bleu = en dessous
Répartition par Phase
| Phase | Buts | % | Par match |
|---|---|---|---|
| 0-15' | 932 | 12.8% | 0.40 |
| 16-30' | 1,035 | 14.2% | 0.44 |
| 31-45' | 1,199 | 16.5% | 0.51 |
| 46-60' | 1,287 | 17.7% | 0.55 |
| 61-75' | 1,180 | 16.2% | 0.51 |
| 76-90' | 1,641 | 22.6% | 0.70 |
⏱️ Temps additionnel MT1 (45+)
7.7%
de tous les buts — 560 buts
⏱️ Temps additionnel MT2 (90+)
6.7%
de tous les buts — 487 buts
⚽ Effet Premier But — Marquer en Premier Fait-il Gagner?
Based on 1,300 matches with valid scorelines and commentary. Outcome from the perspective of the team that scored first.
A gagné le match
38.2%
(496 matchs)
Match nul
32.9%
(428 matchs)
Remontée (adversaire a gagné)
28.9%
(376 matchs)
Taux de Victoire selon Quand Vous Avez Marqué en Premier
| Premier but à | N | Vict.% | Nul% | Remontée% |
|---|---|---|---|---|
| 0-15' | 515 | 41.0% | 26.6% | 32.4% |
| 16-30' | 366 | 39.9% | 28.4% | 31.7% |
| 31-45' | 206 | 36.4% | 36.4% | 27.2% |
| 46-60' | 129 | 33.3% | 45.7% | 20.9% |
| 61-75' | 58 | 29.3% | 56.9% | 13.8% |
| 76-90' | 26 | 15.4% | 76.9% | 7.7% |
⚡ But précoce (0-15') — le plus décisif
41.0%
taux victoire. Mais 32,4% de ces matchs voient encore une remontée — marquer tôt N'EST PAS une garantie.
🕐 But tardif (76-90') — rarement décisif
15.4%
taux victoire. 76,9% des matchs se terminent nul après un premier but 76-90' — marquer tard mène rarement à un second.
🔄 Meilleure équipe de remontée
Manchester City
50,9% de remontée quand menés (29/57 matchs). Sheffield United le pire à 11,5% (3/26).
🕰️ Profil Temporel des Buts par Équipe
Average goal minute and late/early scoring tendencies. Sorted by average goal minute — teams at the top score later in matches.
🔴 Équipes qui marquent le plus tard
Brentford27.4%
Everton26.7%
Leicester City25.8%
Bournemouth25.8%
Nottingham Forest25.2%
% buts marqués en 76-90'
🟢 Équipes qui marquent le plus tôt
Burnley15.5%
West Bromwich Albion15.4%
Crystal Palace15.2%
Manchester City15.1%
Wolverhampton Wanderers14.9%
% buts marqués en 0-15'
| Équipe | Buts | Min moy. | % en 76-90' | % en 0-15' |
|---|---|---|---|---|
| Sunderland | 32 | 55.0' | 21.9% | 9.4% |
| Luton Town | 65 | 53.9' | 23.1% | 9.2% |
| Nottingham Forest | 163 | 52.3' | 25.2% | 12.3% |
| Brentford | 252 | 52.2' | 27.4% | 10.7% |
| Newcastle United | 348 | 51.9' | 23.6% | 11.5% |
| Watford | 107 | 51.8' | 24.3% | 13.1% |
| Brighton and Hove Albion | 307 | 51.8' | 24.1% | 12.1% |
| Arsenal | 353 | 51.6' | 22.1% | 11.3% |
| Leicester City | 256 | 51.5' | 25.8% | 10.2% |
| Chelsea | 326 | 51.5' | 24.8% | 9.8% |
| Bournemouth | 213 | 51.4' | 25.8% | 11.3% |
| Everton | 292 | 51.1' | 26.7% | 14.0% |
| Manchester United | 356 | 50.8' | 21.9% | 11.5% |
| Sheffield United | 144 | 50.7' | 20.8% | 9.7% |
| Crystal Palace | 302 | 50.5' | 23.5% | 15.2% |
| Leeds United | 185 | 50.5' | 23.8% | 13.0% |
| West Ham United | 359 | 50.0' | 19.8% | 11.7% |
| Aston Villa | 354 | 50.0' | 20.6% | 13.0% |
| Tottenham Hotspur | 364 | 49.8' | 21.2% | 14.0% |
| Wolverhampton Wanderers | 308 | 49.8' | 23.1% | 14.9% |
| Fulham | 226 | 49.6' | 19.5% | 11.9% |
| Liverpool | 372 | 49.4' | 23.9% | 14.8% |
| West Bromwich Albion | 52 | 48.4' | 19.2% | 15.4% |
| Norwich City | 108 | 48.4' | 18.5% | 13.9% |
| Ipswich Town | 58 | 48.2' | 22.4% | 12.1% |
| Burnley | 220 | 47.9' | 20.5% | 15.5% |
| Manchester City | 430 | 47.8' | 21.2% | 15.1% |
| Southampton | 252 | 47.4' | 17.5% | 14.7% |
🎯 Efficacité des Grosses Occasions — Qui Convertit Sous Pression?
bigChanceScored ÷ bigChanceCreated from Opta team_stats across 1,299 matches with recorded big chance data. A 'big chance' is a situation where a player is expected to score.
Le plus prolifique
Manchester City
2.72
grosses occasions / match
Meilleur taux de conversion
Bournemouth
70.5%
(échantillon 86+ matchs)
Pire taux de conversion
Brighton
54.8%
(129 matchs, significatif)
Taux de conversion global
61.4%
1 grosse occasion sur 1,6 devient un but
| Équipe | M | Créées | Marqués | Conv.% | GO/match | Vict.% |
|---|---|---|---|---|---|---|
| Luton Town | 28 | 36 | 33 | 91.7% | 1.29 | 10.7% |
| Leicester City | 86 | 154 | 110 | 71.4% | 1.79 | 33.7% |
| Bournemouth | 87 | 146 | 103 | 70.5% | 1.68 | 31.0% |
| Newcastle United | 115 | 216 | 149 | 69.0% | 1.88 | 37.4% |
| Southampton | 90 | 127 | 87 | 68.5% | 1.41 | 18.9% |
| Nottingham Forest | 55 | 97 | 66 | 68.0% | 1.76 | 29.1% |
| Wolverhampton Wanderers | 104 | 159 | 108 | 67.9% | 1.53 | 28.8% |
| Arsenal | 114 | 217 | 147 | 67.7% | 1.90 | 50.9% |
| Fulham | 85 | 139 | 94 | 67.6% | 1.64 | 35.3% |
| West Ham United | 127 | 225 | 146 | 64.9% | 1.77 | 33.9% |
| Brentford | 93 | 205 | 132 | 64.4% | 2.20 | 36.6% |
| Crystal Palace | 105 | 198 | 123 | 62.1% | 1.89 | 27.6% |
| Leeds United | 66 | 113 | 70 | 61.9% | 1.71 | 30.3% |
| Everton | 107 | 204 | 125 | 61.3% | 1.91 | 29.0% |
| Tottenham Hotspur | 131 | 275 | 163 | 59.3% | 2.10 | 38.2% |
| Sheffield United | 45 | 66 | 39 | 59.1% | 1.47 | 13.3% |
| Burnley | 77 | 108 | 62 | 57.4% | 1.40 | 16.9% |
| Watford | 33 | 61 | 35 | 57.4% | 1.85 | 24.2% |
| Manchester United | 131 | 270 | 153 | 56.7% | 2.06 | 46.6% |
| Manchester City | 116 | 316 | 179 | 56.6% | 2.72 | 59.5% |
| Chelsea | 127 | 287 | 161 | 56.1% | 2.26 | 43.3% |
| Brighton and Hove Albion | 129 | 250 | 137 | 54.8% | 1.94 | 31.8% |
| Liverpool | 131 | 341 | 185 | 54.3% | 2.60 | 53.4% |
| Aston Villa | 122 | 251 | 136 | 54.2% | 2.06 | 39.3% |
| Norwich City | 23 | 32 | 17 | 53.1% | 1.39 | 17.4% |
| Ipswich Town | 21 | 37 | 19 | 51.4% | 1.76 | 14.3% |
⚠️ Données grosses occasions disponibles pour ~53% des matchs. Luton Town (28 matchs) a un taux de conversion anormalement élevé — traiter comme valeur aberrante (petit échantillon).
🔑 Conclusions Clés
- Early PL goals are the most volatile, not the most decisive — when the first goal arrives before 30', the cascade risk is 32.1%: the trailing team wins outright 1 in 3 times. This is the Open Game regime. The "score first, win first" narrative simply doesn't hold in the data.
- After 60', matches lock in — the Locked Game regime produces only 11.9% cascade risk. 63.1% of such matches end as draws. Late opening goals don't win games; they negotiate the result. Scoring after 75' locks in 76.9% draws.
- Goals are back-loaded — 22.6% of all PL goals arrive in the final 15 minutes (76-90'), but these are predominantly equalizers and late winners in already-decided matches. The 86-90' window alone accounts for 11.9% of all goals.
- Man City exploits Open Game regimes better than anyone — 50.9% comeback rate when trailing (29/57 games), highest in the league. Their early-scoring tendency (avg min 47.8) means they frequently operate in Open Game conditions where cascades are structurally probable.
- Brentford lives in the Locked regime — 27.4% of their goals arrive in 76-90' (highest in the league), avg min 52.2. Their scoring pattern means matches where Brentford lead are structurally unlikely to cascade. Sheffield United (11.5% comeback rate when trailing) confirms the same pattern from the defensive side.
- Big chance volume dominates efficiency at scale — Man City creates 2.72 BC/match (highest) at 56.6% conversion and wins 59.5%. Brighton creates 1.94 BC/match at 54.8% conversion and wins 31.8%. The gap isn't conversion rate — it's volume. Creating more big chances per match is the structural advantage, not finishing them at a higher rate.
Méthodologie
- Source: premierleague.com commentary JSON — 2,464 matches, GW1–GW38, seasons 2019–20 through 2024–25.
- Goal detection: events where comment starts with 'Goal!' or event type contains 'goal'. Own goals excluded from first-scorer logic; penalties flagged separately.
- Stoppage time normalised: 45+2 → 45, 90+4 → 90 for phase/bucket classification. Actual minute preserved for per-team profiling.
- First-goal effect: 1,300 matches with valid final scores (homeTeam.score / awayTeam.score fields). Outcome classified from perspective of first-scoring team.
- Comeback teams: computed from 'fell behind' perspective — team that conceded first, then won. Min 20 matches to qualify.
- Big chance data (Opta team_stats.json): bigChanceCreated, bigChanceScored available for ~53% of the dataset. Conversion rate = bigChanceScored ÷ bigChanceCreated. bigChanceMissed tracked separately (Opta counts independently). Min 20 matches per team.
- Team goal timing: goals from open play + penalties (excluding own goals) assigned to scoring team via 'Goal! TeamName N,' regex. Per-team stats use base minute (clamped stoppage).