Player Tracking Research Articles

Individual risk factors and prediction of gambling disorder in online sports bettors - the longitudinal RIGAB study.

While research in online sports betting is dominated by studies using objective player tracking data from providers to identify risky gambling behavior, basicresearch has identified various putative individual risk factors assumed to underlie the development of gambling disorder across all types of gambling. This study aims to examine individual risk factors and their longitudinal clinical relevance in online sports bettors. German online sports bettors (N = 607, Mage = 34, 92% male) from a provider based sample took part in an online survey. The study team randomly preselected customers to be invited. N = 325 (53,45%) of the participants also took part in an online follow-up survey one year later. Crosssectional and longitudinal associations of putative risk factors and DSM-5 gambling disorder in online sports bettors were analyzed. These risk factors include alcohol and tobacco use, impulsivity, difficulties in emotion identification, emotion regulation strategies, comorbid mental disorders and stress. We found more pronounced impulsivity, difficulties in emotion identification, emotion suppression, comorbid mental disorders and stress were cross-sectionally associated with gambling disorder, and longitudinally predicted gambling disorder in online sports bettors (with the exception of emotion suppression). In an overall model only lack of premeditation and perceived helplessness remained significant as predictors for gambling disorder. Online sports bettors with gambling disorder predominantly showed more pronounced risk factors, which were also confirmed longitudinally as relevant for the maintenance of gambling disorder. Risk factors such as impulsivity and stress and appropriate coping mechanisms should consequently be integrated not only into prevention efforts to identify individuals at risk early, but also into intervention efforts to tailor treatment.

A machine learning framework for quantifying in-game space-control efficiency in football

Analysis of player tracking and event data in football matches is used by the coaching staff to evaluate team performance and to inform tactical decision-making, whereas using Machine Learning methods to gain useful insights from the data is still an open research question. The objective of our research is to discover the football team's space-control efficiency using a novel Machine Learning approach and evaluate the team performance based on its space-control efficiency. We develop a novel Possession Evaluation Model through deep generative machine learning to predict the football team's space-control capability utilising tracking and event data. The developed model is used to quantify the efficiency of attacking and defending for a given sequence of play. Performance analysis results demonstrate that this novel method of space-control efficiency quantification is objective and precise. The superior performance of the model is attributed to the utilization of deep generative modelling on image datasets and conditioning in the prediction with contextual factors. This study presents a novel approach to football analysis in evaluating team performance and providing tactical insights for the coach to make data-informed adjustments.

THE ROLE OF ARTIFICIAL INTELLIGENCE IN VIDEO GAME DEVELOPMENT

In this paper, we study the vital role of artificial intelligence (AI) in the development of video games, with a focus on various aspects of AI application in this industry. In the introduction, we discuss both the development of video games and the role of AI systems in the user experience, defining the progression of AI’s role in video games. In the following section, we investigate how in-game entities and AI collaborate. Here, we analyze basic concepts such as Non-Playable Characters (NPCs) and how AI enhances their intelligence and reactivity in the game. The mechanisms of AI in video games are a crucial point of consideration in the next part of the paper. We explain how various AI techniques are used for decision-making, player tracking, and adapting the game to their actions. Furthermore, we explore the use of AI in video games beyond NPC control, examining examples such as procedurally generated worlds and player experience modeling. This application of AI contributes to a deeper and more dynamic player experience. In the modern application of AI in video games, we delve into advanced uses of machine learning and deep neural networks in game development. Here, we consider how AI is used for game personalization, user data analysis, and enhancing graphics and sound. Finally, we discuss the future of video games and the role of neural networks in their development. We predict the growth of AI in various aspects of games and how it will shape the future player experience. In conclusion, we assert that artificial intelligence has become an indispensable part of video game development, and its impact will inevitably expand in the future, enabling increasingly rich, dynamic, and personalized games for players worldwide.

Characteristics and prediction of risky gambling behaviour study: A study protocol.

This study protocol describes the RIGAB study, a prospective case-control-study assessing online sports betting behaviour and underlying risk factors for the development of gambling disorder (GD). It has two aims: (1) to characterise sports bettors concerning putative risk factors and their gambling behaviour, and (2) to predict the development of GD from these factors. At baseline, online sports bettors took part in an online survey comprising a GD screening (DSM-5), questions on gambling behaviour and on the putative risk factors emotion regulation, impulsivity, comorbidities, stress, and substance use. Participants were reinvited for a 1-year follow-up online survey. In a nested design, a subsample was invited in-person to take part in a cognitive-behavioural task battery and a clinical interview. Of the initial 6568 online sports bettors invited, 607 participated at baseline (rate: 9.2%), 325 took part in the 1-year follow-up and 54 participated in the nested in-person assessment. The RIGAB study combines different fields of GD studies: player tracking data and putative risk factors from self-report and behavioural tasks. The results of this study will support the development of preventive measures for participants of online gambling based on the combined findings from previously rather distinct research fields.

Age-Related Reference Intervals for Physical Performance Test Outcomes Relevant to Male Youth Middle Eastern Football Players.

To develop age-specific reference intervals for physical performance test outcomes relevant to male youth Middle Eastern football players. We analyzed mixed-longitudinal data (observations range: 1751-1943 assessments) from a sample of 441 male youth outfield football players (chronological age range: 11.7-18.4y) as part of the Qatar Football Association and Aspire Academy development program over 14 competitive seasons. Semiparametric generalized additive models for location, scale, and shape estimated age-specific reference centiles for 10-m sprinting, 40-m sprinting, countermovement jump height, and maximal aerobic speed variables. The estimated reference intervals indicated that the distribution of the physical performance test scores increased monotonically and nonlinearly with advancing chronological age for sprinting and countermovement jump outcome measures, reaching a plateau after 16years common to each of these performance variables. The maximal aerobic speed median score increased substantially until ∼14.5years, with the nonlinear trend flattening off toward relatively older chronological ages. We developed age-related reference intervals for physical performance test outcomes relevant to youth Qatari football players. Country-wide age-specific reference intervals can assist in the longitudinal tracking of the individual players' progress over time against benchmark values derived from the reference population.

Average demands and most demanding passages of national-level female soccer matches: do small- and large-sided games replicate match demands?

This study aims to (1) determine the average and most demanding passage (MDP) load of national-level female soccer matches and (2) evaluate the relationship between average and MDP load during small-sided games (SSGs), large-sided games (LSGs), and matches. A total of 37 national-level female soccer players from a single club senior team and the U18 team participated. The average and 1-, 3-, and 5-min MDP external (total, high-speed, and very-high-speed running distances, acceleration and deceleration distances, average metabolic power, and high-metabolic load distance) and internal loads (average heart rate, rate of perceived exertion) of the 29 league matches, ten 4 vs. 4 + goalkeepers SSGs, and six 8 vs. 8 + goalkeepers LSGs were analyzed by the Polar Team Pro player tracking system. In matches, the external load variables during 1-, 3-, and 5-min MDPs were 167%-1,165%, 135%-504%, and 126%-359%, of match average values, respectively. In LSGs, all external load variables reached higher average values compared with those during matches, except for the very-high-speed running distance; however, no variable reached higher values of 1-min MDP compared with those during the matches. In SSGs, the average acceleration and deceleration distances were higher compared with those during the matches. The findings from the present study suggest that LSGs and SSGs can be used to overload the average values of the selected external load variables compared with those during the matches; however, other training options must be explored to overload 1-min match MDPs.

An empirical attempt to identify binge gambling utilizing account-based player tracking data

Binge gambling is a relatively under-explored area and the few published studies have all used self-report data (i.e. surveys and interviews). The use of account-based tracking data has increasingly been used to identify indicators of problem gambling. However, no previous study has ever used tracking data to operationalize and explore binge gambling. Therefore, the present study investigated whether it is possible to identify behavioral patterns that could be related to binge gambling among a real-world sample of online gamblers. The authors were given access to an anonymized secondary dataset from a British online casino operator comprising 150,895 online gamblers who gambled between January and March 2023. Using 14 parameters of gambling (e.g. total number of gambling days, total number of gambling sessions, average amount of money spent per game), six distinct clusters of gamblers were identified. Two clusters – Cluster 2 (n = 22,364) and Cluster 5 (n = 12,523) – gambled on a relatively low number of days during three months, but displayed a high gambling intensity on those days compared to the other four clusters. These two profiles could potentially match the habits of binge gamblers. The majority of players retained their behavior in the following three months between April and June 2023 and were consequently assigned to the same cluster in the latter time period. A total of 17% of gamblers in Cluster 3 and 29% of gamblers in Cluster 5 stopped gambling entirely between April and June 2023. The findings suggest that binge gambling may be able to be identified by online gambling operators using account-based tracking data and that targeted interventions could be implemented with binge gamblers.

A Wireless Self-Powered Sensor Network Based on Dual-Model Convolutional Neural Network Algorithm for Tennis Sports

More and more scholars turn their attention to the use of digital and systematic management methods to further improve the safety and effectiveness of tennis. At present, most of these systems are based on video monitoring technology, and their actual operation process is limited by the deployment environment, and the cost is very high. In order to obtain the track of tennis players, this paper uses the sensor array in the intelligent insole to collect the original data. A Dual-model Convolutional Neural Network (DMCNN) structure is designed to further improve the computational efficiency and accuracy of tennis players’ gait recognition. Combined with quaternion and complementary filtering algorithm, the rotation matrix is established to convert the data in the sensor coordinate system to the reference coordinate system. Use multiple sensors in the smart insole to collect linear and curve track data from tennis players. After the unit conversion and normalization of the original data, the acceleration and angular velocity data in the sensor coordinate system are converted to the reference coordinate system by using quaternions to represent the rotation matrix and using complementary filtering algorithm to update the quaternions. The experimental results show that the displacement error of the scheme is small and the performance is good when the track tracking of tennis players is realized. The error between the estimated distance and the actual distance is within 6%, which proves the accuracy of the linear trajectory experiment.

Here Comes the STRAIN: Analyzing Defensive Pass Rush in American Football with Player Tracking Data

In American football, a pass rush is an attempt by the defensive team to disrupt the offense and prevent the quarterback (QB) from completing a pass. Existing metrics for assessing pass rush performance are either discrete-time quantities or based on subjective judgment. Using player tracking data, we propose STRAIN, a novel metric for evaluating pass rushers in the National Football League (NFL) at the continuous-time within-play level. Inspired by the concept of strain rate in materials science, STRAIN is a simple and interpretable means for measuring defensive pressure in football. It is a directly observed statistic as a function of two features: the distance between the pass rusher and QB, and the rate at which this distance is being reduced. Our metric possesses great predictability of pressure and stability over time. We also fit a multilevel model for STRAIN to understand the defensive pressure contribution of every pass rusher at the play-level. We apply our approach to NFL data and present results for the first eight weeks of the 2021 regular season. In particular, we provide comparisons of STRAIN for different defensive positions and play outcomes, and rankings of the NFL’s best pass rushers according to our metric.

Positional Movement Demands during NFL Football Games: A 3-Year Review

Understanding the in-game demands placed on athletes may allow practitioners to design improved training protocols to prepare athletes for competitive demands. This study aimed to quantify the competitive movement demands of professional American football athletes and to determine any inter-positional differences that may exist. Player tracking data were collected from 2018 to 2020 regular season games of the National Football League. Distance, maximum velocity, high-velocity efforts and distance, and acceleration and deceleration efforts and distance were used to evaluate competitive movement demands. To determine position-specific demands, each player was classified by their designated position, and velocity data from competitive games were used to develop position-specific velocity thresholds. One-way ANOVA and post hoc Bonferroni statistical analysis were used to determine inter-positional difference. Significant (p < 0.05) positional differences were found for all load metrics with respect to competitive game demands. Generally, wide receivers and defensive backs had faster maximum velocities, higher distances, and more acceleration and deceleration efforts and distance than other positions. Linebackers accumulated the most high-velocity efforts and distance. Lineman had the lowest values for all assessed metrics. These findings may assist the performance staff in developing improved training and return-to-play protocols with the aim of improving player performance and mitigating injury.

Analysis of player tracking data extracted from football match feed

Analysis of player tracking data extracted from football match feed

Automated player identification and indexing using two-stage deep learning network

American football games attract significant worldwide attention every year. Identifying players from videos in each play is also essential for the indexing of player participation. Processing football game video presents great challenges such as crowded settings, distorted objects, and imbalanced data for identifying players, especially jersey numbers. In this work, we propose a deep learning-based player tracking system to automatically track players and index their participation per play in American football games. It is a two-stage network design to highlight areas of interest and identify jersey number information with high accuracy. First, we utilize an object detection network, a detection transformer, to tackle the player detection problem in a crowded context. Second, we identify players using jersey number recognition with a secondary convolutional neural network, then synchronize it with a game clock subsystem. Finally, the system outputs a complete log in a database for play indexing. We demonstrate the effectiveness and reliability of player tracking system by analyzing the qualitative and quantitative results on football videos. The proposed system shows great potential for implementation in and analysis of football broadcast video.

Review of computer vision in sports

All industries employ machine learning extensively, and one of the most promising fields is computer vision. Computer vision is a simulation of the human visual system that uses cameras and computers to take the role of the human eye to find the target, follow it, and gather data from it so that a decision may be made on whether to take further action or provide recommendations. The various uses of computer vision in sports are covered in this paper. Currently, computer vision is mostly utilized for broadcast enhancement, tracking and detection of players and balls. Although the games graphics has been substantially improved by this technology, there are still several flaws. For instance, some areas are not suited to employ this technology. Another is the issue of players being blocked in multiplayer sports. For broadcasters, computer vision has significant commercial value. For athletes, this technique can improve their performance.

Development of the Online Problem Gaming Behavior Index: A New Scale Based on Actual Problem Gambling Behavior Rather Than the Consequences of it.

Many items in current problem gambling screens focus on negative consequences of gambling and gambling-related harms. However, few problem gambling screens comprise items that are totally based on actual gambling behavior such as gambling duration, gambling frequency, or gambling late at night. The aim of the present study was to develop and validate the 12-item Online Problem Gambling Behavior Index (OPGBI). A total of 10,000 online Croatian gamblers were administered the OPGBI alongside the nine-item Problem Gambling Severity Index (PGSI), as well as questions regarding types of gambling engaged in and socio-demographic factors. The 12 OPGBI items mainly concern actual gambling behavior. The correlation between OPGBI and PGSI was highly significant (r = 0.68). Three latent factors in the OPGBI were identified (gambling behavior, limit setting, communication with operator). The three factors all significantly correlated with the PGSI score (R2- = 51.8%). The fact that pure gambling behavior related items explained over 50% of the PGSI score strengthens the idea that player tracking could be an important approach in identifying problem gambling.

Is a compact organization important for defensive success in elite soccer? – Analysis based on player tracking data

The interest in tactical analysis in soccer has increased in the latest years, especially with the growing availability of player tracking data. With it, the defending team's compact organization, which is considered by practitioners to be an important factor in defense, was repeatedly examined. However, the connection between this defensive principle of play and the defending success remains unclear. Therefore, this study aims to investigate the relation of the principle of playing defensive compact organization to the success of the defense. Based on tracking and event data of 153 games of the German Bundesliga (season 2020/21), the compactness (surface area, spread of the team, and of defending subgroups) and the organization (distances between formation lines) of the defending team was compared between successful and unsuccessful defensive plays. There were almost no differences in the compactness of the whole team, and the organizational measures between successful and unsuccessful defensive plays. The defending subgroup of five defenders closest to the ball showed a higher compactness (smaller surface area and smaller spread) in successful defensive plays compared to unsuccessful ones (−0.08 ≤ d ≤ −0.16). Our results indicate that the compactness of players in areas close to the ball seems crucial for defensive success. However, the compact organization of the entire team does not seem important to regain the ball in defense.

Predicting High-Risk Gambling Based on the First Seven Days of Gambling Activity After Registration Using Account-Based Tracking Data

AbstractIn recent years, several European regulators have introduced mandatory player tracking to identify potentially problematic online gambling. The present study’s aim was to investigate the possibility of predicting future high-risk gambling based on a short time window (i.e., the first seven days after the registration for an online gambling site). The authors were given access to a secondary dataset comprising 37,986 gamblers who registered at a European online gambling operator between January 1 and April 30, 2022. The study examined the association between gambling behavior during the first week after registration and high-risk gambling during the first 90 days after registration. A logistic regression model with high-risk gamblers (based on the first three months of gambling data after initial registration) as the dependent variable and age, gender, and the first week’s gambling behavior as independent variables explained 40% of the variance. Age, gender, and seven player tracking features from the first week after registration were significant. Machine learning models confirmed the high correlation between the first week of gambling and a high-risk classification during the first three months after registration. The most important features reported by a Random Forest and a Gradient Boost Machine model were the total amount of money deposited, the number of deposits, the amount of money lost, and the average number of deposits per session. The study showed that high-risk gambling during the first three months of a player’s lifetime can be predicted very early after registration. These findings suggest that gambling operators should initiate preventive measures (such as limit setting, mandatory play-breaks, personalized messaging) and monitor gambling behavior at a very early stage after a gambler’s initial registration.

Individual attention capacity enhances in-field group performances in soccer

ABSTRACT Visual attention capacity impacts performance in different laboratory and real-world tasks positively. The present field-study extends for the first time performance gains from individuals to team performance. Our study examined the attention capacity of 88 soccer players to create low vs. high attention capacity teams (LACT vs. HACT) who competed in a controlled field-study. Positional data were gathered using player tracking systems (1 Hz) in a standardised 11 vs. 11 soccer game. Key performance indicators (KPIs), which have been established in match analysis research, were measured to reflect tactical performance of the teams. As predicted from attention frameworks, HACT outperformed LACT in five out of seven KPIs, suggesting that attention capacity might play a crucial role for/in searching and gathering important space on the pitch. These findings provide evidence for attention being a predictor of team success. Practice task designs with an emphasis on attention capacity therefore could be a useful tool to study the emergent behavioural dynamics in a real-world environment.

Analysis of Movement and Activities of Handball Players Using Deep Neural Networks

This paper focuses on image and video content analysis of handball scenes and applying deep learning methods for detecting and tracking the players and recognizing their activities. Handball is a team sport of two teams played indoors with the ball with well-defined goals and rules. The game is dynamic, with fourteen players moving quickly throughout the field in different directions, changing positions and roles from defensive to offensive, and performing different techniques and actions. Such dynamic team sports present challenging and demanding scenarios for both the object detector and the tracking algorithms and other computer vision tasks, such as action recognition and localization, with much room for improvement of existing algorithms. The aim of the paper is to explore the computer vision-based solutions for recognizing player actions that can be applied in unconstrained handball scenes with no additional sensors and with modest requirements, allowing a broader adoption of computer vision applications in both professional and amateur settings. This paper presents semi-manual creation of custom handball action dataset based on automatic player detection and tracking, and models for handball action recognition and localization using Inflated 3D Networks (I3D). For the task of player and ball detection, different configurations of You Only Look Once (YOLO) and Mask Region-Based Convolutional Neural Network (Mask R-CNN) models fine-tuned on custom handball datasets are compared to original YOLOv7 model to select the best detector that will be used for tracking-by-detection algorithms. For the player tracking, DeepSORT and Bag of tricks for SORT (BoT SORT) algorithms with Mask R-CNN and YOLO detectors were tested and compared. For the task of action recognition, I3D multi-class model and ensemble of binary I3D models are trained with different input frame lengths and frame selection strategies, and the best solution is proposed for handball action recognition. The obtained action recognition models perform well on the test set with nine handball action classes, with average F1 measures of 0.69 and 0.75 for ensemble and multi-class classifiers, respectively. They can be used to index handball videos to facilitate retrieval automatically. Finally, some open issues, challenges in applying deep learning methods in such a dynamic sports environment, and direction for future development will be discussed.

Metabolic power and energy expenditure in the German Bundesliga.

The aims of the study were to analyze metabolic power (MP) and MP derived parameters for different positions in the German Soccer Bundesliga and to evaluate if classification of high-intensity is more suited using the metabolic power approach instead of using traditional speed-based methods. 1,345 video match analysis (25Hz) datasets from 380 players in 96 matches of the German first Bundesliga were gathered by an automatic player tracking system. Displacement (speed, acceleration, distance) and energetic (MP, energy expenditure) variables were determined. Intensity was classified utilizing conventional thresholds. Metabolic and running profiles were compared among six positional groups and between the halves of the match respectively (one-way ANOVA). Further, time spent, distance covered and energy expended at high speed (>15.5kmh-1) and high acceleration (>2ms-2) were compared to those at high MP (>20Wkg-1) (one-way ANOVA) for evaluating if metabolic power is more suited to describe intensity in team sports. Main findings are that central-attacking midfielders (CAM) and central midfielders (CM) expended more energy (CAM: 59.8 ± 4.2kJkg-1; CM: 59.6 ± 3.6kJkg-1) and covered more distance (CAM: 11,494 ± 765m; CM: 11,445 ± 638m) than all other positions (p < 0.001). In the whole group of players, less time (t) and less energy were expended (EE) in high speed (t: 302 ± 84s; EE: 10.1 ± 2.9kJkg-1) and at high acceleration zones (t: 147 ± 24s; EE: 5.5 ± 1.2kJkg-1) when compared to high MP zone (t: 617 ± 141s; EE: 20.0 ± 4.4kJkg-1) (p < 0.001). Furthermore, players covered more distance at high MP (2,464 ± 597m) than at high speed (1,638 ± 458m) and much less at high acceleration (359 ± 67m) (p < 0.001). The higher activity profiles of CAM and CM compared to the other positions indicate the need for higher developed physiological performance in players of these positions. High intensity activities should be interpreted differently when using MP and displacement parameters as indicators of high intensity in soccer.

Position specific physical demands in different phases of competitive matches in national level women's football.

The purpose of the present study was twofold: to investigate position-specific physical match demands of national level women's football; and to examine if demands change during a match (comparison between first and second half and in 15-minute intervals). Seven teams from the Finnish National League participated in the study. Eighty-five players met the inclusion criteria, and a total of 340 individual match observations from 68 individual matches were included for analysis. The Polar Team Pro -player tracking system (with 10 Hz GPS units, including 200 Hz tri-axial accelerometer, gyroscope, magnetometer and HR monitor) was used to assess positional data and HR response of the players. This study demonstrated that women's national level football matches place a range of physical demands on players, which in general were highest for wide midfielders, and lowest for central defenders. Wide midfielders and forwards performed significantly more 'very high-speed' running, sprinting, accelerations, and decelerations than other outfield positions (p < 0.05). HRmean varied from 84-87% of HRmax and was significantly lower for central defenders than central midfielders (p < 0.001). External load variables varied during a match and generally decreased especially after 60 minutes of play compared to first 15-min period of the match. Present study showed that national level women football players' positional differences in match demands are similar to those reported with elite players in previous studies. On national level, players' physical performance tended to decrease towards the end of the match, especially in terms of total distance (~10%), high-speed running (~20%), and decelerations (~20%).