Skilled soccer players do not simply make better predictions than novices about where a pass will go — the way they move a computer cursor before committing to an answer reveals a hidden layer of preparation that conventional tests of anticipation have never captured. That is the central finding of a new proof-of-concept study published in Frontiers in Sports and Active Living, in which researchers at the University of Tokyo and Waseda University used continuous cursor tracking to read out the preparatory state of the mind in the moments before a decision was required.

Standard measures of perceptual-cognitive expertise in sport — response time, accuracy, or which option a player chooses — are collected only once, at the moment a response is made. They reveal the outcome of anticipation but say nothing about the process leading up to it. Two observers who eventually select the same option may differ in how far, and how early, they had already committed to it, and such differences cannot be recovered from accuracy or response time alone. The new study set out to capture this pre-response preparatory state by borrowing a method from basic research on reaching movements, where hand and cursor trajectories are known to reflect how competing options are weighed over time.

Twenty-four male Japanese university students took part: twelve skilled players from Division 1 of the Kanto Men’s University Soccer League, with an average of 12.9 years of soccer experience and roughly two hours of team practice six days a week, and twelve novices with no competitive soccer experience, though some had backgrounds in sports such as baseball, tennis, track and field, or sumo. Participants watched bird’s-eye video clips of real Kanto University Soccer League matches, recorded from an elevated view showing all 22 players on the field. Each clip ended at the moment the passer planted their standing foot immediately before making a pass, and viewers had to anticipate which teammate would receive the ball.

The key technical innovation was that participants were not passive observers. Each held a stylus on a pen tablet whose position was mapped in real time to a cursor on a monitor, with cursor positions recorded at 30 Hz. They were free to move the cursor throughout the 10-second clips and were instructed to predict the pass receiver. After the video stopped, the final frame appeared as a still image for two seconds, during which participants moved the cursor to their predicted receiver. Because the hand begins to move before a decision is finalized, the trajectory up to the stopping point offered a continuous window onto preparation.

The researchers defined two complementary measures relative to response onset — the moment the video stopped and a response was required. The first was spatial: the Euclidean distance, in pixels, between the cursor and the eventual correct receiver at response onset, indicating where the participant was poised. The second was kinematic: the slope of cursor velocity fitted by least-squares linear regression over the 1.5 seconds immediately before response onset, indicating how much cursor velocity increased as the moment of commitment approached. Sensitivity analyses varying the window from 1.0 to 2.0 seconds confirmed the results did not depend on this choice.

The expertise differences were clear on all fronts. Skilled players anticipated more accurately than novices — 71.64 percent correct versus 43.13 percent, a large effect of d = 2.33 with strong Bayesian support. They were also poised closer to the correct receiver at response onset (315.58 versus 347.97 pixels on average), and they showed a much larger velocity increase before responding (75.69 versus 11.75 pixels per second squared). Control analyses argued that the velocity difference was not a by-product of how far participants had to move: peak velocity after the video stopped and movement amplitude did not differ significantly between groups, total trajectory length actually tended to be greater in novices, and within the skilled group larger movement amplitudes were associated with smaller, not larger, velocity increases.

The exploratory analyses across repeated viewings produced perhaps the most intriguing pattern. Participants watched the same 19 analyzed clips three times without being told the videos would repeat. The positional difference between groups emerged only across repeated presentations and was largest in the final block, where it reached 59.4 pixels — driven by skilled players becoming progressively poised closer to the correct receiver while the poise of novices stayed roughly constant. The velocity difference, by contrast, was already present in the first presentation and showed no significant change across blocks. This dissociation suggests the two aspects of preparation are at least partly separable, though the authors caution the design cannot determine whether they are complementary facets of a single preparatory state or two distinct processes.

A supplementary confidence rating task reinforced the interpretation. Presented with the pre-pass still images and no time limit, participants rated the likelihood that each teammate would receive the ball. Skilled players assigned higher confidence to the correct receiver than novices did, mirroring the direction of the movement-based differences. The authors suggest the larger pre-response acceleration in skilled players may reflect greater confidence in, or earlier commitment to, the anticipated response — consistent with basic research showing that movement vigor scales with the subjective value of the chosen option — but they note that confidence was not measured trial by trial during the anticipation task itself, so this interpretation remains to be tested.

The authors are careful to frame the study as a proof of concept. With twelve participants per group contrasting two clearly separated skill levels, sensitivity to smaller effects was limited, and replication in larger, more diverse samples spanning a continuous range of skill is needed. The number of stimulus clips was small, and the videos were not corrected for perspective, so a given pixel distance does not correspond to the same field distance across the image — although, because both groups viewed identical frames, any distortion affects group and block comparisons equally. The block-wise analyses were exploratory, and the design cannot separate whether skilled players extracted clip-specific structure more efficiently or memorized their own previous selections, since the correct receiver was never shown to participants.

Even with those caveats, the method opens a distinctive window on sports anticipation. Because the positional and kinematic measures revealed group differences for which no counterpart appeared in the subsequent response — peak velocity after the stop did not differ between groups, and accuracy improved similarly in both groups across blocks while positional poise changed only in skilled players — continuous cursor tracking captures something discrete end-point responses cannot. Future extensions could analyze trajectory shape to track how competition among candidate receivers is resolved, or examine changes of mind and hesitation during anticipation, for example when an observer responds to a deceptive action. The study provides a behavioral basis for measuring the decision-making processes that underlie anticipation in sport, under a single continuous measure.

Subject of Research: Continuous cursor tracking as a measure of pre-response preparation in soccer anticipation

Article Title: Continuous cursor tracking as a process-oriented measure of anticipation in soccer: expertise differences in spatial and temporal preparation

Article References: Nakamura, O., Onagawa, R., Kudo, K., & Hirose, N. (2026). Continuous cursor tracking as a process-oriented measure of anticipation in soccer: expertise differences in spatial and temporal preparation. Frontiers in Sports and Active Living, 8, 1991767. https://doi.org/10.3389/fspor.2026.1991767

Image Credits: AI Generated

DOI: 10.3389/fspor.2026.1991767

Keywords: soccer, anticipation, cursor tracking, perceptual-cognitive expertise, decision-making, motor preparation, sports psychology, velocity, preparatory state, proof of concept, Continuous, cursor