The Basketball Pass: A Systematic Review

HoopTransformer: Advancing NBA Offensive Play Recognition with Self-Supervised Learning from Player Trajectories

BACKGROUND AND OBJECTIVE

Understanding and recognizing basketball offensive set plays, which involve intricate interactions between players, have always been regarded as challenging tasks for untrained humans, not to mention machines. In this study, our objective is to propose an artificial intelligence model that can automatically recognize offensive plays using a novel self-supervised learning approach.

METHODS

The dataset was collected by SportVU from 632 games during the 2015-2016 season of the National Basketball Association (NBA), with a total of 90,524 possessions. A multi-agent motion prediction pretraining model was built on the basis of axial-attention transformer and trained with different masking strategies: motion prediction (MP), motion reconstruction (MR), and MP + MR joint strategy. A downstream play-level classification task and similarity search were used to evaluate the models' performance.

RESULTS

The results showed that the MP + MR joint masking strategy maximized the ability of the model compared with individual masking strategies. For the classification task, the joint strategy achieved a top-1 accuracy of 81.5% and top-3 accuracy of 97.5%. In the similarity search evaluation, the joint strategy attained a top-5 accuracy of 76% and top-10 accuracy of 59%. Additionally, with the same MP + MR joint masking strategy, our HoopTransformer model outperformed the two baseline models in the classification task and similarity search.

CONCLUSION

This study presents a self-supervised learning model and demonstrates the effectiveness and potential of the model in accurately comprehending and capturing player movements and complex interactions during offensive plays.

Levels of Physical Activity at Age 10 Years and Brain Morphology Changes From Ages 10 to 14 Years

Importance

Physical activity may promote healthy brain development in children, but previous research was predominantly cross-sectional and included small samples, providing limited knowledge.

Objective

To investigate the longitudinal associations of physical activity with brain morphology changes.

Design, Setting, and Participants

A 4-year longitudinal population-based cohort study in Rotterdam, the Netherlands, embedded in Generation R, a cohort from fetal life onward. From the women enrolled during pregnancy, children who had repeated measures of brain structure at ages 10 (range 8 to 12) years and 14 (range 13 to 15) years were included. Data were collected from March 2013 to November 2015 (baseline) and from October 2016 to January 2020 (follow-up). Data were analyzed from April to December 2022.

Exposure

At age 10 years, both the child and their primary caregiver reported the child's levels of physical activity with regard to sport participation, outdoor play, and total physical activity. Primary analyses were based on an average multi-informant report.

Main outcomes and measures

Brain morphology was quantified by magnetic resonance imaging. Hypothesized regions of interest were the bilateral amygdala and hippocampal volumes. Global brain measures were studied to test the specificity of the hypothesis.

Results

Data were available for 1088 children (566 girls [52%]; 693 [64%] Dutch). Their mean (SD) age at baseline was 10.1 (0.6) years. For amygdala volume change, positive associations with multi-informant reports of total physical activity (β = 2.6; 95% CI, 0.3-4.9) were found. Total physical activity was associated with hippocampal volume increases only when reported by the child (β = 3.1; 95% CI, 0.4-5.8). No robust associations with global brain measures were found.

Conclusions and relevance

In this cohort study of 1088 children, more physical activity at 10 years was consistently associated with an increase in amygdala volume in children aged 10 to 14 years. Physical activity and increases in hippocampal volume were found using child reports of physical activity only. These findings suggest physical activity in late childhood was prospectively associated with volumetric changes in specific subcortical structures, but not to global brain development, from late childhood to early adolescence. These findings may inform the design of future public health interventions to best facilitate neurodevelopment with physical activity.

A comparative study of Chinese women 3 × 3 basketball players exercise load in Tokyo Olympic preparation cycle

Objective: The aim of this study was to compare the variances in-game loads exhibited by Chinese women's 3 × 3 basketball team across different stages of the preparation cycle for the Tokyo Olympic Games, and to summarize the fundamental regulations governing Chinese women's 3 × 3 basketball training and games, in order to establish a theoretical research foundation for the team's new preparation cycle. Methods: This study measured load-related data during the preparation and main competition periods of the 2019-2021 Tokyo Olympics, from April to August 2019 and from April to June 2021. The aim was to compare the changes and differences in a load of competition during different stages and to explore patterns of load changes during the preparation period. This study used wearable devices authorized by FIFA and NBA, along with the Catapult GPS performance monitoring system from Australia (Catapult&Polar Team) as instruments for collecting sports load data. The OptimEye S5 device was worn around the athlete's neck to collect data prior to the game, while the Open Field™ system was utilized for data editing and report generation post-game. Results: Compared to the primary competition load during the 2019 preparation period, the 2021 preparation period exhibited significant increases and decreases (p < 0.05) in competition load, high-intensity load, the number of explosive moves, the number of high-intensity acceleration, several changes to the left and right, and the number of explosive jumps. During the 2021 preparation period, the mean heart rate, mean heart rate percentage, and mean speed of the race demonstrated significant decreases in comparison to the race during the 2019 preparation period (p < 0.05). Throughout the training period spanning from 2019 to 2021, no significant differences were observed in running distance and maximum speed (p > 0.05). Conclusion: The findings of this study reveal that the national training team has fostered positive adaptive changes in athletes, resulting in a significant enhancement in both load and sports performance science data during competition from 2019 to 2021.

Wheelchair skill tests in wheelchair Basketball: A systematic review

OBJECTIVES

The aim of this study was to identify, describe and synthesize the skill tests used in wheelchair basketball.

METHOD

A systematic review was carried out in the databases: PubMed/Medline, ScienceDirect, Scopus, Web of science and Google Scholar from inception to January 2021 with up to date in January 2022. the eligibility criteria used were Inclusion: (i) evaluation of wheelchair basketball athletes; (ii) using skill tests (defined as agility, speed, ball maneuverability, slalom, etc.) and (iii) papers needed to be written in English and published in peer-reviewed journals. Exclusion: (i) papers with poor description of the test methodology, (ii) participants not classified as wheelchair basketball athletes (less than one year of practice) and (iii) Participants were not people without disabilities.

RESULTS

Our main findings were: a) the most explored skills were pass and speed, and the most frequent test was the pass tests and sprint tests, b) Strong associations were found between sports classes and performance in field tests, c) The most used tests for each skill were: pass = pass accuracy and maximum pass; speed = 20m sprint test with and without the ball; agility = slalom test; dribbling = obstacle dribbling tests and throw = free throw and spot shot.

CONCLUSION

The most explored skills were passing and speed, and to evaluate these skills we highlight the two-handed chest pass test, 20m sprint test with ball and the WMP test. The use of specific tests can facilitate the creation of reference standards and possible comparison of athletes and, thus, enable better training conditions, aiming to meet the specific demands of each athlete and team.

Match Analysis in Team Ball Sports: An Umbrella Review of Systematic Reviews and Meta-Analyses

Background

Match analysis has evolved exponentially over the past decades in team sports resulting in a significant number of published systematic reviews and meta-analyses. An umbrella review of the available literature is needed to provide an integrated overview of current knowledge and contribute to more robust theoretical explanations of team performance.

Methods

The Web of Science (all databases), PubMed, Cochrane Library (Cochrane Database of Systematic Reviews), Scopus, and SPORTDiscus databases were searched for relevant publications prior to 19 February 2021. Appraisal of the methodological quality of included articles was undertaken using the tool for Assessing the Methodological Quality of Systematic Reviews (AMSTAR-2). Twenty-four studies were reviewed that met the following criteria: (1) contained relevant data from match analyses in team ball sports; (2) were defined as systematic reviews or/and meta-analyses; and (3) were written in the English language.

Results

The overall methodological quality of the 24 included reviews, obtained through the AMSTAR-2, revealed very low confidence ratings (Critically Low, n = 12) for the results of most systematic reviews of match analyses in team ball sports. Additionally, the results showed that research is focused mainly on four levels of analysis: (1) dyadic (microlevel); (2) individual (molecular level; predominant); (3) group (mesolevel), and (4) team dynamics (macrolevel). These levels of analysis included tactical, technical, physical, and psychosocial variables. Team performance was contextualized at two levels, with reference to: (1) match context (e.g. match status, match location, match period, quality of opposition) and (2) sociodemographic and environmental constraints (sex, age groups, competitive level, altitude, temperature, pitch surface).

Conclusions

The evolution of methods for match analysis in team ball sports indicates that: (1) an individual-level performance analysis was predominant; (2) the focus on intermediate levels of analysis, observing performance in dyadic and group interactions, has received less attention from researchers; (3) neglected areas of research include psychosocial aspects of team sports and women’s performance; and (4) analyses of match contexts need greater depth.

Registration: The protocol was registered in the International Platform of Registered Systematic Review and Meta-Analysis Protocols with the number 202080067 and the DOI number https://doi.org/10.37766/inplasy2020.8.0067.

The Q-Pass Index: A Multifactorial IMUs-Based Tool to Assess Passing Skills in Basketball

Despite being a key sport-specific characteristic in performance, there is no practical tool to assess the quality of the pass in basketball. The aim of this study is to develop a tool (the quality-pass index or Q-Pass) able to deliver a quantitative, practical measure of passing skills quality based on a combination of accuracy, execution time and pass pattern variability. Temporal, kinematics and performance parameters were analysed in five different types of passes (chest, bounce, crossover, between-the-leg and behind-the-back) using a field-based test, video cameras and body-worn inertial sensors (IMUs). Data from pass accuracy, time and angular velocity were collected and processed in a custom-built excel spreadsheet. The Q-pass index (0–100 score) resulted from the sum of the three factors. Data were collected from 16 young basketball players (age: 16 ± 2 years) with high (experienced) and low (novice) level of expertise. Reliability analyses found the Q-pass index as a reliable tool in both novice (CV from 4.3 to 9.3%) and experienced players (CV from 2.8 to 10.2%). Besides, important differences in the Q-pass index were found between players’ level (p < 0.05), with the experienced showing better scores in all passing situations: behind-the-back (ES = 1.91), bounce (ES = 0.82), between-the-legs (ES = 1.11), crossover (ES = 0.58) and chest (ES = 0.94). According to these findings, the Q-pass index was sensitive enough to identify the differences in passing skills between young players with different levels of expertise, providing a numbering score for each pass executed.

Decision-Making Skills in Youth Basketball Players: Diagnostic and External Validation of a Video-Based Assessment

Decision-making is a central skill of basketball players intending to excel individually and contribute to their teams’ success. The assessment of such skills is particularly challenging in complex team sports. To address this challenge, this study aimed to conceptualize a reliable and valid video-based decision-making assessment in youth basketball. The study sample comprised youth basketball players of the German U16 national team (n = 17; M_Age = 16.01 ± 0.25 years) and students of a sports class (n = 17; M_Age = 15.73 ± 0.35 years). Diagnostic validity was tested by determination of the performance levels according to response accuracy as well as response time in the assessment. External validity was examined by investigation of the correlation between the diagnostic results of the elite athletes and their real game performance data associated with passing skills. Logistic regression analysis revealed that the diagnostic results discriminate between performance levels (χ²(2) = 20.39, p < 0.001, Nagelkerke’s R² = 0.60). Multiple regression analysis demonstrated a positive relationship between the diagnostic results and assists (F(2,10) = 4.82, p < 0.05; R² = 0.49) as well as turnovers per game (F(2,10) = 5.23, p < 0.05; R² = 0.51). However, no relationship was detected regarding the assist-turnover ratio. Further, response time discriminated within the elite athletes’ performance data but not between performance levels while for response accuracy the opposite is the case. The results confirm the diagnostic and external validity of the assessment and indicate its applicability to investigate decision-making skills in youth basketball.

Determinants of Dribbling and Passing Skills in Competitive Games of Women's Basketball

This work aimed to identify the influence of selected endogenous (internal load) and exogenous (possession duration, game quarter, and defensive pressure) factors in natural game conditions on the efficiency of dribbling and passing skills. Dribbling and passing skills were assessed during four games of U19 female basketball players and five games of senior (2nd division) female basketball players. In total, 551 dribbling and 914 passing executions were evaluated. Binary logistic regression identified defensive pressure of the opponent as a predictor of dribbling and passing skills effectivity. When the defensive pressure of the opponent was medium, the chances for the ineffective pass were 1.997 times more likely (95% CI; 1.179-3.383), as it is at the minimum defensive pressure. When the defensive pressure of the opponent was high, the chances for ineffective dribbling were 7.45 times more likely (95% CI; 3.672-15.113) and for ineffective pass were 8.419 times more likely (95% CI; 4.6-15.409), as it is at minimum defensive pressure. The game quarter and the internal load were not identified as the predictors influencing the dribbling and passing effectivity. Possession duration was also an insignificant predictor of dribbling and passing skills effectivity. However, the passing skill effectivity decreases when the shot clock is winding down. These findings confirm the importance of transferring game situations into the training process. Coaches should take into account these factors when they want to stimulate determinants of player performance in a balanced and systematic way.

Predicting field-goal success according to offensive, defensive and contextual variables in elite men's wheelchair basketball

The purposes of this study were to (i) develop a field-goal shooting performance analysis template and (ii) explore the impact of each identified variable upon the likely outcome of a field-goal attempt using binary logistic regression modelling in elite men’s wheelchair basketball. First, a field-goal shooting performance analysis template was developed that included 71 Action Variables (AV) grouped within 22 Categorical Predictor Variables (CPV) representing offensive, defensive and game context variables. Second, footage of all 5,105 field-goal attempts from 12 teams during the men’s 2016 Rio De Janeiro Paralympic Games wheelchair basketball competition were analysed using the template. Pearson’s chi-square analyses found that 18 of the CPV were significantly associated with field-goal attempt outcome (p < 0.05), with seven of them reaching moderate association (Cramer’s V: 0.1–0.3). Third, using 70% of the dataset (3,574 field-goal attempts), binary logistic regression analyses identified that five offensive variables (classification category of the player, the action leading up to the field-goal attempt, the time left on the clock, the location of the shot, and the movement of the player), two defensive variables (the pressure being exerted by the defence, and the number of defenders within a 1-meter radius) and 1 context variable (the finishing position of the team in the competition) affected the probability of a successful field-goal attempt. The quality of the developed model was determined acceptable (greater than 65%), producing an area under the curve value of 68.5% when the model was run against the remaining 30% of the dataset (1,531 field-goal attempts). The development of the model from such a large sample of objective data is unique. As such it offers robust empirical evidence to enable coaches, performance analysts and players to move beyond anecdote, in order to appreciate the potential effect of various and varying offensive, defensive and contextual variables on field-goal success.

Behavioral and Cognitive Electrophysiological Differences in the Executive Functions of Taiwanese Basketball Players as a Function of Playing Position

The effect of the predominant playing position of elite basketball players on executive functions using both behavioral and electrophysiological measurements was investigated in the present study. Forty-six elite basketball players, including 27 guards and 19 forwards, were recruited. Event-related potential (ERP) signals were simultaneously recorded when the athletes performed the visual Go/NoGo task. Analyses of the results revealed that the guards and forwards groups exhibited comparable behavioral (i.e., reaction time (RTs) and accuracy rates (ARs)) performance. With regards to the electrophysiological indices, the guards relative to the forwards exhibited a shorter N2 latency in the Go condition, a longer N2 latency in the NoGo condition, and a smaller P3 amplitude across the two conditions. These results suggested that although the guards and forwards exhibited similar abilities in terms of behavioral inhibition, different neural processing efficiencies still exist in the basketball playing positions, with guards showing divergent efficiencies in the target evaluation and response selection of the target and non-target stimuli and fewer cognitive resources during premotor preparation and decision-making as compared to the forwards.