Waleed Shour: how often he scans before the ball arrives

How often he looks around before the ball reaches him, how his body is turned when it arrives, and whether that decides what happens next. Built from SkillCorner body pose (29 joints, 25 fps) joined to our event data for all 58 of his receptions.

One match. Treat every split below as a first look, not a settled number.

ONE MATCH, 58 RECEPTIONS WITH BODY POSE

How many times does he scan before receiving?

"How many times does he look around before the ball reaches him?"

Across the 38 receptions where the pose covers at least three quarters of the 10 seconds before the ball arrives, he makes 10.2 scans on average (median 9), and 3.8 of them come in the last 3 seconds. Among the 26 players in this match with 10 or more receptions, that ranks him 2nd.

SCANS IN THE 10 SECONDS BEFORE EACH RECEPTION

One column per number of scans. A scan is one distinct head-turn of at least 30°, measured against the direction of the ball so that a head simply following the ball counts as nothing.

WHEN THE SCANS HAPPEN

Every scan by every reception, stacked on a shared clock that ends at the moment the ball reaches him. The pass that feeds him is marked where it is played.

SCAN RATE

How his body is turned when the ball arrives

"Is he open when he gets it, or does he have his back to the play?"

His shoulders sit a median 68° away from the goal he is attacking when the ball reaches him: 27 receptions open, 17 half-open, 14 with his back to it.

WHERE HE RECEIVES, AND HOW HE IS TURNED

Every reception, drawn where it happened, with a tick showing which way his shoulders point. Attacking upwards.

OPENNESS AT THE MOMENT THE BALL ARRIVES

AND HOW FAR HE GETS ROUND IN THE NEXT 2.5 S

Openness is the angle between his shoulders and the goal he is attacking. 0° is square to that goal, 180° is his back to it.

Does his body position change what happens next?

"Does his body position influence whether he turns forward afterwards?"

Yes, clearly. Received open he plays forward 52% of the time (14 of 27). Received closed, 14% (2 of 14), and the median ball from a closed reception travels -6.3 m up the pitch, which is to say it goes backwards.

HOW HE WAS TURNED, AND WHETHER THE BALL WENT FORWARD

Forward = his carry or his release pass moved the ball more than 5 m up the pitch. Small cells are labelled with their count, not dressed up as a rate.

AND DOES SCANNING MORE CHANGE IT?

The same question asked of the scan count instead of the body position.

The moments, frame by frame

Every reception, replayed from ten seconds before the ball arrives. The blue cone is where his head is pointing, the green one his shoulders, and each orange tick on the timeline is a scan, labelled with how far his head travelled.

Everyone else on the pitch

The same three measures for every player in the match with at least 10 receptions that carry body pose, both teams.

Every player with at least 10 receptions that have body pose, both teams, same definitions. One match, so treat the order as indicative.

How this was built

Three sources on one clock. Our event data for this match, SkillCorner's open broadcast tracking (10 fps) and SkillCorner's body pose (25 fps, 29 joints per player). The clocks were aligned per half by searching the offset that minimises the distance between each event's own coordinates and the passer's tracked position: an offset of 2.3 s in the first half and 1.95 s in the second, leaving a median error of 2.63 m and 2.67 m across 3,136 events. Pose joins to tracking at 2.5 pose frames per tracking frame.

Facing. Shoulder yaw is the perpendicular to the line from his left to his right shoulder, head yaw the same from his ears. Sanity check: when he runs above 3.5 m/s his shoulders sit within 45° of his running direction 97% of the time, so the left/right sign is right.

A scan is one distinct head movement of at least 30°, separated from the next by a reversal of at least 10°, measured against the bearing to the ball so that a head simply tracking a moving ball counts as nothing. The detector was calibrated against SkillCorner's own reference animation of phase 406 of this match, which reports 7 scans at 34.0 per minute; on the same frames this finds 7 at 34.1 per minute.

What to distrust. This is one match: 58 receptions, and the three-way splits land in cells of 10 to 27. Pose exists for 55% of his frames, because it needs him visible in the broadcast picture, so scan counts are only quoted over the 38 receptions whose ten-second window is at least three quarters covered. Openness needs one frame rather than a window, so all 58 receptions are used for it. The direction of the body-position finding is strong; the exact percentages are not.