Maarten Paes watched about 50 Sébastien Haller penalties because the save often starts before the shot. The run-up helps a goalkeeper pick a side early; the ball flight usually tells him whether his hands go high or low.

In De Telegraaf’s Kick-off, Mike Verweij said Haller’s pause was the detail Paes used, according to Voetbalzone’s May 26th report. Haller stops in his run-up, looks at what the goalkeeper does, and then chooses his shot; Paes used that pause before saving the penalty.

Paes’s own account narrows and confirms the scouting detail. Voetbal International’s interview with Paes quotes him saying he studies this work for every match, gets a list of penalties, free kicks and player focus, reviews it with his goalkeeper coach three days before the game, and looks again before the match. On Haller, Paes said you can watch about 50 of his penalties and he almost always scores, so Paes needed to invent something; he knew Haller pauses and used a fake movement. Ajax Showtime’s ESPN quotes add that Paes tried saying things to Haller, knew he always makes that stop, made “a kind of fake movement,” and waited for Haller to look down and shoot.

Voetbalzone, VI and Ajax Showtime support Paes’s preparation and explanation; they do not provide a frame-by-frame proof of the exact visual trigger. Paes’s football read is still clear enough from those reports: Haller had a known pause, checked the goalkeeper, and could be nudged by a small movement while the strike path was still hidden.

Why the side is chosen early

Noël, van der Kamp and Klatt’s 2021 Frontiers study puts the time problem in match footage: 395 shootout penalties from FIFA World Cups and UEFA European Championships between 1984 and 2016, with ball flight times sometimes 500 ms or less. The study classified goalkeepers starting 200 ms or more before contact as early responders and those starting 160 ms before contact or later as late responders.

A keeper who waits sees more of the strike, but gives up distance to the ball. Paes’s fake movement made sense because Haller’s pause came before contact, where the keeper still had some time to influence the taker and start his own dive.

The run-up cues are in the body

Causer, Smeeton and Williams’ 2017 PLOS One study tested 12 skilled and 12 less-skilled male goalkeepers with 48 penalties filmed from the goalkeeper’s view and occluded at 160 ms before contact, 80 ms before contact and ball contact. The skilled group averaged 85% correct direction, compared with 67% for the less-skilled group, and skilled keepers stayed above chance in the hip-region-only condition across all occlusion points.

Lower-body information becomes stronger as contact approaches. The same study found later occlusion improved correct response and correct direction, while correct height improved less. Savelsbergh et al.’s 2005 research record says successful expert goalkeepers waited longer before initiating a response and appeared to fixate longer on the non-kicking leg. Hip region, plant foot and kicking-leg movement work as a cue family, and deception can still corrupt any one signal.

Side first, height later

Left-right decisions and high-low adjustments are timed differently. Higueras-Herbada and colleagues’ 2020 Frontiers paper found that goalkeepers’ horizontal hand positions already differed for left and right penalties about 115 ms before average ball contact, while their vertical hand positions for high and low penalties diverged about 245 ms after contact. The hand-timing analysis used 88 penalties with an average ball flight of 525 ms.

Pre-contact kinematics were much clearer for horizontal direction than height. The same paper says earlier horizontal work found correlations above 0.8 for dominant-foot angle, hip angle and dominant-foot movement direction, while the 2020 height analysis found individual kinematic correlations never above about 0.3 and explaining less than 10% of height variance. A goalkeeper can start the dive side from the run-up and still use the first ball flight to lift or drop the hands.

Scouting changes probabilities

In the same 2021 Frontiers study, a keeper-independent taker had chosen the corner before the run-up, while a keeper-dependent taker waited for the goalkeeper to commit. Keeper-independent takers scored 60.00% when goalkeepers moved early and 79.69% when they moved late; keeper-dependent takers scored 81.81% against early movement and 63.63% against late movement. The same early move helps against one taker profile and hurts against the other.

Luis-del Campo et al.’s 2024 Frontiers case report first summarizes prior work that football goalkeepers often begin moving between 50 ms and 350 ms before contact, then gives a small training example. The case used eight sessions and 80 kicks with one U10 novice goalkeeper; in four training sessions, the coach told him that eight of the next 10 penalties would go to one side. His mean movement onset shifted from 145.50 ms before contact at baseline to 9 ms after contact in retention, and performance rose from 2.60 to 3.45 on a five-point scale, although the performance difference was not statistically significant. The evidence is narrow because it comes from one goalkeeper; it supports timing adaptation rather than a general increase in save rate.

What to tag on video

Pinheiro et al.’s 2022 OpenPose study shows how video can make some of these reads measurable. It analyzed 34 elite penalties from major European competitions between 2017 and 2020 with one standardized viewing angle, measured body orientation and goalkeeper anticipation, and found a mean OpenPose confidence score of 0.80 ± 0.14 with retest reliability of 0.976 ± 0.03. Adding goalkeeper anticipation increased the goalkeeper-strategy model’s classification accuracy to 97.0%. The authors also describe the study as preliminary and note the small sample and single viewing angle.

Coaches tagging an opponent’s penalties should log the last pause in the run-up, whether the taker looks at the goalkeeper, the final plant-foot angle, the hip and kicking-foot direction in the last 160 ms, and the goalkeeper’s movement onset relative to contact. The goalkeeper should leave the meeting with one instruction: wait through the pause, show a small fake and still attack the planned corner, or start earlier because the taker rarely checks the goalkeeper.

Paes’s Haller save fits the research because the run-up gave him a cue before the shot did. The clip review should stop at the pause, the plant foot and the first step of the dive: can the keeper affect the taker’s choice and still keep enough hand movement for the ball’s height?