ECSS Paris 2023: OP-AP17
INTRODUCTION: Sprint distance is a key indicator of high external load in football [1], closely linked to injury risk and decisive match events [2]. However, commonly used fixed absolute thresholds misrepresent individual sprint exposure [3]. Normalizing sprint detection to maximal sprinting speed (MSS) improves individualization but may underestimate sprint exposure by excluding its acceleration phase [4]. Incorporating individual acceleration–speed profiles (ASP) may address this limitation by additionally accounting for accelerated sprinting [5]. However, it remains unclear whether these methods yield significantly different sprint exposure metrics during match play. Therefore, this study compared absolute, MSS-normalized, ASP-based, and combined thresholds (ASP+MSS) for quantifying sprint distance and sprint events during elite football matches. METHODS: Eighteen elite male football players from a club competing in the Slovenian First Division and UEFA Conference League were monitored across 158 match observations (3–16 matches per player; >80 min played) during the 2025/2026 season. Speed, acceleration, and distance were collected using a GNSS device. MSS and ASP were defined using rolling match maximum values. Sprint distance and sprint events were calculated using: (a) absolute threshold >25 km·h-1 (ABS); (b) speed >90% MSS (N-MSS); (c) acceleration >90% ASP at lower speeds (N-ASP); and (d) a combined MSS+ASP approach (COM). Values were averaged across matches per player. Methods were compared using RM-ANOVA with Bonferroni post-hoc tests, and associations were assessed using Pearson’s r. RESULTS: ABS, N-MSS, N-ASP, and COM calculated 114.7±44.7, 24.7±8.1, 46.5±12.8, and 63.9±12.8 m of sprint distance, and detected 9.2±2.9, 2.4±0.7, 14.5±3.4, and 15.2±3.3 sprint events, respectively. A significant effect of method was found for sprint distance (F=56.8; p<0.001) and sprint events (F=118.4; p<0.001). ABS yielded greater sprint distance than N-MSS (MD=91.4 m), N-ASP (MD=50.4 m), and COM (MD=50.4 m). N-MSS showed lower sprint distance compared to N-ASP (MD=-21.8 m) and COM (MD=-39.1 m). Sprint events were fewer with N-MSS than ABS (MD=-6.9), N-ASP (MD=-12.1) and COM (MD=-12.8). Conversely, N-ASP and COM detected more sprint events than ABS (MD=5.3 and MD=5.9). Sprint distance correlated between ABS, N-MSS and COM (r=0.51–0.66), whereas N-ASP correlated only with COM (r=0.83). Sprint events correlated between ABS and N-MSS (r=0.61) and between N-ASP and COM (r=0.99). CONCLUSION: Absolute thresholds substantially overestimate sprint distance, whereas MSS-normalized thresholds may underestimate sprint exposure. Incorporating ASP (N-ASP or COM) identifies additional sprint distance and sprint events and may provide a more meaningful representation of sprint-related external load for individualized monitoring and injury-risk management. [1] Gualtieri 2023; [2] Vermeulen 2024; [3] Pimenta 2024; [4] Pimenta 2025; [5] Miguens 2024
Read CV Matic SašekECSS Paris 2023: OP-AP17
INTRODUCTION: Corner kicks have transitioned from routine set-pieces into high-value attacking opportunities as modern defenses increasingly compress open-play. Despite their tactical importance, traditional performance analyses primarily rely on basic event statistics and overlook the rich spatiotemporal information embedded in player and ball tracking data. This study addresses the gap by developing machine learning model to predict corner kick threat, defined as the likelihood of a subsequent shot, and examining key spatiotemporal determinants that distinguish Shot and No-Shot outcomes. METHODS: Utilizing StatsBomb event and tracking data, a total of 3,025 corner kick sequences from 780 matches across the Premier League (2022-2024), Bundesliga (2023/24), World Cup (2022), and UEFA Euro (2020/2024) were extracted, incorporating tracking frames at the moments of delivery and reception. Samples were categorized into Shot and No-Shot groups based on the sequence outcome. Twenty-one metrics were constructed across four dimensions: (i) delivery quality, (ii) collective tactical formation, (iii) recipient opportunity, and (iv) goalkeeper positioning. A two-stage analytical framework was employed: first, Mann-Whitney U tests identified significant feature differences between outcome groups; second, five machine learning models were trained, optimized using Optuna Bayesian optimization and refined via threshold-moving techniques. RESULTS: Significant spatiotemporal differences were observed between Shot and No-Shot sequences (effect sizes reported as non-parametric r). The Shot group exhibited significantly higher recipient space quality (45.99±41.81 vs 32.81±49.66, p<0.001,r=0.34), despite facing higher defense density (5.92±2.83 vs 5.10±3.08, p<0.001, r=0.15) and closer proximity to the nearest defender (2.10±2.97 vs 2.45±4.41, p<0.001, r=0.17).Shot corners featured longer pass lengths (37.39±10.70 vs 35.02±13.15, p<0.001, r=0.12) against a more compact defensive shape, indicated by lower defensive average pairwise distance (7.57±1.70 vs 8.05±2.09, p<0.001, r=0.12). Among all models tested, the optimized LightGBM achieved the highest predictive performance (AUC=0.80, Accuracy=76.1%, Precision=0.84), with recipient space quality and tactical combination strategy emerging as the most influential predictors. CONCLUSION: The findings indicate that recipient space quality, defensive intensity, and tactical combination strategies are pivotal determinants of corner kick shots. High-threat corners are not merely those with open space, but those that create high-quality space under congestion. Practitioners should therefore emphasize strategies that maximize spatial advantage in tight defensive zones and refine delivery precision to enhance shot conversion rates.
Read CV Haiyu LiuECSS Paris 2023: OP-AP17
INTRODUCTION: In invasion games, players spend a large proportion of playing time without ball possession. Thus, off-the-ball movement, such as creating and exploiting space, is as important as ball-related skills. However, off-the-ball movement has mainly been assessed qualitatively, and quantitative methods remain insufficiently established. Previous studies developed the Game Test Situation (GTS “off-the-ball movement”), a task designed to elicit tactical behaviors common to invasion games. However, GTS has been used mainly for qualitative assessment, and its quantitative application has been limited. Moreover, the influence of task constraints, especially goal-oriented objectives, is understudied. This study aimed to examine task conditions suitable for quantitative assessment of off-the-ball movement. METHODS: Nineteen university students performed the GTS (three-on-three passing game) under two rule conditions with different winning criteria (rule A: number of successful passes; rule B: number of goals scored). Twelve games per condition were conducted with varying team configurations. Off-the-ball movement was evaluated from video recordings by GTS experts and classified as Good, Average, or Poor. Participants wore trunk-mounted accelerometers and Ultra-Wideband (UWB) positioning tags. Transfer entropy, an information-theoretic measure, was calculated from accelerometer data to estimate causal relationships between players’ movements. Based on this analysis, eight network-based indices were derived to quantify the amount and direction of information flow (incoming or outgoing) as well as its source (teammates or opponents). In addition, positional data obtained from the UWB tracking system were used to calculate five positional metrics: total distance, mean speed, speed SD, mean acceleration, and acceleration SD. Due to the small sample size of the Poor group, quantitative data were compared between the Good and Average groups. RESULTS: Under rule A, only the All-Path index was higher in the Good group (p < .01), while no significant differences were observed in the positional metrics. In contrast, under rule B, the In-Path and All-Path indices were higher in the Good group (p < .01), as were the InOut-Path, TeamIn-Path, and OppIn-Path indices (p < .05). Conversely, total distance, mean speed, speed SD, and acceleration SD were lower in the Good group (p < .05). CONCLUSION: Under the goal-oriented condition (rule B), the network indices indicated that Good players adjusted their movements in response to others, reflecting greater information inflow from teammates and opponents. Positional metrics suggested that these players exhibited more selective movement patterns, with lower total distance and reduced variability in speed and acceleration. Quantitative differences between qualitative groups were clearer under rule B. These findings indicate that goal-oriented game tasks are more suitable for quantitative assessment of off-the-ball movement in invasion games.
Read CV Rintaro HamamuraECSS Paris 2023: OP-AP17