This study examined the differences in passing network between two- and three-center-back (2-CB and 3-CB) formations and their predictive relevance for match outcomes using machine learning models. The dataset comprised 256 matches (7328 player observations) from FIFA Soccer World Cups from 2010-2022. A novel subgroup entropy metric was introduced to quantify the distribution of ball circulation within positional units (defenders, midfielders and forwards). Results showed that conventional network metrics failed to distinguish formation-specific features, whereas the subgroup entropy metric revealed significant differences: 2-CB formations exhibited higher forward- and defender-entropy but lower midfielder-entropy compared with 3-CB formations (p < 0.001). Among the tested models, XGBoost achieved the best performance (accuracy = 0.599 ± 0.040, F1 = 0.737 ± 0.033). SHAP-based explainability analysis indicated that the key network metrics associated with match success varied by formation: midfielder-entropy and forward-entropy were most influential in 2-CB formations, while average weighted degree, average path length, and Density dominated in 3-CB formations. These findings suggest that different formations shape distinctive patterns of passing flow and team coordination. The proposed subgroup entropy provides a novel framework for linking network structure, tactical organization, and match outcomes in football analytics.

This study examines positional differences in running performance during the 2025 FIFA Club World Cup, focusing on the moderating effects of team strength and confederation. The sample comprised 63 matches. Eight running performance variables were analysed, including total distance (TD), distances at five speed zones (DZ1-DZ5), number of sprints, and maximum speed. One-way and two-way ANOVAs were conducted to examine the effects of player position, confederation, and team strength on running performance. Players position significantly influenced all running performance variables. Significant position and confederation interactions were found in DZ1, DZ2-DZ5 and sprints (p < 0.05). For instance, CONMEBOL players exhibited higher DZ1 values across most positions (p < 0.05), while UEFA players had higher DZ4 in central backs (p < 0.05), but lower sprint frequency in wide midfielders (p < 0.01). Significant position and team strength interactions were observed in DZ3-DZ5 and sprints (p < 0.05). Top-teams had greater sprint distances and frequencies in defensive positions, whereas wide midfielders from low-teams performed more high-speed actions than those from top-teams. Running performance in elite football is shaped not only by positional roles but also by confederation and team strength. Coaches can design targeted training based on the physical characteristics of the opponents to improve the players' running performance during the game.

This study examined whether teams using identical formations display similar playing styles by integrating Phase of Play (PoP) indicators with clustering and explainable machine learning. Ninety-three observations from teams employing a 4-3-3 formation in the FIFA World Cup 2022 (n = 58) and FIFA Club World Cup 2025 (n = 35) were analyzed. K-means clustering identified distinct styles, while XGBoost with SHapley Additive exPlanations (SHAP) interpreted key differentiators. Two contrasting styles emerged despite identical formations: a possession-oriented and a defending-oriented style. SHAP revealed attacking transition and both opposed and unopposed build-up as decisive in-possession indicators, and low block and high press as main out-of-possession factors. The possession-oriented cluster was characterized by structured build-up play, higher possession, total passes, pass completion ( p < 0.001, large), and more defensive line breaks ( p < 0.01, moderate). The defending-oriented cluster demonstrated deeper defensive organization, with greater low-block, more pushing-on-into-pressing actions and defensive pressures ( p < 0.001, moderate), and longer ball recovery time ( p < 0.001, large). These findings show that teams sharing the same formation can differ substantially in style, emphasizing the limitations of formation-based analysis and the value of PoP indicators for understanding functional behaviors in elite soccer.

The Football Video Support (FVS) system was officially introduced into football in 2025 to improve the referees’ officiating accuracy. This study aims to investigate the influence of FVS on elite football matches compared to Video Assistant Referee (VAR). This study compared 52 matches played in the 2023 FIFA U20 World Cup with VAR implementation and 52 matches played in the 2025 FIFA U20 World Cup following the introduction of FVS using Generalized Linear Model (GLM) and means comparison. Eleven variables were gathered for each game: first half match time, second half match time, full match time, fouls, penalties, goals, yellow cards, red cards, offsides, corners and free kicks. The results showed that after the introduction of FVS, the total duration of the first half, the total duration of the second half, the overall match duration, and the number of corners all decreased significantly. Notably, the reductions in match time were less pronounced in the first two rounds of the group stage than in the final round, suggesting that the effect of FVS on match time may vary according to match round, which may offer guidance for future investigations into FVS and VAR. The findings may help football practitioners and fans better understand the effect of FVS on elite football matches and the key differences between FVS and VAR, and inform FVS related regulations.

Using machine learning, the technical effectiveness of table tennis players and their winning probability were modeled. This study adopted a novel algorithm, SHapley Additive exPlanation (SHAP), to analyze the important features based on the gradient boosting + categorical features-tree-structured parzen estimator (Catboost-TPE) with the four-phase evaluation theory, an analysis framework in a table tennis match. A total of 110 singles’ matches (9536 rallies) were analyzed, and 59 male players’ winning rates from 2018 to 2022 were categorized into three levels (high, medium, low) by k-means cluster analysis. The results showed that Catboost-TPE has the best performance (MSE = 7.5e-05, MAE = 0.006, RMSE = 0.008, R 2 = 0.99, and adjusted R 2 = 0.989) among six hybrid machine learning algorithms. Using Catboost-TPE to calculate the SHAP value of each feature, the global interpretation and multiple local interpretations found that the performance of receive-attack and serve-attack phases have essential impacts on the winning probabilities in current matches. Besides, four and three phase evaluation theory are all import framework in table tennis match analysis. To further deepen the theoretical and applied value of the four-phase evaluation theory, this study derived the mathematical equations for converting indicators from the four-phase evaluation theory into the new three-phase evaluation theory. These results provided quantitative references to table tennis matches’ characteristics and winning phases. These methods used in the study can be widely applied to other sports performance analyses, and the equations derived in this study are also instructive for relative sports.

This study investigated positional differences in physical capacities and match running performance, and explored how physical capacities influence match running metrics in elite U17 football players. Data were collected from 212 players across 12 teams and 23 competitive matches, classified into five positions: central defenders (CD), full backs (FB), central midfielders (CM), wingers (W), and central forwards (FW). All the players come from the same league level. Analysis using one-way ANOVA and multiple regression revealed significant positional distinctions. FW and CD showed the highest height and weight, while W were significantly shorter and lighter (p < 0.01). CD achieved the best countermovement jump scores, while W outperformed others in 30 m speed, agility, and the Yo-Yo Intermittent Recovery Test Level 1 (YYIR1), indicating superior speed and endurance. Conversely, FW had the lowest YYIR1 performance (p < 0.05). Match data showed CM covered the most total, mid-speed, and quick-speed running distances (p < 0.001), while W recorded the highest maximum speed, high-speed running, and sprint distances (p < 0.001). Regression analysis identified 30 m speed as the main predictor of high-speed running, sprint distance, and maximum speed (p < 0.05), while YYIR1 predicted running performance in worst-case scenarios (p < 0.05). Overall, the findings emphasize the importance of position-specific physical training, focusing on speed and agility for attacking players and endurance development for central midfielders and forwards.

This study investigated the impact of winning determinants in two professional soccer leagues. The sample was composed of 1,440 Chinese Super Football League (CSL) and Chinese Football Association China League (CFACL) matches (CSL matches = 720; CFACL matches = 720) during the 2017–2019 seasons. The study employed eXtreme Gradient Boosting (XGBoost) to assess the importance of 25 indicators exhibiting significant differences (p < 0.05) in their association with match outcomes, and the SHapley Additive explanations (SHAP) was utilized to interpret these findings. The results showed that scoring performance indicators, such as Shots On Target Inside Box (SOTIB), Shots, and Shots On Target (SOT), significantly influenced outcomes in both the CSL ( S_G =37.854 S_G =38.934

Due to the dynamic and complex nature of soccer, match-running performance (MRP) is highly influenced by match content. This study aimed to examine the interaction between possession status (PS) and possession percentage (PP) in relation to match-running performance (MRP) and to quantify MRP in each PS while considering multiple contextual variables. MRP indicators, including total distance (TD) and high-intensity running distance (HID), were collected from 8,468 observations of 412 outfield male players in the 2018-2019 Spanish LaLiga, excluding matches with red cards. This study set PS, possession percentage (PP), effective playing time, match location, quality of opposition, and match results as fixed effects, and set players and teams as random effects. Results indicated: i) PP interacted with PS, negatively affecting TD (r =-0.26, p < 0.05) and HID (r =-0.11, p < 0.05) during IP but positively influencing TD (r = 0.24, p < 0.05) and HID (r = 0.28, p < 0.05) during OP; ii) MRP during in-possession exceeded out-of-possession when PP was below 36% for TD and 36.4% for HID; iii) PP thresholds for MRP shifts varied by position, with forwards requiring higher PP (TD: 61.8%, HID: 68.6%) compared to central defenders (TD: 28.3%, HID: 9.2%). This study reveals the interaction effects of PS and PP on MRP, emphasizing the complexity of multivariate relationships in soccer. It underscores the importance of multivariate approaches over traditional methods like t-tests, which provide only partial insights.

This study introduces two novel metrics within table tennis technical-tactical networks: technical decision-making style (TDS) and connecting technical style (CTS), inspired by the entropy concept, to quantify players' technical-tactical styles. This study proposes a multilayer technical-tactical network framework to capture interactive information and technique-to-technique confrontations between players. Additionally, we develop four new metrics—absorbing rate, releasing rate, stalemate rate, and usage rate—based on three states within table tennis matches: scoring, losing, and stalemate, to analyze inter-links within these networks. The champion table tennis player, who won gold medals in both the 2016 Rio Olympics and the 2020 Tokyo Olympics, and his opponents were analyzed in 5054 technical actions during these events. Metrics such as TDS, CTS, out-degree centrality (ODC), and in-degree centrality (IDC) within the networks were calculated. We also created datasets for TDS, CTS, ODC, and IDC and employed six machine learning algorithms for modeling. The results indicate that nodes utilizing TDS and CTS demonstrate superior predictive accuracy for game outcomes compared to those using ODC and IDC. SHAP analysis revealed the feature importance in the best-performing models for TDS and CTS, revealing non-linear relationships between the TDS and CTS values of each key node and game outcomes. The analysis of the multilayer network offers insights into the dynamic interactions between the champion player and his opponents, enhancing our understanding of the key factors influencing match victories. By integrating network science, entropy, and machine learning, this study presents a comprehensive framework and practical metrics for match analysis, with potential implications in performance analyses of other racket sports.

To determine the effective attacking patterns of corner kicks in women’s football, this study used associate rule to analyse the corner kicks in the Chinese Women’s Super League 2019. A total of 343 corner kicks from 56 matches were studied to examine the effective characteristics of corner kicks and strategies of strong and weak teams using associate rule analysis. In total, 17 goals were scored via corner kicks, accounting for 9.4% of the total goals (180) of the Women’s Super League in 2019. The result of associate rule analysis showed that the pattern of indirect, outside the box and eight attackers was more likely to create shots when a strong team (top 4 teams) faced a weak team (bottom 4 teams; Lift: 2.9). When a weak team faced a strong team, it was easier to create a shot by delivering the corner kick through an outswing ball directly to the near post (Lift: 6.091). These findings can help coaches to understand the key successful patterns of corner kicks in women’s football and choose a more effective attacking strategy of corner kicks according to the quality of both teams.

China has promoted campus soccer for over a decade due to its potential health benefits. The study aimed to explore soccer knowledge (SK), soccer attitude (SA), soccer practice (SP), and health status among Chinese freshmen and sophomore undergraduates who had received campus soccer education. Of the 7419 participants, 1,069 were valid and included in the analysis. Structural equation modeling (SEM) results indicated SK is positively associated with SA (p < 0.001), but negatively with SP (p < 0.01). SA was positively linked to SP (p < 0.001). SK indirectly affected SP through SA (Z = 13.677). Random forest-tree-structured Parzen estimators (RF-TPE) with SHAP indicated SP holds primary importance with a strong negative impact on health. Additionally, differences in rankings for SK, SA, and SP were observed among gender and urban-rural groups. These results reveal current campus soccer education is suboptimal to health promotion.

This study aimed to investigate the tactical factors influencing shooting success in the Chinese Football Association Super League (CSL). The data set comprised 5914 shooting actions collected from all 240 matches in the CSL 2018 season, featuring 16 different teams. The variables assessed included pass number (PN), playing minute (M), match status (MS), final passing area (FPA), regaining possession position (RPP), offense category (OC), and shooting position (SP). To distinguish the attacking patterns of teams of varying strengths, all 16 teams were clustered into two groups based on two contextual factors: their starting season budgets and their rankings at the end of the season. Firstly, a descriptive analysis was conducted, followed by the application of a binomial-logit model (p < 0.05). The Maximum Likelihood Estimation (MLE) method was employed to evaluate the factors impacting shooting success. The results revealed that: (1) Match status (p = 0.0412), regaining possession position (p = 0.0002), and shooting position (p = 0.0001) play important roles in shooting success for all teams. (2) For strong teams, the offense category (p = 0.0031) and the final passing area (p = 0.0027) also have an important influence on shooting success. (3) The tactics of the weak teams should focus especially on the final passing area (p = 0.0388) before shooting at the goal. These findings can provide guidance for coaches in developing more effective scoring strategies or defensive tactics and tailoring them appropriately to the strengths and weaknesses of their respective teams.

The purpose of this study was to analyse the influence of the new substitution option (NSO) on the Chinese Soccer Super League (CSL). After exclusion of matches included red cards, a total of 1129 match observations from 2018 to 2020 were analysed. Linear mixed models were built to analyse the impact of NSO on physical and technical performance. The study found that after the introduction of NSO: (i) substitution distributions changed, as substitution were made earlier by coaches than the past especially in the period from 46 to 75 min; (ii) although there was no significant changes on physical performance, technical performance showed significant changes especially in organizing performance and defending performance; (iii) specifically, teams used NSO showed significantly higher possession (p = 0.022), passes (p < 0.001), pass accuracy (p < 0.001), passes in attacking third (p = 0.004), and pass accuracy in attacking third (p < 0.001), whereas lower breakthroughs (p = 0.028), tackles (p = 0.002), clearances (p < 0.001), and yellow cards (p = 0.002). The findings of the study can help coaches, especially in CSL, to better understand NSO and improve decision-making when replacing players.

This study developed two action recognition models using the YOLOv8-Alphapose two-stream spatial temporal graph convolutional networks (2s-STGCN), and the networks were used to recognize technical actions in table tennis. This study proposed a novel framework that merges dynamic and static complex network analysis with a community detection algorithm aimed at evaluating table tennis players' techniques, tactical patterns and styles. Two datasets that contain 8015 high-definition action videos of 37 elite players were constructed: a front-facing player technical action dataset (4154 videos) and a backwards-facing player technical action dataset (3861 videos). The results showed that YOLOv8-Alphapose-2s-STGCN achieved better recognition performance than seven other YOLOv8-Alphapose-based artificial intelligence algorithms (transformer, BiGRU, BiLSTM, GRU, LSTM, TCN and RNN algorithms) on both datasets and exhibited robust performance in practical applications. In the case study, multiple indicators were used to measure the importance of nodes (players' techniques) within the serving and receiving networks and within the two-round (winning and losing) networks. Dynamic complex network analysis was adopted to evaluate tactical styles and patterns. Furthermore, this study examined whether players and their opponents exhibit variability or similarity in their tactical patterns, focusing on the player networks and the two-round winning and losing networks. By integrating action recognition with process-focused match analysis, this study explored an innovative and comprehensive way to analyse matches, with implications for the performance analysis of table tennis players and players in related racket sports.

ObjectiveThis study investigates the causal relationship between moderate to vigorous physical activity and cognitive performance.MethodsGenetic loci strongly related to moderate to vigorous physical activity from genome-wide association studies were used as instrumental variables. These were combined with genetic data on cognitive performance from different Genome-Wide Association Study (GWAS) to conduct a two-sample Mendelian randomization analysis. The primary analysis used inverse variance weighting within a random effects model, supplemented by weighted median estimation, MR-Egger regression and other methods, with results expressed as Beta coefficient.ResultsThis study selected 19 SNPs closely related to physical activity as instrumental variables. The multiplicative random-effects Inverse-Variance Weighted (IVW) analysis revealed that moderate to vigorous physical activity was negatively associated with cognitive performance (Beta = −0.551; OR = 0.58; 95% CI: 0.46–0.72; p < 0.001). Consistent results were obtained using the fixed effects IVW model (Beta = −0.551; OR = 0.58; 95% CI: 0.52–0.63; p < 0.001), weighted median (Beta = −0.424; OR = 0.65; 95% CI: 0.55–0.78; p < 0.001), simple mode (Beta = −0.467; OR = 0.63; 95% CI: 0.44–0.90; p < 0.001), and weighted mode (Beta = −0.504; OR = 0.60; 95% CI: 0.44–0.83; p < 0.001). After adjusting for BMI, smoking, sleep duration, and alcohol intake frequency, the multivariate MR analysis also showed a significant association between genetically predicted MVPA and cognitive performance, with Beta of −0.599 and OR = 0.55 (95% CI: 0.44–0.69; p < 0.001).ConclusionThe findings of this study indicate that genetically predicted moderate to vigorous physical activity may be associated with a decline in cognitive performance.

This study investigates the evolution of passing networks (PN) at both team and player levels in the FIFA World Cups (WC) from 2010 to 2022. Analyzing 256 matches (7328 player observations) using a multiple-camera tracking system across four WCs, we considered six playing positions: goalkeeper (n = 521), central defender (n = 1192), fullback (n = 1223), midfielder (n = 2039), winger (n = 1320), and central forward (n = 1033). We used 17 network metrics and considered contextual variables such as team formation, and team ranking. Linear mixed-effect models analyzed differences in team and player PN parameters by year and team strength. Results showed a shift from possession-play to direct-play from the 2010 to 2018 WCs, with possession-play returning in 2022. Specifically, high- and low-quality teams significantly decreased their density, average degree (AD), modularity, and average path length in 2018 (p < 0.05). High-quality teams showed increased density, AD, and average weighted degree in 2022 (p < 0.05). Midfielders and central forwards exhibited significantly lower centrality parameters, whereas central defenders and goalkeepers showed increased centrality parameters (p < 0.05). This study highlights the evolutionary trends of passing relationships from a network analysis perspective over twelve years, providing insights into the changing dynamics of team interactions and positional prominence in elite soccer.

Background Substitutions are generally used to promote the match performance of the whole team. This study aimed to analyze the performance of substitute players and explore the performance difference among substitute players, completed players, and replaced players across each position. Methods Chinese Super Soccer League (CSL) matches in the season 2018 including 5871 individual observation from 395 professional soccer players were analyzed by establishing linear mixed models to quantify the performance difference among substitute players (SP) ( n = 1,071), entire match players (EMP) ( n = 3,454), and replaced players (RP) ( n = 1,346), and then separately for each position (central defenders, fullbacks, central midfielders, wide midfielders, and attackers). Results The results show SP display higher high intensity distance and sprint distance significantly ( p < 0.05) relative to playing time than RP and EMP. SP in offensive positions (attackers, wide midfielders) showed significantly higher ( p < 0.05) passing and organizing performance such as passes, ball control, short passes, and long passes than RP or EMP. The scoring performances of central midfielders of SP including goals, shots, and shots on target are significantly higher ( p < 0.05) than RP or EMP. Central defenders of SP showed higher shot blocks and pass blocks ( p < 0.05) while lower passing and organizing performance ( p < 0.05). Conclusion Depending on different playing positions, substitute players could indeed improve physical and technical performance related to scoring, passing, and defending as offensive substitute players can boost organizing performance and substitute defenders enhance defending performance. These could help coaches better understand substitute players’ influence on match performance and optimize the substitution tactic.

This study adopted a novel algorithm, SHapley Additive exPlanation (SHAP), to analyze the table tennis matches based on a hybrid gradient boosting + categorical features-tree-structured parzen estimator (Catboost-TPE) with the four-phase evaluation theory. 110 singles’ matches (9536 rallies) were analyzed, and 59 elite male players’ winning rates from 2018 to 2022 were categorized into three levels (high, medium, low) by k-means cluster analysis. The results showed that Catboost-TPE has the best performance (MSE = 7.5e-05, MAE = 0.006, RMSE = 0.008, \({\text{R}}^{2}\)=0.99 and adjusted \({\text{R}}^{2}\)=0.989) among six hybrid machine learning algorithms. Using Catboost-TPE to calculate the SHAP value of each feature, the global interpretation and multiple local interpretations found that the performance of receive-attack and serve-attack phases have essential impacts on the winning probabilities in current matches. Besides, this study derived the mathematical equations for converting the scoring rate (SR), usage rate (UR) and technique effectiveness (TE) from the four-phase evaluation theory into the new three-phase evaluation theory to further deepen the theoretical and applied value of the four-phase evaluation theory used in this study. These results provided quantitative references to table tennis matches' characteristics and winning phases. These methods used in the study can be widely applied to other sports performance analyses, and the equations derived in this study are also instructive for relative sports.