📊 statistics

Movement synchronization in complex and dynamic team coordination: Relationship with mutual anticipation

This study utilizes Bayesian inference on 3-on-3 basketball field experiments to demonstrate that coaching advice significantly enhances the probability of movement synchronization among offensive players, thereby quantitatively linking improved team coordination to mutual anticipation.

Original authors: Jun Ichikawa, Mizuki Yokoyama, Soma Ishida, Genki Ichinose

Published 2026-06-23

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Jun Ichikawa, Mizuki Yokoyama, Soma Ishida, Genki Ichinose

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine a group of friends trying to dance together without music. If they wait to see what the others do before moving their own feet, they will constantly bump into each other or leave awkward gaps in the rhythm. But if they can "feel" the dance and predict where their partners will step next, they move as one fluid unit. This paper is about how a basketball team learns to do exactly that.

Here is a breakdown of the study using simple analogies:

The Problem: The "Wait-and-See" Trap

In team sports like basketball, players have to make split-second decisions. If a player waits to see their teammate move before they start moving themselves, it takes too long. It's like trying to walk through a crowded hallway while only stepping when you see someone else move; you'll get stuck or cause a pile-up. Opponents (the other team) are waiting for these gaps to sneak in and steal the ball.

To be fast and smooth, players need mutual anticipation. They need to know what their teammates are going to do before it happens, so they can move in perfect sync.

The Experiment: Coaching as a "Rehearsal"

The researchers studied a small group of female university basketball players playing 3-on-3 games. They set up a specific scenario:

The Goal: The offensive team had to create an open space for a shot within 15 seconds.
The Roles: Each player had a different job. One was the main dribbler, one was a helper, and the third was a "choreographer" who decided when to step in or step back to make space.
The Test: They played a series of games (the "pre-test") to see how they moved naturally. Then, a coach gave them specific advice on how to coordinate their roles. Afterward, they played again (the "post-test").

Think of the coach's advice as giving the dancers a specific routine to practice. Before the advice, they were improvising; after the advice, they had a plan.

The Measurement: The "Invisible String"

How do you measure if people are dancing in sync? The researchers used a mathematical trick called Relative Phase.

Imagine two people walking.

In-Phase (0 degrees): They step forward and backward at the exact same time.
Anti-Phase (180 degrees): When one steps forward, the other steps back. This is like a perfectly balanced seesaw or a dance where one partner moves in while the other moves out to make space.

The researchers tracked the players' movements frame-by-frame to see if their steps were "locked" into these patterns.

The Math: A "Gamble" Instead of a Coin Flip

Usually, scientists use strict tests (like flipping a coin) to say, "Yes, this change happened." But this study used a different approach called Bayesian Inference.

Think of it like this: Instead of asking, "Did the team get better?" (Yes/No), they asked, "What is the probability that the team is now moving in a more synchronized way?"
They calculated the odds that the players' movements after the coaching were more "in sync" than before.

The Results: The Team Found Their Rhythm

The study found that after the coach gave advice:

The Main Dribbler and the Helper started moving in Anti-Phase. When the dribbler moved in to attack, the helper moved out to clear space. They were perfectly balanced, like a seesaw, creating a smooth flow.
The Main Dribbler and the "Choreographer" started moving in In-Phase. They moved toward or away from each other at the same time, showing they were anticipating each other's moves perfectly.
The Probability: The researchers calculated that there was a 70% or higher chance that this improvement in synchronization was real and not just a lucky accident.

The Big Picture

The paper concludes that when a team shares a strategy (the "common ground" established by the coach), they stop waiting to see what others do. Instead, they start anticipating each other. This turns a group of individuals into a synchronized unit where movements flow without gaps.

The study didn't just say "they played better"; it used math to show how their bodies moved in a more coordinated, predictable pattern, proving that good coaching helps a team "dance" together rather than just run around.

Technical Summary: Movement Synchronization in Complex and Dynamic Team Coordination

Problem Statement
In team sports involving direct competition, efficient coordination relies on mutual anticipation, allowing players to interact implicitly and non-verbally without the delays caused by observing teammates before acting. While movement synchronization—a temporally stable structure in interactions—has been studied in simple systems (e.g., clapping, walking) and some sports, there is a lack of research investigating synchronization in real-world coordination characterized by heterogeneous roles and evolving interactions. Furthermore, existing evaluation methods often rely on null-hypothesis significance testing (e.g., t-tests, ANOVA) to compare frequency distributions against a baseline, which carries a risk of Type I errors and fails to provide a probabilistic quantification of how specific coordination structures differ between conditions.

Methodology
The study utilized a dataset from a field experiment involving 3-on-3 basketball mini-games conducted with six female university players. The experiment consisted of 42 trials (21 pre-advice and 21 post-advice) where an offensive team attempted to create scoring opportunities within a 15-second limit. A guest coach provided specific advice on offensive coordination strategies (e.g., "hand-off," "pick and roll," and specific positioning for an intervention player) between the pretest and posttest phases.

The analysis focused on extracting temporal movement components using the following steps:

Relative Phase Calculation: For each offensive pair, movement vectors were projected onto the inter-player distance vectors to determine if players were approaching or moving away from one another. These time series were standardized, and the Hilbert transform was applied to extract instantaneous movement phases. The relative phase ( $\Delta\phi$ ) between pairs was calculated, ranging from $-\pi$ to $\pi$ , where 0 degrees indicates in-phase synchronization (moving together) and $\pm180$ degrees indicates anti-phase synchronization (moving in opposition).
Bayesian Inference: Instead of traditional significance testing, the study employed Bayesian inference to estimate the frequency distributions of relative phases across 20-degree bins. A multinomial likelihood model with a Dirichlet prior was used to model the observed data. Markov chain Monte Carlo (MCMC) methods were utilized to estimate the posterior distributions of the average frequency vectors for both pretest and posttest conditions.
Probabilistic Comparison: The study calculated the probability $p(\delta_k > 0 | \text{data})$ for each bin, representing the likelihood that the frequency in the posttest condition exceeded that of the pretest condition. This allowed for a direct probabilistic assessment of synchronization trends without relying on null hypotheses.

Key Results
The analysis revealed distinct changes in movement synchronization following the coaching intervention:

Offensive #1 and #2: The probability that the posttest frequency exceeded the pretest frequency reached 70% or higher in bins around $\pm180$ degrees. This indicates a stronger tendency toward anti-phase synchronization, suggesting that when one player approached a space, the other moved away, facilitating effective space creation.
Offensive #1 and #3: The probability reached 70% or higher in bins around 0 degrees, indicating a shift toward in-phase synchronization. This suggests these players began moving in unison, likely reflecting the established strategy where the intervention player (Offensive #3) coordinated closely with the primary ball-handler (Offensive #1).
Offensive #2 and #3: While less distinct than the other pairs, there was a notable tendency (probability $\approx$ 0.95) for in-phase synchronization in the 20-degree bin during the posttest.
Overall: The probability of a synchronization trend emerging after advice compared to before advice was consistently 70% or higher for the key offensive pairs.

Significance and Claims
The authors claim that this study contributes to a quantitative understanding of coordination processes in complex, dynamic team environments. By introducing a statistical method that combines relative phase calculations with Bayesian inference, the study successfully probabilistically examined differences in synchronization without the limitations of null-hypothesis significance testing.

The findings suggest that the observed increase in synchronization is related to mutual anticipation based on team strategies established through coaching. The study posits that when players share a common ground (team strategy), they can anticipate each other's behaviors, resulting in smoother interactions and synchronized movements. The authors note that while the data comes from a single team, the approach offers a framework for quantitatively evaluating coordination that could be applied to other sports with similar tracking data. The study concludes that these findings may provide insights for sports practice, potentially serving as a guideline for determining the effectiveness of team strategies, though the authors remain modest regarding generalizability, calling for future measurements across various sports and competitive levels.