henhaobio
Sign in

Frontiers in Sports and Active LivingAnimal Behavior & CognitionPeer-reviewed journal article

Analyzing coordinated group behavior through role-sharing: A pilot study in female 3-on-3 basketball with practical application

Jun IchikawaiD, Masatoshi Yamada, Keisuke FujiiiD

PDFDOIPreprint
In-depth full-text digest available · Jump to digest ↓

Abstract

A group often shares a common goal and accomplishes a task that is difficult to complete alone by distributing roles. In such coordination, the nonverbal behavior among three or more members complicates the explanation of the mechanism due to complex and dynamic interactions. In cognitive science, a crucial role is indicated: to intervene moderately with others and adjust the whole balance without interrupting their main smooth interactions, using an experimental task. The findings suggest that resilient helping actions in the third role support to coordination. These actions are similar to off-ball movements in team sports, which involve an on-ball player and have recently been the focus of sports science because their characteristics are not represented in common statistical data, such as a shooting success rate. Hence, a new perspective for discussing coordination has emerged, as existing theories, such as synchronization--where movements between players are spontaneously matched and organized--cannot explain the mentioned role. However, there is a lack of investigation and discussion regarding whether these findings are applicable to real-world activities. Therefore, this study applied the experimental findings to the field of sports. We developed a 3-on-3 basketball game in which the offensive role of intervention decision and adjustment is key for winning and introduced it to the practice of a female university team as a pilot study. Participants repeatedly engaged in the mini-game, and the playing was compared before and after receiving tips of this role. Consequently, in the bins of the relatively large distance between the participant required to the relevant role and each defensive player, the frequencies after receiving these tips were significantly higher. Furthermore, the winning rate on the offensive team improved temporarily; however, the effects were not maintained. These suggest that spacing skill, which maintains reasonable distances from the other players, creates favorable situations for coordination. This study may bridge the gap between controlled experiments and real-world applications and make an educational contribution; it may recommend practice design for the acquisition of spacing skill related to the crucial role.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 3 (8 Apr 2025) · CC BY 4.0

Coach tips on an intervene-and-adjust offensive role in 3-on-3 basketball shifted the key player toward larger distances from defenders, yet offensive winning rate rose only temporarily.

Earlier cognitive science work found that a third triad member who intervenes sparingly and rebalances the group without disrupting the main pair supports coordination, a role synchronization accounts cannot explain, yet transfer to real-world settings was untested. This pilot maps that role onto off-ball offensive play in 3-on-3 basketball and tests whether coaching it alters spacing and winning.

Key results

  1. SuggestedAfter role-specific tips, offensive #3 spent proportionally less time near 400 cm from defenders and more time at 1200 and 1400 cm, consistent with keeping space away from defensive pressure.

    400 cm bin: t(20) = 2.739, p = 0.013, d = 0.894, power 0.973 (lower after tips); 1200 cm: t(20) = −2.269, p = 0.034, d = 0.732, power 0.891; 1400 cm: t(20) = −2.581, p = 0.018, d = 0.668, power 0.829 (both higher); 21 trials per half, 200 cm bins, no Bonferroni correction. · 3.2 Analysis Procedures of Playing Related to…; 3.3 Results of the Distance with Each Defensive…

  2. ShownOffensive team success did not change significantly between halves; the winning rate climbed after the tips, peaking in Session 5, but fell again in Session 6, so the performance gain was not sustained.

    Offensive wins 9 vs 11 over 21 trials per half; χ²(1) = 0.095, p = 0.757, φ = 0.048. Sessions 1 and 2 each had four wins in seven trials, Session 3 dropped, Session 5 reached six wins, Session 6 declined. · 3.1 Team Performance

  3. SuggestedDistances from #3 to the other players showed only non-significant trends forming a U-shape (more time at short and long distances, less at mid-range after tips), which the authors read as #3 holding position without crowding teammates.

    200 cm: t(20) = −2.002, p = 0.059, d = 0.517, power 0.616; 800 cm: t(20) = 1.856, p = 0.078, d = 0.628, power 0.782; 1600 cm: t(20) = −1.745, p = 0.096, d = 0.572, power 0.703; authors state no clear differences overall. · 3.3 Results of the Distance with Each Other…; 4 Discussion

  4. ClaimedThe late drop in offensive success was attributed to a defensive countermeasure: defensive #3 pretended to help teammates while staying on the role player, exploiting a standardized post-tip offensive pattern.

    Qualitative report from defensive players questioned after the practice; no quantitative measure of the countermeasure. Authors also cite a single practice and third-division skill level as factors. · 3.1 Team Performance; 4 Discussion

  5. ShownVideo-based tracking with manual correction gave position errors small relative to the 200 cm histogram bins, supporting the reliability of the distance analyses.

    ByteTrack on YOLOX applied to 20 fps, 1280 px × 720 px video, corrected in Labelbox; mean absolute error 3.165 cm × 2.102 cm vs 2.3 cm × 4.6 cm reported in a prior soccer triad study. · 3.2 Analysis Procedures of Playing Related to…

Methods, data and reproducibility

Pre/post field pilot: in a 3-on-3 mini-game won when a shooter is freed within 15 seconds, offensive #3 must judge when to intervene; 21 trials ran before and 21 after a guest coach's role tips. Bird's-eye video was tracked to player coordinates, and per-bin histogram frequencies of #3's distances were compared by t-tests, Cohen's d and power.

  • Female university basketball players (five regulars and the sixth man) from a Tokai league third-division team in Japan; mean age 19.17, height 165.08 cm, experience 10.58 years · human · n = six participants (fixed offensive and defensive trios)
  • Single-camera bird's-eye videos (20 fps, 1280 px × 720 px) of mini-game trials, three sessions of seven trials per half, before and after coaching tips · human · n = 21 trials per half
  • Tracked, manually corrected x-y position time series of all players after real coordinate transformation; the paper says only that relevant data are in Supplementary Datasets · human
Methods & tools
ByteTrack (multi-object tracking), YOLOX (object detector), Labelbox (manual correction of detections), MATLAB R2021b, R 4.3.3, G*Power 3.1, Sony HDR-CX680 video camera, Chi-square test on win/loss cross-tabulation, Per-bin t-tests with Cohen's d and statistical power
Reagents & environment
MATLAB R2021b (distance computation), R 4.3.3 (statistical tests), G*Power 3.1 (power analysis); ByteTrack on YOLOX with Labelbox correction, versions not stated
Reproducibility gaps
Data are stated to be in Supplementary Datasets with no URL or accession; no tracking or analysis code is shared; ball positions were not extracted; statistics for non-significant bins appear only in Supplementary Materials.
Limitations (6)
  • Only one female university team at third-division level took part, in a single practice; the authors call for replication with other teams of similar skill and with expert teams. · stated by the authors
  • Bonferroni correction was deliberately skipped across many histogram bins to avoid type II error, so per-bin t-tests carry inflated type I error risk; effect sizes and power were used to temper interpretation. · stated by the authors
  • The ball could not be detected in the recordings, so passing, dribbling and shooting contexts were not analyzed, and spacing skill was inferred from distance histograms rather than measured directly (e.g., Voronoi occupied space). · stated by the authors
  • Offensive gains faded once defenders adopted a countermeasure; the authors cite a single practice session and limited individual skill as reasons coordination did not fully develop. · stated by the authors
  • No control group or counterbalanced order: pre/post differences are confounded with session order and practice effects, and the defense also received help-defense tips at the same break. · noted in this summary
  • All distance findings rest on one player occupying the #3 role, with trials treated as replicates, so they describe one individual's adaptation rather than a generalizable effect. · noted in this summary

Read this if you study team coordination or off-ball movement and want an example of turning a lab triad role-sharing finding into a coached 3-on-3 basketball drill with tracking-based spacing analysis.

Citations
9
FWCI
8.21
Field percentile
top 2%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISChackochan '19Fanghella '26Yamamoto '23Kijima '26Ogino '26Lübbert '24
62% semantic matchNeuroscience
Vinil T. Chackochan, Vittorio Sanguineti·PLoS Computational Biology·Dec 12, 2019· 41 citations