Experimental protocol. Actions within SSGs were captured using two cameras: a primary camera on a DJI® Mini 2 SE drone positioned diagonally above the playing area to capture the entire field, and a secondary Canon® EOS REBEL T6I camera placed on the ground opposite the drone for supplementary analysis. SSGs were incorporated into regular training sessions, forming an integral part of the training regimen and preceding other activities scheduled for the day. Upon players' arrival, they were equipped with GPS devices and assigned numbered vests for team identification and player recognition. Subsequently, a standardized 10-minute warm-up led by one of the evaluators was conducted to prepare players for the upcoming activities.
Two game formats, GK + 4v4 + GK and GK + 5v5 + GK, were played twice by each player, once on small fields and once on large fields, resulting in four games per individual (SSG4v4-Small, SSG4v4-Large, SSG5v5-Small, and SSG5v5-Large). The dimensions of the fields were determined based on the relative area per player, with only outfield players considered in the calculation of the playing area. In previous studies, small fields were designed to provide approximately 64.80m2/player, a dimension associated with the emergence of creative actions in previous studies8. Large formats adhered to the area/player ratio suggested in previous research on SSGs, adopting a value of 272.80m2/player12. The length-to-width ratio for all setups was maintained at approximately 1.36:1. The dimensions of the penalty area were calculated proportionally, based on the standard dimensions of the field (105.00m x 68.00m) and the standard dimensions of the FIFA World Cup penalty area (16.50m x 40.00m).
Furthermore, tests were conducted on natural grass fields under rain-free weather conditions to mitigate environmental discrepancies across the proposed configurations. Throughout the data collection period, minimum and maximum temperatures ranged between 11–18 and 19–26 degrees Celsius, respectively, with relative humidity varying between 72–90% and average wind speeds between 3–5 m/s.
Participants. The study comprised 36 young male soccer players selected from the youth divisions of a well-established professional soccer club in Florianópolis, Brazil. These players were divided into two distinct age categories: U-10 and U-11. The U-10 category (n = 19; mean age 10.28 ± 0.58 years) consisted of players occupying various positions on the field, including defenders (n = 6), midfielders (n = 7), and forwards (n = 6), with a higher representation of right-footed players compared to left-footed (n = 15 vs. n = 4, respectively). Conversely, the U-11 category (n = 17; mean age 11.40 ± 0.27 years) comprised defenders (n = 4), midfielders (n = 9), and forwards (n = 4), with a similar prevalence of right-footed players over left-footed (n = 15 vs. n = 2, respectively). Goalkeepers were initially considered for inclusion but were ultimately excluded from the analyses due to their positional constraints, limited involvement in offensive play, and distinct spatial movement patterns to prevent potential misinterpretations. All participants maintained regular attendance at training sessions and conducted three times a week and actively participated in state and national competitions organized by the Club.
Inclusion criteria were as follows: (i) absence of injuries throughout the data collection period; (ii) attendance of at least one training session per week throughout the data collection period; (iii) participation in at least one session of all structured SSGs; (iv) provision of signed Informed Consent Form and Informed Assent Form by the young participants and their legal guardians, respectively. Participants were also informed of their right to withdraw from the study at any time. The human research investigation adhered to all national regulations and institutional policies, followed the principles outlined in the Helsinki Declaration, and received approval from the ethics committee of the Federal University of Santa Catarina, Brazil (approval number: Proc. 11.501).
Original actions in SSGs. Players’ originality was assessed using notational analysis via the Creative Behavior Assessment in Team Sports (CBATS), which was processed using computerized notation software, Lince Plus (version 2.1.0), and organized into a pre-prepared spreadsheet (Excel for Windows®). The tool was designed to assess players' creativity during the offensive phase of the game based on four components: (i) attempts, (ii) fluency, (iii) versatility, and (iv) originality32.
The distinctions among these components are based on their success (i.e., successful actions, fluency, versatility, and originality were considered, while unsuccessful actions were regarded as attempts) and whether they were standardized actions. Standardized actions include common passing, dribbling, or shooting, which occur when: (a) a pass is executed with the inside of the foot; (b) dribble to the right or left, touching the ball only to one side with the inside or outside of the foot; and (c) shooting with the inside of the foot. Thus, non-standardized actions encompass different executions from those mentioned above. Therefore, successful passing, dribbling, and shooting actions are classified as fluency (if standardized) or versatile (if non-standardized). Conversely, unsuccessful actions are tallied as attempts (if non-standardized) or simply considered execution failures when errors occur during a standard execution. Originality is a variable that encompasses the rarity of observations, counted only in actions performed by a maximum of 5% of players or representing up to 5% of the total actions32. Of the four indicators in the observational matrix (attempts, fluency, versatility, and originality), only original actions were considered for analysis. Table 1 presents the description of the original actions that occurred in the game with their descriptions and frequencies (absolute and relative).
Table 1
Description and analysis of original actions.
Pass Actions | Definitions | Absolute Frequency (n) | Relative Frequency (%) |
---|
Back Pass | Touching the ball to teammate in opposite side to which the body adjustment is oriented. | 3 | 1.44 |
Dropped Pass | Type of pass taken while a player is falling or off-balance. | 3 | 1.44 |
Chest Pass | The player in possession sends the ball to the teammate using the chest. | 1 | 0.48 |
Heel Pass | The player in possession sends the ball to the teammate using the heel of the foot. | 5 | 2.39 |
Outside of the foot (“trivela”) | The player uses the outside part of the foot, exerting an effect on the ball so that it takes a curved trajectory toward the teammate. | 6 | 2.87 |
Pass using knee | The player in possession sends the ball to the teammate using the knee. | 2 | 0.96 |
Rabona pass | Pass with the heel where there is a crossing between the leg of the motor action and the supporting leg. | 1 | 0.48 |
Touch and pass | The player takes an inside touch on the ball, inducing the opponent to one side before passing to the other. | 5 | 2.39 |
Dribble Actions | Definitions | Absolute Frequency (n) | Relative Frequency (%) |
“L” dribble | Quick change of direction with a dribbling movement resembling the letter "L". | 2 | 1.54 |
Dummy | One side of the opponent touches the ball, and the attacker goes around the other, catching the ball at the next point. | 5 | 3.85 |
Pinball dribble | Quick dribbling using touch with the inside of one foot and another touch with the inside of the other foot with agility. | 2 | 1.54 |
Chip | Lifts the ball into the air with a short, controlled motion, aiming to cover the legs of the approaching defender | 6 | 4.62 |
Roulette (Zidane turn) | Stop the ball with the sole, and go around an opponent player using the other sole. | 4 | 3.08 |
Scissor | The player quickly and alternately swings each leg over the ball to deceive an opponent and change direction. | 3 | 2.31 |
U-turn | A quick U-shaped turn to change direction and evade a defender. | 3 | 2.31 |
Inside-outside | The inside of one foot is used to make a short touch on the ball in one direction, and then quickly, the outside of the same foot is used to push the ball in the opposite direction. | 2 | 1.54 |
Lift above | Lift the ball above the opponent. | 6 | 4.62 |
Lift above outside of the foot | Lift the ball above the opponent using the outside of the foot. | 2 | 1.54 |
Outside cut | The player uses the outside of their foot to make a sharp and quick cut or change direction while controlling the ball. | 2 | 1.54 |
Nutmeg (“Caneta”) | The player passes the ball through an opponent's legs and retrieves it on the other side. | 3 | 2.31 |
Shot Actions | Definitions | Absolute Frequency (n) | Relative Frequency (%) |
Chip | Lifts the ball into the air with a short, controlled motion, aiming to clear an oncoming goalkeeper. | 1 | 0.32 |
Bicycle kick | The player quickly flipped their body backward in mid-air and made contact with the ball with both feet while upside down. | 1 | 0.32 |
Falling shot | Type of shot taken while a player is falling or off-balance. | 1 | 0.32 |
Heel kick | The player strikes the ball with the heel of their foot | 1 | 0.32 |
Volley | The technique of kicking the ball while it is in the air, usually when the ball is coming from above without touching the ground | 1 | 0.32 |
Note: Values considered for originality: < 5% of the total versatility shares for each share class. |
Reference versatility values: Pass (n = 209; limit considered: up to 10 actions); Dribble (n = 130; limit considered: up to 6 actions); Shot (n = 32; limit considered: 1 action.
Team Defensive dispersion. Players were tracked during SSGs using WIMU PROTM RealTrack systems from Almeria, Spain, operating at a frequency of 10Hz. Defensive dispersion was verified based on the original actions (n = 70 actions), these being passes (n = 26) and dribbling (n = 44) actions, extracted from the notational analysis, capturing two moments: those that precede (pre) and succeed (post) the exact moment of the original action. The pre-moment is characterized by the time interval that includes the pass preceding the original action or the moment of ball possession recovery (if the action was not preceded by a pass but by regaining possession) up to the moment when the original action occurs. The post-moment begins immediately after the original action and ends with one of three possible outcomes: (i) a shot on goal – when the sequence following the original action results in an attempt on target; (ii) a change in ball possession – when possession is lost following the original action; or (iii) circulation/maintenance of possession – when neither of the previous outcomes is observed within a 10-second interval, indicating that the attacking team merely retained and circulated the ball.
Based on the duration of the pre- and post-original actions recorded in the videos, synchronization with the official local time was performed to extract positional data. Positional coordinate data (latitude and longitude) were exported for further analysis using dedicated routines in MATLAB R2015b (The MathWorks, Inc., Natick, MA, USA). The following variables were analyzed, following literature recommendations33,34: (i) Stretch Index, calculated as the average distance of a team’s players and the Geometrical Center, giving a notion of the compactness of the team (Fig. 1); (ii) Surface Area, calculated as the area of the polygon defined as the convex hull with the least number of vertices that can encompass all of the team players, and where the potential vertices are defined as the positions of the players (Fig. 2); (iii) Team Length, which represents the maximum length of a team considering the minimum and maximum position of a player in the longitudinal (goal-to-goal) direction (Fig. 2); (iv) Team Width, determined by the distance between the rightmost and leftmost players in width direction (side-to-side) (Fig. 2); (v) Individual Player Area (Rectangular), calculated considering the rectangular area formed by the players of the defending team and determined by dividing the playing area by the number of players (Fig. 3).
Statistical analysis. Of the total 1454 actions recorded in the games (considering attempts, fluency, versatility, and originality), 367 records (~ 25%) were analyzed by a second evaluator with experience in observational methodology (ICC = 0.829 [IC 95% = 0.755–0.882]; F(98,98) = 10.845; p < 0.001) and reanalyzed by the primary evaluator after 4 weeks from the end of the initial measurement (ICC = 0.857 [IC 95% = 0.794–0.901]; F(98,98) = 12.831; p < 0.001) to ensure the reliability of the 70 original actions obtained. The intraclass correlation coefficient (ICC) indicated good inter- and intra-rater reliability, according to the following classifications: unacceptable (< 0.70); fair (0.70–0.79); good (0.80–0.89); excellent (≥ 0.90)35.
The normality of data was obtained through Kolmogorov-Smirnov test. To verify the effects of the original actions performed by the team with ball possession on the player’s defensive dispersion, a General Linear Model with Repeated Measures was applied after validating the data sphericity through the Mauchly test. This analysis considered pre- and post-comparison of the defensive player’s behavior dispersion within the different original actions (passing and dribbling) in different game formats (SSG4v4-Small, SSG4v4-Large, SSG5v5-Small, and SSG5v5-Large). All analyses were conducted using SPSS v.29 software, with a significance level set at 5% (p < 0.05).