Next Article in Journal
Fiber-Optic Pressure Sensors: Recent Advances in Sensing Mechanisms, Fabrication Technologies, and Multidisciplinary Applications
Previous Article in Journal
“Speed”: A Dataset for Human Speed Estimation
Previous Article in Special Issue
Validation of Aerobic Capacity (VO2max) and Pulse Oximetry in Wearable Technology
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sensors
Volume 25
Issue 20
10.3390/s25206334
sensors-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

The Relationship Between External Load and Player Performance in Elite Female 3 × 3 Basketball Games: A Markerless Motion Capture Approach

by
Mingjia Qiu
1,2,†,
Rui Dong
3,†,
Junye Tao
4,
Zhaoyu Li
5,
Wen Zheng
1,2 and
Mingxin Zhang
1,2,*
1
School of Athletic Performance, Shanghai University of Sport, Shanghai 200438, China
2
Shanghai Key Lab of Human Performance, Shanghai University of Sport, Shanghai 200438, China
3
School of Physical Education and Sports Science, South China Normal University, Guangzhou 510631, China
4
College of Physical Education and Sports, Beijing Normal University, Beijing 100875, China
5
Institute of Physical Education and Training, Capital University of Physical Education and Sports, Beijing 100091, China
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work.
Sensors 2025, 25(20), 6334; https://doi.org/10.3390/s25206334
Submission received: 11 September 2025 / Revised: 3 October 2025 / Accepted: 8 October 2025 / Published: 14 October 2025
(This article belongs to the Special Issue Sensors for Performance Analysis in Team Sports: Second Edition)
Browse Figures
Versions Notes

Abstract

Background: This study employed a markerless motion capture system to quantify the external game load of elite 3 × 3 basketball players and evaluated its association with game performance. Methods: Twenty-four female 3 × 3 basketball games from the 2024 Paris Olympic Games were analyzed, involving 32 players from eight national teams. A markerless motion capture system was used to collect six categories of external load metrics during games, and 22 types of technical statistics were gathered to determine performance. Collected data were standardized according to live game time (min−1). Repeated-measures correlation analysis was applied to examine the relationships between external load and performance, while mixed-effects models were used to compare external load differences between better- and worse-performing groups (classified by Player Value). Results: The correlations between external load and performance indicators were trivial to small. Accelerations (ACC) were significantly associated with the greatest number of performance indicators (e.g., points, rebounds, 1-point made, key assists), while rebounds were significantly correlated with the largest number of external load metrics (e.g., total distance, low-intensity active distance, high-intensity active distance); however, all correlations remained at the small level (r = 0.16–0.24). No significant differences in external load were observed between players of differing performance groups (p > 0.05). Conclusions: In elite 3 × 3 basketball, external load reflects players’ involvement and effort rather than serving as a primary determinant of game performance. This study provides new empirical evidence on the characteristics of 3 × 3 basketball, suggesting that coaches and strength and conditioning practitioners should adopt a comprehensive perspective when evaluating performance, with external load being more suitable for training regulation and fatigue monitoring.

1. Introduction

In April 2025, the International Olympic Committee announced that the number of participating teams in 3 × 3 basketball would be expanded from 8 to 12 per gender at the upcoming 2028 Los Angeles Olympic Games. This initiative not only provides more nations with the opportunity to compete on the Olympic stage, but also further consolidates 3 × 3 basketball’s status as a flagship project promoted by the International Basketball Federation (FIBA). The rules of 3 × 3 basketball (10-min games, 12-s shot clock), along with simplified court dimensions and reduced player numbers, accelerate game tempo and increase the frequency of transitions, making it a high-intensity intermittent sport [1,2]. During games, players perform numerous accelerations, decelerations, changes in direction, jumps, and physical contests. On average, 3 × 3 basketball involves ~24% of playing time at high intensity, compared with ~9% in traditional 5 × 5. In terms of event density, players in 3 × 3 execute approximately 1.1–2.1 accelerations·min−1, 1.5–2.8 decelerations·min−1, and 3.9–5.9 changes in direction·min−1, which is consistently higher than the corresponding 0.6–0.8, 1.1–1.7, and 3.6–4.5·min−1 observed in 5 × 5 basketball [1]. This dense physical output pattern results in heart rates consistently exceeding 90% of maximum [3], accompanied by a rapid increase in blood lactate concentration. On average, blood lactate values in 3 × 3 basketball (~6.1–6.3 mmol·L−1) are broadly comparable to those in 5 × 5 basketball (≈6.5  ±  2.1 mmol·L−1 in females; up to ≈6.8  ±  2.8 mmol·L−1 in males) [4,5]. It should be noted that 3 × 3 games are of much shorter duration than 5 × 5, and that lactate values are influenced by the timing of sample collection. These studies reflect the high intensity reached during competition in this sport and its heavy reliance on anaerobic metabolic capacity. Moreover, the scheduling of 3 × 3 basketball is often organized as short-term tournaments, where players frequently play multiple games within a single day and are required to perform high-intensity tasks under conditions of incomplete recovery [6]. In the absence of scientific load monitoring, players are prone to fatigue or overload, which may impair tactical execution and competitive performance. Therefore, establishing a scientific framework for load monitoring and performance evaluation is crucial for this emerging Olympic discipline. However, research on the characteristics and demands of 3 × 3 basketball has remained limited over the past decade, highlighting the urgent need for empirical studies based on real-game contexts to better inform researchers and practitioners.
External load generally refers to the objective physical stimuli experienced by players during competition or training [7,8]. The monitoring and quantification of external load have been widely applied in team sports, providing important references for training regulation and performance evaluation. By tracking external load, coaches can ensure that players receive appropriate stimuli, thereby reducing the risks of non-functional overtraining, injuries, and illness [9,10]. With technological advancements, researchers can objectively measure players’ physical workload at both macro levels (e.g., PlayerLoad, running distance) and micro levels (e.g., acceleration changes, directional shifts). Traditionally, inertial measurement units (IMUs), local positioning systems (LPS), and integrated microsensor devices (e.g., Catapult) have been the primary tools, and their validity has been demonstrated in basketball training and certain official games [11,12]. Similarly, the application of these technologies has provided valuable tools for researchers and practitioners of 3 × 3 basketball to further understand the characteristics of the sport [4,5]. However, the application of traditional load-monitoring technologies in official 3 × 3 basketball games faces environmental limitations and regulatory constraints.
In official 3 × 3 basketball games, due to complex venue environments and restrictions at major international competitions (e.g., the Olympic Games) prohibiting players from wearing sensors, external load data can only be obtained through time-consuming and inefficient time-motion analysis [13,14]. In contrast, advances in computer vision and artificial intelligence over the past decade have driven the emergence of markerless motion capture systems. These systems do not require players to wear devices; instead, they rely on multi-view video and algorithms for human keypoint detection and 3D pose reconstruction, and they have been applied in the sports domain [15,16]. Initially, this approach was primarily used in sports biomechanics laboratories to support precise motion capture, posture estimation, and gait analysis in clinical or laboratory settings, and multiple studies have confirmed its good agreement with traditional systems such as IMUs and Vicon [17,18,19,20]. Subsequently, this approach was extended to performance analysis. Researchers incorporated it into training and real-game contexts, and it has been applied in sports such as badminton and tennis [21,22]. In these studies, researchers extracted valid external load data using markerless motion capture systems. With advantages such as deployment flexibility, efficiency, and non-invasiveness, markerless motion capture systems are gradually becoming a feasible alternative for load monitoring in real-game environments of team sports.
In basketball research, both external and internal load measures have been examined in relation to player performance [23,24,25,26]. External load is commonly quantified using indicators such as PlayerLoad, total distance covered, accelerations/decelerations, and actions across different intensity zones, while internal load is typically assessed through session rating of perceived exertion (sRPE). Player performance has generally been evaluated using box-score statistics, efficiency ratings, or performance index rating (PIR). Evidence from 5 × 5 basketball suggests that external load during official games shows small-to-moderate associations with performance, varying across competition level, age, and sex [23,24,25]. Vázquez-Guerrero [26] further applied cluster analysis to classify performance levels (better, intermediate, worse) and reported no significant differences in training load across these groups. Regarding internal load, studies indicate moderate associations with performance outcomes [25]. However, this body of evidence is derived primarily from 5 × 5 basketball. Although recent research has demonstrated strong correlations between external load (Movement Load) and internal load (sRPE, summated-heart-rate-zones) in 3 × 3 basketball, to our knowledge no study has yet systematically examined the relationship between external load and performance in 3 × 3 players.
Considering the differences between 3 × 3 and 5 × 5 basketball in game rules and characteristics, as well as in technical–tactical, physiological, physical, perceptual, and competition scheduling demands [1], findings from 5 × 5 cannot be directly generalized to 3 × 3, as such structural and demand-related differences may alter load–performance relationships and physiological responses across formats. In fact, existing studies have already highlighted the particular significance of external load in 3 × 3 basketball. Sansone and colleagues [2] recorded the “Movement Load” of adult male players across multiple official games and examined its correlations with heart rate load and ratings of perceived exertion (RPE). They found strong positive correlations with both physiological and subjective responses (r = 0.57–0.76), revealing the important feedback role of external load in monitoring players’ status in 3 × 3 basketball; however, its relationship with game performance was not further tested. Particularly given existing evidence suggesting that sex influences players’ physical output during competition [13], targeted studies on female samples have long been neglected. Such information is crucial for 3 × 3 basketball practitioners, guiding them in selecting workload thresholds in practice and providing gender-sensitive data to inform training optimization and game management.
Therefore, the purpose of this study was to quantify players’ external load using a markerless motion capture system in real 3 × 3 basketball games and to systematically examine its relationship with game performance. Specifically, this study (1) analyzed the correlations between external load indicators (e.g., PlayerLoad, running distance, number of accelerations) and performance metrics; and (2) classified players’ performance into high and average groups based on overall performance levels, and compared the external load differences between groups. We hypothesized that external load indicators would show significant associations with game performance metrics, but that the magnitude of these relationships would be small-to-moderate, reflecting the complex and multifactorial nature of basketball performance. Additionally, we expected that players with better performance would show higher external loads than those with worse performance.

2. Materials and Methods

2.1. Participants and Design

Twenty-four group-stage games from the female 3 × 3 basketball tournament at the 2024 Paris Olympic Games were monitored, covering all players from eight national teams (n = 32; age = 29.4 ± 4.7 years; height = 182.6 ± 7.4 cm). This study involved no intervention, as players competed in publicly accessible game settings. Strict anonymization procedures were applied during data processing to ensure that no personally identifiable information of players was included. The study strictly adhered to the principles of the Declaration of Helsinki, complied with ethical guidelines for sport and exercise science, and was approved by the Ethics Committee of Shanghai University of Sport [approval number: 102772025RT096].
This study employed a descriptive, observational, and retrospective design to explore the relationship between external load and players’ game performance. External load data were extracted using a markerless motion capture system, while performance data were obtained from the official statistics provided on the FIBA website.

2.2. Data Collection

To quantify external load during competition, this study employed a video-based markerless motion capture method. Researchers recorded multi-angle video of the 3 × 3 basketball games at the Paris Olympics: three Apple iPhone 13 devices (iOS 18, Apple Inc., Cupertino, CA, USA) were mounted at elevated positions in the front and side stands, approximately 6–7 m above court level and about 10 m from the court boundary, ensuring full-court coverage. Each device recorded in 1080p at 25 fps, with all other camera parameters maintained at default settings. A schematic of the setup is presented in Figure 1 to illustrate the measurement environment and ensure reproducibility. The recorded videos were subsequently edited following established procedures in basketball time–motion analysis, retaining only periods when the game clock was running (i.e., live play), and excluding stoppages due to fouls, timeouts, free throws, or out-of-bounds situations [27,28]. Video processing and analysis were conducted by an experienced basketball video analyst, with over three years of professional club experience, using SportsCode Elite software (v12, Hudl, Australia). For reliability assessment, the analyst reprocessed the game videos, yielding excellent intra-observer consistency (ICC = 0.92). Player performance data were collected from the official FIBA website, with indicators defined by the FIBA 3 × 3 statistics manual and previously applied in 3 × 3 basketball research [29], as shown in Table 1.

2.3. Markerless Motion Capture System

The external load data in this study were obtained using a markerless motion capture system based on Real-Time Models for Pose estimation (RTMPose) technology. RTMPose is a rapidly developing real-time pose estimation technique that can simultaneously identify skeletal keypoints of multiple players and track their dynamic movement trajectories. This method detects 17 key body points in video, aligned with pose detection from the authoritative Microsoft Common Objects in Context (MS COCO) dataset [30], including joints or endpoints of the head, trunk, and limbs [31]. In frame-by-frame computation, the two-dimensional coordinates of keypoints are first extracted, and then the three-dimensional coordinate system is reconstructed using triangulation under multi-view geometric constraints [32], thereby obtaining the complete skeletal information of players in the scene. Players’ external load was calculated based on the trajectory changes of these keypoints. Previous studies have demonstrated that RTMPose performs well on public datasets, achieving a mean accuracy of 75.8% on the MS COCO validation set [31,33]. The technique has been validated in applied studies on gait and human kinematics [34,35,36], with reliability and accuracy comparable to traditional methods such as inertial measurement units (IMUs) [18]. The RTMPose technology was implemented using Python 3.6 (Python Software Foundation, version 3.6). The detailed code implementation of RTMPose can be found on GitHub (https://github.com/open-mmlab/mmpose/tree/1.x/projects/rtmpose (accessed on 10 March 2024)).
The validity and reliability of the RTMPose-based markerless motion capture system used in this study were supported by internal comparison tests and external published evidence. We conducted comparison tests with the widely recognized Vicon marker-based motion capture system, which is considered the gold standard in biomechanical analysis. During the execution of typical basketball movement tasks by participants, we employed a custom marker set (ankle, knee, hip, shoulder, elbow, and wrist) rather than standardized models such as Plug-in-Gait or Helen Hayes, which are primarily designed for gait analysis and full-body kinematics/kinetics. This configuration focused on the joint trajectories most relevant to basketball-specific tasks, avoiding redundant inverse-dynamics modeling that was not aligned with our study aims. Simultaneous data collection revealed high spatial agreement between the two systems, with a mean intra-class correlation coefficient (ICC) of 0.89 (range: 0.77–0.97), indicating strong consistency in joint displacement trajectories. These results are consistent with recent validation work [37], which evaluated the RTMPose-based markerless motion capture system during standardized 3 × 3 basketball tasks. Their study reported that, compared with tape-measure benchmarks, displacement estimation bias and standard error of estimate (SEE) were below 5%, and velocity estimation showed an ICC range of 0.97 to 1.00 compared with manual time–motion analysis. In addition, recent work [38] further demonstrated the robustness of this system under complex game conditions, showing that it could reliably capture 3D trajectories of the head, trunk, and limbs even during high-speed movements, body occlusions, or overlapping trajectories of multiple players. Overall, the RTMPose-based markerless motion capture system provides an efficient, reliable, and ecologically valid tool for extracting external load in real basketball game environments.

2.4. External Load Metrics

Based on indicators previously used in traditional basketball and 3 × 3 basketball studies [4,5,11,39,40,41], six categories of external game load metrics were collected, as follows: (1) PlayerLoad, a reliable measure of external load, defined as the instantaneous rate of change in acceleration along the transverse (x), coronal (y), and sagittal (z) axes. This is calculated as the square root of the sum of squared changes in acceleration across the x, y, and z planes, multiplied by a scaling factor of 0.01. (2) Total distance, referring to the distance covered by players during the game. (3) Low-intensity, (4) moderate-intensity, and (5) high-intensity active distances (LIA, MIA HIA), representing movements distinct from standing, walking, or jogging, performed at low intensity (<2.5 m/s), moderate intensity (2.5–3.5 m/s), and high intensity (>3.5 m/s), respectively. (6) Accelerations (ACC): high-intensity forward or backward movements characterized by effort and intent approaching maximal levels (>3.5 m/s). All metrics were normalized to game time (min−1) to account for variations in game duration and playing time. Intra-class correlation coefficient (ICC) calculations for repeated measures of external load indicators showed excellent consistency (ICC = 0.98).

2.5. Statistical Analysis

Multiple repeated-measures correlation analyses were conducted between external load metrics and game performance to evaluate their relationships. Because players were monitored across multiple games, this approach explicitly accounted for repeated measures within players. The method was implemented using the rmcorr package in RStudio (version 4.1). According to Hopkins’ definitions [42], correlation coefficients (r) were interpreted as follows: trivial (r < 0.1), small (0.1 < r < 0.3), moderate (0.3 < r < 0.5), large (0.5 < r < 0.7), very large (0.7 < r < 0.9), almost perfect (r > 0.9), or perfect (r = 1).
To determine whether external load could differentiate player performance, k-means cluster analysis was applied to classify each player’s game Player Value (P-VAL) into better performance (n = 63, P-VAL = 8.58 ± 2.67) and worse performance (n = 128, P-VAL = 2.82 ± 1.81). This method is commonly used in performance analysis research as an unsupervised method of dividing data into two categories [43,44,45]. Independent linear mixed-effects models were then constructed, accounting for repeated measures within individuals. Specifically, each external game load metric was modeled as the dependent variable, with player performance as a fixed effect (two levels: better vs. worse performance) and players as a random effect (random intercept and fixed slope). p-values were adjusted using Bonferroni post hoc tests. Estimated marginal means and their 95% confidence intervals were calculated for each external load metric as descriptive statistics. Statistical significance was set at p < 0.05. All statistical analyses were performed using IBM SPSS software (version 24; IBM Corp, Armonk, NY, USA).

3. Results

The correlations between players’ external loads and game performances are presented in Figure 2. PlayerLoad showed a small positive correlation with 1PT% (r = 0.16, p = 0.042). Total Distance demonstrated small positive correlations with REB (r = 0.19, p = 0.018), DREB (r = 0.19, p = 0.014), and TO (r = 0.22, p = 0.005). LIA showed small positive correlations with PTS (r = 0.18, p = 0.021), REB (r = 0.16, p = 0.043), 1PT attempted (r = 0.16, p = 0.042), FT made (r = 0.17, p = 0.031), DREB (r = 0.16, p = 0.038), and TO (r = 0.23, p = 0.004). MIA showed a small positive correlation with 1PT% (r = 0.17, p = 0.034) and a small negative correlation with BZR (r = −0.16, p = 0.040). HIA showed small positive correlations with REB (r = 0.18, p = 0.023) and OREB (r = 0.16, p = 0.044). ACC showed small positive correlations with PTS (r = 0.24, p = 0.003), REB (r = 0.18, p = 0.026), 1PT made (r = 0.20, p = 0.013), 1PT attempted (r = 0.18, p = 0.013), FT attempted (r = 0.16, p = 0.040), KAS (r = 0.19, p = 0.014), BS (r = 0.17, p = 0.033), and OREB (r = 0.19, p = 0.014).
Comparisons of external load between players with better and worse performance are presented in Table 2. No significant differences were found in any external load indicators (p > 0.05).

4. Discussion

The purpose of this study was to evaluate the relationship between external load and game performance in female 3 × 3 basketball players during the Olympic Games, and to determine whether external load can distinguish between players’ performance levels. The results showed that, in elite-level female 3 × 3 basketball, correlations between external load indicators and most game performance metrics were trivial, with only a few significant correlations falling within the small range. Comparatively, ACC was significantly correlated with the largest number of performance indicators, while REB was significantly associated with the greatest number of external load indicators. No significant differences in external load were found between players with better and worse performance. Overall, these findings provide only partial support for our hypotheses: while some small associations were observed, most relationships were weaker than expected, and external load did not distinguish between performance groups. These results may suggest that external load alone cannot accurately predict game performance, and that multidimensional factors and contextual conditions should be considered in evaluating players’ performance.
To our knowledge, this is the first study to evaluate the relationship between external load and game performance in 3 × 3 basketball. Therefore, comparisons can only be made with previous studies conducted in traditional 5 × 5 basketball. Current research distinguishes between training load and game load, with no clear relationship identified between training load and game performance. Specifically, Vázquez-Guerrero found that, in professional basketball, external training load was not related to Winscore, performance index rating, or player total contribution [26]. Two additional studies on semi-professional players reached similar conclusions, showing that higher training loads did not correspond to better game performance [23,46]. In contrast, studies on game load have reported divergent results across different populations. Studies on female and youth players have mostly reported associations between game load and game performance. One study in female collegiate basketball reported moderate-to-large correlations between game load and specific performance indicators, including points, field goals, and free-throw metrics [47]. Similarly, research on semi-professional female players reported moderate-to-large correlations between external load and basic technical statistics (steals, turnovers, points, shot attempts), with additional moderate associations observed in advanced statistics (number of possessions, points per possession, percentage of offensive rebounds) [48]. Furthermore, Piñar reported that players with a higher Performance Index Rating (PIR) accumulated greater loads during games [49]. For youth samples, a study on U19 male players also reported similar findings, identifying a moderate correlation between PlayerLoad and player index rating [25]. Studies on collegiate male players have likewise reported that increased game load was significantly associated with game statistics [50]. However, at the professional men’s level, García collected in-game external load indicators, including total distance, high-speed running, and number of accelerations, along with performance measures such as Performance Index Rating, player total contribution, and points scored, and found correlations between external load and performance indicators ranging from trivial to small [24]. The contrasting findings may be explained by differences in game structure and determining factors across competitive levels. Professional men’s basketball games exhibit greater intensity of physical confrontation and higher tactical complexity compared with semi-professional, collegiate, or female basketball. In elite-level competitive contexts, game performance tends to rely more heavily on players’ control of finer details, such as accuracy of tactical execution, quality of in-game decision-making, and technical skill execution [51], thereby limiting the explanatory power of external load.
Our findings are consistent with those reported in professional 5 × 5 basketball research [24]. Despite differences in court dimensions, player numbers, game characteristics, and rules, the absolute physical output data may differ between the two formats, and existing studies have confirmed differences in physical demands [1,3]. However, it should not be overlooked that both formats are high-intensity intermittent sports and share similar technical and tactical actions during competition [1]. Considering that the sample in this study was drawn from the Olympic Games, it effectively represents the highest competitive level and elite players in 3 × 3 basketball. Therefore, at the elite level, 3 × 3 and 5 × 5 basketball appear to exhibit similar relationships between physical output and game performance. In fact, available studies in 3 × 3 basketball suggest that external load differs only marginally between winning and losing games [14]. Similarly, Sansone found that although players reported higher perceived fatigue in lost games, average and peak heart rates showed little difference between wins and losses [2]. This may indicate that the physical and physiological demands imposed by the high intensity and fast pace of 3 × 3 basketball are experienced by every player on the court, while the smaller court size restricts players’ movement range, making these demands more uniform. Therefore, performance evaluation should adopt a holistic perspective, as external load may only reflect the degree of participation rather than serving as a decisive variable directly determining performance.
Among all external load indicators in this study, ACC showed significant correlations with the greatest number of performance metrics, including PTS, REB, 1PT made, KAS, and DRV. Nevertheless, it should be emphasized that these correlations remained within the small range, limiting their explanatory power. From a practical perspective, ACC may better reflect the high-intensity confrontations and explosive actions specific to 3 × 3 basketball compared with other external load indicators. As this sport heavily relies on neuromuscular actions (e.g., jumping, changes in direction, accelerations, and decelerations), previous studies have also emphasized the importance of instantaneous high-intensity actions and neuromuscular training in 3 × 3 basketball [1,2,6]. Due to the smaller court size and the absence of full-court transitions, sudden stops, starts, and changes in direction appear to play a relatively important role in half-court transitions and scoring opportunities. Meanwhile, REB emerged as the performance variable most correlated with external load indicators, including total distance and distances across different speed zones. This suggests that rebounding is not only related to players’ strength and physical contests but may also be associated with their overall movement coverage on the court. However, similar to ACC, these correlations remained small, implying that external load has limited explanatory power for rebounding performance. Overall, although ACC and REB demonstrated greater sensitivity than other indicators, they should be interpreted with caution in practice, as they are not strong predictive tools and cannot be used independently to assess game performance.
When examining the relationship between external load and performance, we found that external load could not distinguish between players with better or worse performance. This finding contrasts with research in traditional 5 × 5 basketball, where time-window analyses comparing elite and non-elite male players revealed that non-elite players experienced greater external demands, attributed to lower movement efficiency and differences in tactical adjustment capacity [10]. The discrepancy may be due to the characteristics of 3 × 3 basketball, in which every player on the court is required to take on comprehensive roles including playmaking, driving, and defending. Such role blurring and balanced task distribution further contribute to the convergence of players’ physical output patterns. In fact, when quantifying individual differences in external load, traditional 5 × 5 basketball often uses player position as a basis for classification. For example, available studies indicate that perimeter players exhibit significantly lower acceleration and deceleration rates compared with forwards and centers, while forwards display the lowest total external load in competition [52]. Conversely, in youth competitions, frontcourt players were found to sustain higher external loads than backcourt players [53]. Compared with other positions, centers showed statistical performance more closely related to the physical loads sustained in competition, which was explained by their reliance on physical displacement during offensive actions [54]. These findings provide useful references for designing position-specific training programs and contribute to more precise evaluations of player performance. However, such designs are difficult to implement in 3 × 3 basketball research, as players do not have clearly defined court positions. To further clarify the relationship between external load and performance in 3 × 3 basketball, future studies could consider grouping players based on functional roles, tactical–technical tasks, or clustering approaches, thereby providing a better explanation of external load contributions to performance.
Methodologically, this study is the first to apply a markerless motion capture system in real match settings of 3 × 3 basketball. Inertial systems are restricted by regulations in official international competitions, and traditional optoelectronic systems (e.g., Vicon) require lab environments and are expensive. In contrast, markerless motion capture is characterized by its non-invasive nature and operational flexibility, offering researchers and practitioners a feasible means of acquiring external load data during elite competitions like the Olympics. In particular, it has demonstrated strong accuracy and consistency in 3 × 3 basketball tasks [37]. This method is considered a practical solution for dynamic and complex team sport environments, with strong potential for broader implementation and scalability [55].

Practical Applications, Limitations, and Future Directions

From a practical perspective, the findings of this study suggest that coaches and strength and conditioning practitioners should exercise caution when interpreting game performance using external load data. External load alone is insufficient as a determinant of performance quality and should instead be primarily applied to regulate training volume and monitor physical condition. The continuous monitoring of indicators such as PlayerLoad, running distance, and accelerations can assist coaches in appropriately distributing training and recovery cycles, thus preventing fatigue accumulation or overtraining from congested schedules. Moreover, integrating internal load indicators (e.g., heart rate, lactate, RPE) can more comprehensively reflect players’ physiological stress, offering more reliable references for pre-game preparation and in-game load management.
This study has several limitations. First, the sample included only female players from the Olympic Games. Although the sample represents the highest level of 3 × 3 basketball competition, the findings should be cautiously generalized to male players or other competitive levels. Second, no internal load data were collected. Multimodal data would provide deeper insights into the multidimensional determinants of player performance. Furthermore, the moderating effects of game context were not considered, although such factors have been widely applied in prior performance analyses. Future studies should aim to establish a more comprehensive monitoring framework. At the same time, given the specific scheduling characteristics of 3 × 3 basketball, exploring load and performance fluctuations under different contextual conditions may hold greater practical value.

5. Conclusions

This study is the first to use a markerless motion capture system to quantify external load in female 3 × 3 basketball players during the Olympic Games and to evaluate its relationship with game performance. The results provided no clear evidence of an association between external load and game performance, nor were significant differences in external load observed between players of differing performance levels. These results suggest that, in elite-level 3 × 3 basketball, external load primarily reflects participation and effort rather than serving as a key determinant of performance. This study provides new empirical evidence on the characteristics of 3 × 3 basketball, indicating that coaches and strength and conditioning practitioners should be cautious in using external load as a performance evaluation tool, as it may be more appropriate for training regulation and fatigue monitoring.

Author Contributions

Conceptualization, M.Q., R.D. and M.Z.; methodology, M.Q. and J.T.; formal analysis, M.Q., R.D. and J.T.; investigation, R.D. and Z.L.; resources, M.Z.; data curation, M.Q., R.D. and W.Z.; writing—original draft preparation, M.Q. and R.D.; writing—review and editing, all authors; visualization, J.T., Z.L. and W.Z.; supervision, M.Z.; project administration, M.Z. All authors have read and agreed to the published version of the manuscript.

Funding

This research was supported by the Shanghai Municipal Education Commission (Project: Research on Innovative Applications of AI-Enabled Technical–Tactical Analysis in Ball Sports, No. A1-N3017-24-0013-3), the Shanghai Key Laboratory of Human Performance (Shanghai University of Sport, No. 11DZ2261100), and the Young Faculty Research Development Fund of South China Normal University (Grant No. 24SK14).

Institutional Review Board Statement

This study was conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Shanghai University of Sport [approval number: 102772025RT096].

Informed Consent Statement

Not applicable.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors on request.

Conflicts of Interest

The authors declare no conflicts of interest.

References

  1. Sansone, P.; Conte, D.; Tessitore, A.; Rampinini, E.; Ferioli, D. A Systematic Review on the Physical, Physiological, Perceptual, and Technical–Tactical Demands of Official 3 × 3 Basketball Games. Int. J. Sports Physiol. 2023, 18, 1233–1245. [Google Scholar] [CrossRef] [PubMed]
  2. Sansone, P.; Tessitore, A.; Makivić, B.; Ferioli, D.; Conte, D. The Relationships Between Training-Load Models in 3 × 3 Basketball Games. Int. J. Sports Physiol. Perform. 2025, 1, 684–688. [Google Scholar] [CrossRef] [PubMed]
  3. Reina, M.; Rubio, J.G.; Antúnez, A.; Ibáñez, S.J. Comparación de la carga interna y externa en competición oficial de 3 vs. 3 y 5 vs. 5 en baloncesto femenino (Comparison of internal and external load in official 3 vs. 3 and 5 vs. 5 female basketball competitions). Retos 2020, 37, 400–405. [Google Scholar] [CrossRef]
  4. Montgomery, P.G.; Maloney, B.D. 3 × 3 Basketball: Performance Characteristics and Changes During Elite Tournament Competition. Int. J. Sports Physiol. Perform. 2018, 13, 1349–1356. [Google Scholar] [CrossRef]
  5. Montgomery, P.G.; Maloney, B.D. Three-by-Three Basketball: Inertial Movement and Physiological Demands During Elite Games. Int. J. Sports Physiol. Perform. 2018, 13, 1169–1174. [Google Scholar] [CrossRef]
  6. Sansone, P.; Rago, V.; Alonso-Perez-Chao, E.; Zhang, S.; Conte, D. Physical Demands and Acute Neuromuscular Responses Across a Single-Day 3 × 3 Male Basketball Tournament. Sensors 2025, 25, 3296. [Google Scholar] [CrossRef]
  7. Fox, J.L.; Scanlan, A.T.; Stanton, R. A Review of Player Monitoring Approaches in Basketball: Current Trends and Future Directions. J. Strength Cond. Res. 2017, 31, 2021. [Google Scholar] [CrossRef]
  8. Tuttle, M.C.; Power, C.J.; Dalbo, V.J.; Scanlan, A.T. Intensity Zones and Intensity Thresholds Used to Quantify External Load in Competitive Basketball: A Systematic Review. Sports Med. 2024, 54, 2571–2596. [Google Scholar] [CrossRef]
  9. Kanaras, V.; Michailidis, Y.; Mandroukas, A.; Stafylidis, A.; Vardakis, L.; Kyranoudis, A.E.; Christoulas, K.; Gissis, I.; Metaxas, T.I. Weekly External Load Correlation in Season Microcycles with Game Running Performance and Training Quantification in Elite Young Soccer Players. Sensors 2024, 24, 4523. [Google Scholar] [CrossRef]
  10. Antoranz, Y.; Alonso-Pérez-Chao, E.; Tejero-González, C.M.; Salazar, H.; del Campo-Vecino, J.; Jiménez-Sáiz, S.L. Understanding External Peak Demands in Elite vs. Non-Elite Male Basketball Players. Sports 2025, 13, 179. [Google Scholar] [CrossRef]
  11. Castillo, D.; Raya-González, J.; Clemente, F.; Conte, D.; Rodríguez-Fernández, A. The Effects of Defensive Style and Final Game Outcome on the External Training Load of Professional Basketball Players. Biol. Sport 2020, 38, 483–490. [Google Scholar] [CrossRef]
  12. Power, C.J.; Fox, J.L.; Dalbo, V.J.; Scanlan, A.T. External and Internal Load Variables Encountered During Training and Games in Female Basketball Players According to Playing Level and Playing Position: A Systematic Review. Sports Med. Open 2022, 8, 107. [Google Scholar] [CrossRef]
  13. Ferioli, D.; Conte, D.; Rucco, D.; Alcaraz, P.E.; Vaquera, A.; Romagnoli, M.; Rampinini, E. Physical Demands of Elite Male and Female 3 × 3 International Basketball Matches. J. Strength Cond. Res. 2023, 37, e289. [Google Scholar] [CrossRef] [PubMed]
  14. Ferioli, D.; Rampinini, E.; Conte, D.; Rucco, D.; Romagnoli, M.; Scanlan, A. Physical Demands during 3 × 3 International Male and Female Basketball Games Are Partially Impacted by Competition Phase but Not Game Outcome. Biol. Sport 2023, 40, 377–387. [Google Scholar] [CrossRef] [PubMed]
  15. Colyer, S.L.; Evans, M.; Cosker, D.P.; Salo, A.I.T. A Review of the Evolution of Vision-Based Motion Analysis and the Integration of Advanced Computer Vision Methods Towards Developing a Markerless System. Sports Med.-Open 2018, 4, 24. [Google Scholar] [CrossRef] [PubMed]
  16. Suo, X.; Tang, W.; Li, Z. Motion Capture Technology in Sports Scenarios: A Survey. Sensors 2024, 24, 2947. [Google Scholar] [CrossRef]
  17. van der Kruk, E.; Reijne, M.M. Accuracy of Human Motion Capture Systems for Sport Applications; State-of-the-Art Review. Eur. J. Sport Sci. 2018, 18, 806–819. [Google Scholar] [CrossRef]
  18. Ceriola, L.; Taborri, J.; Donati, M.; Rossi, S.; Patanè, F.; Mileti, I. Comparative Analysis of Markerless Motion Capture Systems for Measuring Human Kinematics. IEEE Sens. J. 2024, 24, 28135–28144. [Google Scholar] [CrossRef]
  19. Tang, H.; Munkasy, B.; Li, L. Differences between Lower Extremity Joint Running Kinetics Captured by Marker-Based and Markerless Systems Were Speed Dependent. J. Sport Health Sci. 2024, 13, 569–578. [Google Scholar] [CrossRef]
  20. Edwards, N.A.; Caccese, J.B.; Tracy, R.E.; Hagen, J.; Quatman-Yates, C.C.; Oñate, J. The Validity and Usability of Markerless Motion Capture and Inertial Measurement Units for Quantifying Dynamic Movements. Med. Sci. Sports Exerc. 2025, 57, 641–655. [Google Scholar] [CrossRef]
  21. Emmerson, J.; Needham, L.; Evans, M.; Williams, S.; Colyer, S. Comparison of Markerless and Marker-Based Motion Capture for Estimating External Mechanical Work in Tennis: A Pilot Study. Int. Soc. Biomech. Sports 2023, 41, 27. [Google Scholar]
  22. Santiano, F.; Ison, S.; Emmerson, J.; Colyer, S. Using Markerless Motion Analysis to Quantify Sex and Discipline Differences in External Mechanical Work during Badminton Match Play. J. Sports Sci. 2025, 43, 1158–1166. [Google Scholar] [CrossRef] [PubMed]
  23. Fox, J.L.; Stanton, R.; O’Grady, C.J.; Teramoto, M.; Sargent, C.; Scanlan, A. Are Acute Player Workloads Associated with In-Game Performance in Basketball? Biol. Sport 2021, 39, 95–100. [Google Scholar] [CrossRef] [PubMed]
  24. García, F.; Fernández, D.; Martín, L. Relationship Between Game Load and Player’s Performance in Professional Basketball. Int. J. Sports Physiol. Perform. 2022, 17, 1473–1479. [Google Scholar] [CrossRef] [PubMed]
  25. Li, G.; Shang, L.; Qin, S.; Yu, H. The Impact of Internal and External Loads on Player Performance in Chinese Basketball Association. BMC Sports Sci. Med. Rehabil. 2024, 16, 194. [Google Scholar] [CrossRef]
  26. Vázquez-Guerrero, J.; Casals, M.; Corral-López, J.; Sampaio, J. Higher Training Workloads Do Not Correspond to the Best Performances of Elite Basketball Players. Res. Sports Med. 2020, 28, 540–552. [Google Scholar] [CrossRef]
  27. Conte, D.; Favero, T.G.; Lupo, C.; Francioni, F.M.; Capranica, L.; Tessitore, A. Time-Motion Analysis of Italian Elite Women’s Basketball Games: Individual and Team Analyses. J. Strength Cond. Res. 2015, 29, 144–150. [Google Scholar] [CrossRef]
  28. Conte, D.; Straigis, E.; Clemente, F.M.; Gómez, M.-Á.; Tessitore, A. Performance Profile and Game-Related Statistics of FIBA 3x3 Basketball World Cup 2017. Biol. Sport 2019, 36, 149–154. [Google Scholar] [CrossRef]
  29. Ferioli, D.; Conte, D.; Scanlan, A.T.; Vaquera, A. Technical-Tactical Demands of 3 × 3 International Basketball Games According to Game Outcome, Player Sex, and Competition Phase. J. Strength Cond. Res. 2023, 37, 403–412. [Google Scholar] [CrossRef]
  30. Lin, T.-Y.; Maire, M.; Belongie, S.; Bourdev, L.; Girshick, R.; Hays, J.; Perona, P.; Ramanan, D.; Zitnick, C.L.; Dollár, P. Microsoft COCO: Common Objects in Context. In Computer Vision—ECCV 2014, Proceedings of the Computer Vision—ECCV 2014, Zurich, Switzerland, 8–11 September 2014; Fleet, D., Pajdla, T., Chiele, B., Tuytelaars, T., Eds.; Springer: Cham, Switzerland, 2015. [Google Scholar]
  31. Jiang, T.; Lu, P.; Zhang, L.; Ma, N.; Han, R.; Lyu, C.; Li, Y.; Chen, K. RTMPose: Real-Time Multi-Person Pose Estimation Based on MMPose. arXiv 2023, arXiv:2303.07399. [Google Scholar] [CrossRef]
  32. Hartley, R.; Zisserman, A. Multiple View Geometry in Computer Vision; Cambridge University Press: Cambridge, UK, 2003; ISBN 978-0-521-54051-3. [Google Scholar]
  33. Cai, Y.; Wang, Z.; Luo, Z.; Yin, B.; Du, A.; Wang, H.; Zhang, X.; Zhou, X.; Zhou, E.; Sun, J. Learning Delicate Local Representations for Multi-Person Pose Estimation. In Computer Vision—ECCV 2020, Proceedings of the Computer Vision—ECCV 2020, Glasgow, UK, 23–28 August 2020; Vedaldi, A., Bischof, H., Brox, T., Frahm, J.-M., Eds.; Springer International Publishing: Cham, Switzerland, 2020; pp. 455–472. [Google Scholar]
  34. Cao, D.; Sun, Q.; Cui, T. Research on an Improved RTMPose Model for Evaluating Dance Standardization Scores. In Proceedings of the 14th International Conference on Information Science and Technology (ICIST), Chengdu, China, 6–9 December 2024. [Google Scholar]
  35. Guo, X.; Zhao, H. A Study on Gait Analysis Based on the Pose Estimation Algorithm RTMPose Using Knee Angle as a Metric. In Proceedings of the 2024 IEEE International Conference on Mechatronics and Automation (ICMA), Tianjin, China, 4–7 August 2024; pp. 573–578. [Google Scholar]
  36. He, S.; Wei, M.; Meng, D.; Lv, Z.; Guo, H.; Yang, G.; Wang, Z. Adversarially Trained RTMpose: A High-Performance, Non-Contact Method for Detecting Genu Valgum in Adolescents. Comput. Biol. Med. 2024, 183, 109214. [Google Scholar] [CrossRef] [PubMed]
  37. Zheng, W.; Zhang, M.; Dong, R.; Qiu, M.; Wang, W. Feasibility and Accuracy of an RTMPose-Based Markerless Motion Capture System for Single-Player Tasks in 3 × 3 Basketball. Sensors 2025, 25, 4003. [Google Scholar] [CrossRef] [PubMed]
  38. Yang, Y.; Wang, W.; Liu, Y.; Dong, L.; Wu, H.; Zhang, M.; Zhong, Z.; Sun, X. SGA-INTERACT: A 3D Skeleton-Based Benchmark for Group Activity Understanding in Modern Basketball Tactic. arXiv 2025, arXiv:2503.06522. [Google Scholar]
  39. Ben Abdelkrim, N.; El Fazaa, S.; El Ati, J. Time–Motion Analysis and Physiological Data of Elite under-19-Year-Old Basketball Players during Competition. Br. J. Sport Med. 2007, 41, 69–75. [Google Scholar] [CrossRef]
  40. Ben Abdelkrim, N.; Castagna, C.; El Fazaa, S.; El Ati, J. The Effect of Players’ Standard and Tactical Strategy on Game Demands in Men’s Basketball. J. Strength Cond. Res. 2010, 24, 2652. [Google Scholar] [CrossRef]
  41. Willberg, C.; Wellm, D.; Behringer, M.; Zentgraf, K. Analyzing Acute and Daily Load Parameters in Match Situations—A Comparison of Classic and 3 × 3 Basketball. Int. J. Sports Sci. Coach. 2023, 18, 207–219. [Google Scholar] [CrossRef]
  42. Batterham, A.M.; Hopkins, W.G. Making Meaningful Inferences About Magnitudes. Int. J. Sport Physiol. 2006, 1, 50–57. [Google Scholar] [CrossRef]
  43. Sampaio, J.; Lago, C.; Casais, L.; Leite, N. Effects of Starting Score-Line, Game Location, and Quality of Opposition in Basketball Quarter Score. Eur. J. Sport Sci. 2010, 10, 391–396. [Google Scholar] [CrossRef]
  44. Sampaio, J.; Lago, C.; Drinkwater, E.J. Explanations for the United States of America’s Dominance in Basketball at the Beijing Olympic Games (2008). J. Sport Sci. 2010, 28, 147–152. [Google Scholar] [CrossRef]
  45. Chen, R.; Zhang, M.; Xu, X.; Liu, Y. Game-Related Statistics for Distinguishing Winning and Losing Teams in Olympic Basketball: The Impact of Game Pace. J. Sports Sci. 2024, 42, 2541–2552. [Google Scholar] [CrossRef]
  46. Sansone, P.; Gasperi, L.; Tessitore, A.; Gomez, M. Training Load, Recovery and Game Performance in Semiprofessional Male Basketball: Influence of Individual Characteristics and Contextual Factors. Biol. Sport 2020, 38, 207–217. [Google Scholar] [CrossRef]
  47. Brown, F.S.A.; Fields, J.B.; Jagim, A.R.; Baker, R.E.; Jones, M.T. Analysis of In-Season External Load and Sport Performance in Women’s Collegiate Basketball. J. Strength Cond. Res. 2024, 38, 318. [Google Scholar] [CrossRef]
  48. Espasa-Labrador, J.; Martínez-Rubio, C.; Oliva-Lozano, J.M.; Calleja-González, J.; Carrasco-Marginet, M.; Fort-Vanmeerhaeghe, A. Relationship between Physical Demands and Player Performance in Professional Female Basketball Players Using Inertial Movement Units. Sensors 2024, 24, 6365. [Google Scholar] [CrossRef] [PubMed]
  49. Piñar, M.I.; García, D.; Mancha-Triguero, D.; Ibáñez, S.J. Effect of Situational and Individual Factors on Training Load and Game Performance in Liga Femenina 2 Basketball Female Players. Appl. Sci. 2022, 12, 7752. [Google Scholar] [CrossRef]
  50. Olthof, S.B.H.; Tureen, T.; Tran, L.; Brennan, B.; Winograd, B.; Zernicke, R.F. Biomechanical Loads and Their Effects on Player Performance in NCAA D-I Male Basketball Games. Front. Sports Act. Living 2021, 3, 670018. [Google Scholar] [CrossRef] [PubMed]
  51. Ibáñez, S.J.; Sampaio, J.; Feu, S.; Lorenzo, A.; Gómez, M.A.; Ortega, E. Basketball Game-Related Statistics That Discriminate between Teams’ Season-Long Success. Eur. J. Sport Sci. 2008, 8, 369–372. [Google Scholar] [CrossRef]
  52. Vázquez-Guerrero, J.; Casamichana, D.; Rodas, G. Comparing External Total Load, Acceleration and Deceleration Outputs in Elite Basketball Players across Positions during Match Play. Kinesiology 2018, 50, 228–234. [Google Scholar] [CrossRef]
  53. Sansone, P.; Ceravolo, A.; Tessitore, A. External, Internal, Perceived Training Loads and Their Relationships in Youth Basketball Players Across Different Positions. Int. J. Sports Physiol. Perform. 2022, 17, 249–255. [Google Scholar] [CrossRef]
  54. Espasa-Labrador, J.; Martínez-Rubio, C.; García, F.; Fort-Vanmeergaehe, A.; Guarch, J.; Calleja-González, J. Exploring the Relationship Between Game Performance and Physical Demands in Youth Male Basketball Players. J. Funct. Morphol. Kinesiol. 2025, 10, 293. [Google Scholar] [CrossRef]
  55. Noorbhai, H.; Moon, S.; Fukushima, T. A Conceptual Framework and Review of Multi-Method Approaches for 3D Markerless Motion Capture in Sports and Exercise. J. Sports Sci. 2025, 43, 1167–1174. [Google Scholar] [CrossRef]
Sensors 25 06334 g001
Figure 1. Schematic diagram of video recording setup.
Figure 1. Schematic diagram of video recording setup.
Sensors 25 06334 g001
Sensors 25 06334 g002
Figure 2. Heatmap of the correlation between external load and player performance. * Indicate p < 0.05.
Figure 2. Heatmap of the correlation between external load and player performance. * Indicate p < 0.05.
Sensors 25 06334 g002
Table 1. Game-related statistics used to represent player performance in 3 × 3 basketball.
Table 1. Game-related statistics used to represent player performance in 3 × 3 basketball.
VariableAbbreviationDefinition
PointsPTSNo. of points scored by a team.
ReboundsREBNo. of controlled recoveries of a live ball by a team being entitled to the ball for a throw-in after a missed shot or the last free-throw.
HighlightHGLA comprehensive evaluation variable; the calculation formula is KAS + DRV + DNK + BS + BZR
Shooting efficiencyS-EFFThe realized points-per-shot value for a shot. It is calculated by dividing the points made with the no. of shots attempted; the calculation formula is PTS/(1PTA + 2PTA + FTA)
1-point shots made1PT madeNo. of 1-point shots attempt for which 1 point was awarded.
1-point shots missed1PT missedNo. of 1-point shots missed.
1-point shots attempt1PT attemptNo. of 1-point shots attempted (both made and missed 1-point shots, excluding free-throws).
1-point %1PT%The shooting percentage for 1-point shots; the calculation formula is 1PT made/1PT attempt.
2-point shots made2PT madeNo. of 2-point shots attempt for which 2 points were awarded.
2-point shots missed2PT missedNo. of 2-point shots missed.
2-point shots attempt2PT attemptNo. of 2-point shots attempted (both made and missed 2-point shots, excluding free-throws).
2-point %2PT%The shooting percentage for 2-point shots; the calculation formula is 2PT made/2PT attempt.
Free-throws madeFT madeNo. of free-throw shots attempt for which 1 point was awarded.
Free-throws missedFT missedNo. of free-throw shots missed.
Free-throws attemptFT attemptNo. of free-throw shots attempted (both made and missed free-throw shots).
Free-throw %FT%The shooting percentage for free-throw shots; the calculation formula is FT made/FT attempt
Key assistsKASNo. of passes that give a teammate a positional advantage to score a basket directly from the paint.
DrivesDRVNo. of quick or skillful dribbling moves used to move from behind the two-point arc to directly score a field goal from the restricted area.
DunksDNKNo. of shots made in which the ball is forced by the player downwards into the basket with one or both hands.
Blocked shotsBSNo. of blocked shots that take place when a defending player contacts the ball to alter the flight of the shot attempt of the opponent, and the shot is missed.
Buzzer beatersBZRNo. of last shots of the overtime, or tie-breaking, tie-forcing, or lead-changing last-made shots during the last 5 continuously played seconds of regular time (regardless of whether the game was won at the end of 10 min or before the limit, if a team reached 21 points).
Offense ReboundsOREBNo. of rebounds awarded when the ball possession is retained by the same team who missed the field goal or the free throw.
Defense ReboundsDREBNo. of rebounds awarded when the ball possession is gained by the other team who missed the field goal or the free throw.
TurnoversTONo. of mistakes committed by a team that results in the defensive team gaining possession of the ball.
PossessionsPOSNo. of times a team had the ball and produced one of the following possession outcomes: 1-point shot, 2-point shot, trip to the free-throw line, or turnover.
Points per possessionPPPNo. of points a team scored per possession.
1-point possessionsPOS-1PTNo. of times a team had the ball and attempted a 1-point shot.
1-point points per possessionPPP-1PTNo. of points a team scored per 1-point possession.
Free-throw possessionsPOS-FTNo. of times a team had the ball and secured a trip to the free-throw line.
Free-throw points per possessionPPP-FTNo. of points a team scored per free-throw possession (excluding extra free-throw possessions).
2-point possessionsPOS-2PTNo. of times a team had the ball and attempted a 2-point shot.
2-point points per possessionPPP-2PTNo. of points a team scored per 2-point possession.
Table 2. The difference in external load between better and worse performance.
Table 2. The difference in external load between better and worse performance.
External Load MetricsBetter Performance
(n = 63)		Worse Performance
(n = 128)		Fp
MeanSDMeanSD
PlayerLoad13.6181.30313.8341.5571.5970.211
Total Distance104.7298.848104.97711.222<0.0010.979
LIA62.9266.24262.4958.8860.6600.418
MIA36.3695.36036.8688.9300.2540.615
HIA5.4352.3625.6153.5431.0020.318
ACC4.9892.1254.8882.8570.1390.709
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Qiu, M.; Dong, R.; Tao, J.; Li, Z.; Zheng, W.; Zhang, M. The Relationship Between External Load and Player Performance in Elite Female 3 × 3 Basketball Games: A Markerless Motion Capture Approach. Sensors 2025, 25, 6334. https://doi.org/10.3390/s25206334

AMA Style

Qiu M, Dong R, Tao J, Li Z, Zheng W, Zhang M. The Relationship Between External Load and Player Performance in Elite Female 3 × 3 Basketball Games: A Markerless Motion Capture Approach. Sensors. 2025; 25(20):6334. https://doi.org/10.3390/s25206334

Chicago/Turabian Style

Qiu, Mingjia, Rui Dong, Junye Tao, Zhaoyu Li, Wen Zheng, and Mingxin Zhang. 2025. "The Relationship Between External Load and Player Performance in Elite Female 3 × 3 Basketball Games: A Markerless Motion Capture Approach" Sensors 25, no. 20: 6334. https://doi.org/10.3390/s25206334

APA Style

Qiu, M., Dong, R., Tao, J., Li, Z., Zheng, W., & Zhang, M. (2025). The Relationship Between External Load and Player Performance in Elite Female 3 × 3 Basketball Games: A Markerless Motion Capture Approach. Sensors, 25(20), 6334. https://doi.org/10.3390/s25206334

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop