Serie A · Manager Maurizio Sarri · Nominal 4-3-3 (Opta team sheet)
Sample: last 5 league matches — v Sassuolo 2-1, v Bologna 1-0, @ Roma 1-2, v Cagliari 1-2, @ Juventus 0-2
Build-up3-3-1
Progression3-6
Final third1-3-4
Rest defence4+2 / 3+4
In possession
attack → · shapes are phase snapshots, not formations
IP · 01
Build-up
3-3-1
MEDIUMSTRONGLY INFERRED4-man first line in 2/5 · per match: 4-3, 3-4-1, 4-4-1, 3-2-1, 1-5-1
First line: Kolasinac, Scalvini and Kossounou; Éderson, Gaetano and Bellanova next.
Own-third passes / match 144.8long 10%goal kicks short 43%
IP · 02
Progression
3-6
LOWINFERRED3-man first line in 2/5 · per match: 3-6, 5-4, 3-4-2, 2-7, 2-3-3
6 on the highest line (Éderson, Kessie, Gaetano, Samardzic, De Ketelaere and Bellanova).
Middle-third passes / match 246.2into final third / match 35.4
IP · 03
Final third
1-3-4 · 3 box
LOWINFERRED7-man first line in 2/5 · per match: 2-3-3, 7, 1-4-2, 1-5-3, 7-2
3/5 lanes occupied, width from Bellanova. Box presence: De Ketelaere, Scamacca and Samardzic.
Box touches / match 22.8crosses 10.2shots 11.8
IP · 04
Rest defence
4+2 / 3+4
LOWSTRONGLY INFERRED4+2 in 1/5, 3+4 in 1/5 · consistent across game states (max avg shift 6.6m)
Across 116 sustained final-third attacks, Scalvini, Kossounou, Kolasinac, Éderson, Gaetano and Kessie took the ball outside the final third (pooled 3+3). Per match: 4+2, 3+4, 3+1, 2, 3+2 — 6 behind the ball in 1/5.
Opp. fast-break shots / match 0.6opp. shots 19.2
Out of possession
positions = where each player makes his defensive actions
First engagements high: Samardzic 19%, Éderson 17%, Bellanova 14% of the team's actions in the final 40%. Line structure: insufficient evidence (defensive actions sit where the ball is; no tracking).
PPDA per match 19 / 33 / 27 / 16 / 16high regains / match 8.6actions in final 40% 17%
Deep defending led by Scalvini 19%, Kossounou 17%, Bellanova 13%. Formation: insufficient evidence.
Def. actions in own third 57%per match 57.4
OOP · 04
Rest offence
1 HIGH
MEDIUMMEASURED17 outlet balls in the sample
Outlet after a deep regain: Rowe 17% of the balls played out beyond halfway.
Deep regains / match 28.4played beyond halfway ≤3 passes 12%
Transitions
the first seconds after the ball changes hands
TR · 01
Defensive transition
HYBRID
HIGHINFERREDpooled 244 losses · mode agrees in 4/5 matches
Hybrid (inferred from thresholds): 34% of losses won back within 5 s, opponent reaches their half within 10 s after 20%. Per match: hybrid, hybrid, retreat-first, hybrid, hybrid.
Losses (front 60%) / match 48.8won back ≤5s 34%tactical foul ≤5s 2%
TR · 02
Attacking transition
SECURE-FIRST
LOWINFERREDpooled 213 regains · style agrees in 2/5 matches
Secure-first (inferred): first pass forward 28%, long 6%. Forward receptions: De Ketelaere, Éderson and Bellanova. Per match: direct, secure-first, controlled, secure-first, controlled.
Regains → final third ≤10s 12%shot ≤15s 4%
Involved — median spot of his actions in the phaseOccupation zone (25th–75th pct)Rarely involved — overall positionForward receptionsLoss not won backLoss won back ≤5s
Tactical phases · Atalanta1 / 2
Tactical phases · 2026/2027
PHASES OF PLAY
Set pieces · Tactical summary
Atalanta
Sample: last 5 league matches — v Sassuolo 2-1, v Bologna 1-0, @ Roma 1-2, v Cagliari 1-2, @ Juventus 0-2
Set pieces
goal at the top · corners shown from the right
SP · 01
Corners for
FAR POST
LOWMEASURED
Deliveries: near post 29% · centre 29% · far post 43%; 0% played short.
Corners / match 2.8shots from corners 4goals 0
SP · 02
Corners against
1ST CONTACT 35%
MEDIUMMEASUREDmarking scheme: UNVERIFIED
Win the first contact on 35% of corners against. Event data can't show the marking scheme — the dots show where they clear it.
Corners against / match 5.2shots conceded / match 3.0
SP · 03
Free kicks for
CROSS
LOWMEASURED
Of 10 free kicks in the final 40%: 4 crossed, 4 short, 0 shot direct.
Set-piece shots / match 1.6from these FKs 1
SP · 04
Free kicks against
8 FACED
LOWMEASURED
8 opposition free kicks in the final 40%; 4 led to a shot within 12 s.
Set-piece shots conceded / match 3.8
Involved — median spot of his actions in the phaseOccupation zone (25th–75th pct)Rarely involved — overall positionForward receptionsDelivery → shot / contact wonNo shotOpp. delivery
Tactical summary
phase map · confidence · evidence
Phase
Structure
Recurrence
Confidence
Evidence
Game state
Build-up
3-3-1
4-man first line 2/5
MEDIUM
STRONGLY INFERRED
consistent across game states (max avg shift 5.5m)
Progression
3-6
3-man first line 2/5
LOW
INFERRED
consistent across game states (max avg shift 6.9m)
Final third
1-3-4
7-man first line 2/5
LOW
INFERRED
consistent across game states (max avg shift 7.4m)
Rest defence
4+2 / 3+4
6 behind in 1/5
LOW
STRONGLY INFERRED
consistent across game states (max avg shift 6.6m)
High press
PPDA 21.3 · formation n/a
pooled
LOW
MEASURED
not tested
Mid block
insufficient evidence
pooled
LOW
MEASURED
not tested
Low block
insufficient evidence
pooled
LOW
MEASURED
not tested
Rest offence
1 high
17 outlet balls
MEDIUM
MEASURED
not tested
Defensive transition
hybrid
4/5 matches agree
HIGH
INFERRED
not tested
Attacking transition
secure-first
2/5 matches agree
LOW
INFERRED
not tested
Set pieces
corners mostly far post
14 corners
LOW
MEASURED
not tested
Match sample
opponent context from the Opta feed
Date
Opponent
H/A
Score
Pass share
Nominal (Opta)
Opp. PPDA
Reds for/against
Game-state minutes
2026-08-23
Sassuolo
H
2-1
56%
4-3-3 v 4-3-3
22.9
0/0
level 18′, leading 78′
2026-08-31
Bologna
H
1-0
42%
4-3-3 v 4-2-3-1
14.7
0/0
level 98′, leading 0′
2026-09-05
Roma
A
1-2
34%
4-3-3 v 3-4-1-2
8.8
1/0
level 54′, leading 42′, trailing 2′
2026-09-12
Cagliari
H
1-2
61%
4-3-3 v 4-4-1-1
25.3
0/0
level 40′, trailing 57′
2026-09-20
Juventus
A
0-2
48%
4-3-3 v 3-4-2-1
22.2
0/0
level 28′, trailing 67′
Limitations
No tracking / freeze-frame data: positions are where players take part on the ball, not off-ball spacing.
Out-of-possession formations cannot be read from event data (actions sit at the ball) — reported as insufficient evidence.
5 matches: game-state splits compare player displacement; late-game (75'+) splits are too small to report.
Corner in/out-swing and marking schemes are not in the feed (UNVERIFIED).
Transition labels are threshold classifications calibrated on a small set of teams.
Sources & method
Opta event data (every on-ball event, x/y, qualifiers)
Match data: fixture list, crest, club colour, manager
Shapes = median on-ball location per player per phase; lines by optimal 1-D k-means
Confidence = agreement of single matches with the pooled structure + sample size
