Effects of contextual and tactical dimensions on the creation of goal-scoring opportunities in U12 and U13 elite Spanish soccer teams

Abstract

The aim of this paper was to describe the interactive effects of contextual and tactical dimensions on the creation of goal-scoring opportunities (GSO) in youth elite Spanish soccer, according to the type of possession. A total of 1730 team possessions from 24 qualifying matches of the official Spanish LaLiga Promises U12 and U13 soccer-7 tournaments were analyzed. For the analysis, team possessions were grouped into recoveries and restarts, according to the type of possession start. The study evaluated ten contextual and tactical dimensions by observational methodology. For recoveries, the multilevel binary regression models revealed significant effects of dimensions such as match status, initial field zone, initial penetration, type of attack, number of dribbles and number of passes on the creation of GSO. However, for restarts, only dimensions such as initial penetration, number of dribbles and possession width presented significant effects on the creation of GSO. In conclusion, the effects of different contextual and tactical dimensions in youth soccer teams seem to depend on whether the start of the team possession is a transition or a restart, regardless of the age category. These findings are useful for soccer coaches and analysts, which can help them understand the key dimensions in the design of training sessions in youth soccer teams.

Keywords

Association football performance analysis youth sport

Introduction

Research on match tactical analysis in professional soccer has exponentially grown in the last decade.¹ This evolution is contributing to provide a theoretical understanding of collective and individual human behaviors as they emerge under varying constraints within competitive performance contexts.² Within this progress, the tactical evaluation of how teams achieve offensive performance has been an important field of study in the last years.^3–5

In terms of attacking success in soccer, a very low percentage of team possessions in professional soccer led to score a goal.⁶ Thus, the sole analysis of goals may not truly represent the underlying tactical strategies of a team, i.e., those that are concerned with the actual development of goal-scoring opportunities (GSO).⁷ In this sense, the team possessions that achieve GSO represent successful attacks where the offensive team could penetrate over the defensive organization towards optimal field zones to shoot at goal.⁸ For instance, recent studies in professional soccer have shown that playing tactics such as gaining the ball in advanced zones of the field, performing initial penetrative actions and progressing by counterattacks or fast attacks increased the odds of creating GSO in different competitions.^9–11 These findings highlight the importance of performing quick and vertical collective sequences in ball transition moments to achieve higher effectiveness during the offensive phase.

However, despite the relevance of youth soccer for the learning and development of players, the research on match analysis in youth teams is very scarce, especially in prepubescent players. In this regard, existing studies have observed that very young players (U8, U10) display vertical, offensive and individualistic tactical behavior,¹² while U12 players are able to play a more combinative play,¹³ as well as to cover with more efficiency the available field space.¹⁴

Apart from the mentioned studies, there is very limited research on evaluating competitive tactics in key developmental categories such as U10, U12 or U14 players.¹⁵ In this regard, existing literature in youth soccer has been mainly focused on evaluating modified versions of soccer such as small-sided games^16,17 or analyzing the competitive physical demands.^18,19 Thus, it seems very interesting to evaluate the playing tactics of youth soccer to provide a scientific understanding of the factors that contribute to offensive and defensive success, which can help coaches to design the most accurate training sessions in a crucial period for learning the key tactical principles in soccer. In this sense, one of the barriers that can limit the production of studies in youth soccer, in comparison with professional soccer, is the lack of data or technological systems integrated into the day-to-day activities of the teams. In this scenario, the design of specific observational tools may be a very useful technique to analyze technical–tactical actions.²⁰ In fact, systematic observation is a suitable methodology for analyzing tactical behaviors in sport,²¹ because it permits the inclusion of categorical data from the qualitative evaluation of different dimensions of the match performance and may improve the ability to describe soccer match play actions.²²

Therefore, the first aim of this paper was to describe the collective playing tactics implemented by elite youth Spanish U12 and U13 soccer teams. The second aim of this study was to determine the interactive effects of contextual and tactical dimensions on the creation of GSO, according to the type of possession. It is hypothesized that tactical dimensions such as starting in advanced zones of the field, penetrating during the initial action and progressing by counterattacks have higher odds of achieving GSO during recoveries. For restarts, the hypothesis is that team sequences with a greater number of passes and more possession width present higher odds of achieving GSO.

Methods

Sample

A total of 24 knockout matches from the U12 (n = 12) and U13 (n = 12) soccer-7 tournament organized by the Spanish LaLiga in 2021 were evaluated (Table 1).

Table 1.

Matches, teams and scores of the U12 and U13 soccer-7 tournaments evaluated.

	U12 tournament	U13 tournament
Round of 16	Real Betis 3–0 Real Sociedad	Real Madrid CF 0–2 Sevilla CF
	Atlético de Madrid 2- 1 Granada CF	Granada CF 4–0 Deportivo Alavés
	Sevilla CF 1–0 Villarreal CF	Atlético de Madrid 1–0 Getafe Cf
	RC Celta 3–0 Levante UD	FC Barcelona 2–0 Real Sociedad
	FC Barcelona 2–1 SD Huesca	RCD Espanyol 2–0 Real Sociedad
	Valencia CF 3 – CA Osasuna	Valencia CF 2–1 Cádiz CF
	RCD Espanyol – 1–0 Cádiz CF	Real Betis 2–0 Levante UD
	Real Madrid CF 3–0 Real Valladolid CF	Villareal CF 4–0 RC Celta
Quarter finals	Real Betis 2–0 Valencia CF	Sevilla FC – 2–0 Valencia CF
	Atlético de Madrid 0–0 FC Barcelona	RCD Espanyol 0–0 Granada
	Sevilla FC 0-0 Real Madrid CF	Atlético de Madrid 1–1 Villarreal
	RC Celta 1–0 RCD Espanyol	FC Barcelona 0–0 – Real Betis CF

The duration of the matches was 24 min with 2 halves of 12 min each. The videotapes of the matches were obtained from the live TV broadcast. La Liga Promises is the most popular U12 and U13 soccer-7 tournament in Spain and one of the most popular in Europe. This competition has been held for 30 years and since its first editions, dozens of players that afterwards become top professional players participated when they were kids. The format of the tournament includes the 20 clubs that at the specific year take part in the national first division, so that this pool of players offers a representative selection of the elite U12 and U13 soccer players in Spain.

The unit of analysis was the ‘team possession’ according to the definition of Pollard and Reep.²³ The sample comprised 1730 team possessions that were grouped into recoveries (n = 841; 48.6%), restarts (n = 715; 41.3%) and set plays (n = 174; 10.1%) according to the type of possession start⁸ (Table 2). The key inclusion criterion for the analysis was that team possession had to present a tactical build-up and development to create GSO by means of recovering the ball in open play or restarting the team possession (throw in, goal kick, etc.). In these two scenarios, the organization of both teams is not prepared to finish or to defend a possible shot or a cross to the box since the start of the team possession. According to this criterion, set plays (n = 174) were excluded from the study.

Table 2.

Definition and categories of the dimension ‘possession type’, based on the REOFUT observational tool.⁸

Possession type: way to start a team possession according to if the ball is in play or out of play. Three categories were considered:

a) Recovery: when a player gains the possession of the ball by any means other than from a player of the same team with the ball in play.

b) Re-starts: the re-start takes place in any part of the field, (2) the tactical situation of the attacking team is not prepared to try to shoot at goal in one or two passes (player positions do not change) and (3) the attacking team try to pass the ball and build up a ball possession. (Goal kicks, free kicks, kick off, throw in).

c) Set-plays: (1) the restart takes place in the opponents’ half, (2) the tactical situation of the attacking team is prepared to try to shot at goal (both teams group players into or just in front of the box, as well as player positions change because some of the defenders move forward to try to shot at goal), and (3) the attacking team try to cross the ball into the box or shot at goal in one or two passes. (All corner kicks, all penalty kicks and those free kicks with the above characteristics are considered in this category).

Variables

The study used the REOFUT theoretical framework based on observational methodology and multidimensional analysis.⁸ This tool describes how to analyze multiple tactical and technical dimensions related to the start, development and the end of teams’ possessions, as well as their association with achieving offensive performance. This tool combines a field format with systems of categories²⁴ and it has been used in multiple investigations in professional soccer.^4,11,25 The design and protocol of this study was based on previous investigations that explored the effects of several tactical dimensions related to the start and development of team possessions on the offensive performance in professional soccer competitions.^4,9,26,27

Particularly, this study focuses on analyzing the interactive effect of seven independent tactical dimensions (Table 3) related to the possession start (initial zone, initial action and initial defensive pressure) and development (type of attack, possession width, dribbles per possession and passes per possession) on the dependent dimension (offensive performance). This analysis also included the interactive effects of three independent contextual dimensions (age: U12 vs. U13; match status: losing vs. drawing vs. winning; and match half: first half vs. second half).

Table 3.

Operational definitions of the dimensions and categories analyzed in this study.

1. Initial zone: zone on the field of play where the possession starts (Figure 1). Four categories were considered:

a) Defensive sector, b) Pre-defensive sector, c) Pre-offensive sector, d) Offensive sector.

2. Initial action: degree of offensive directness in the first three seconds of the team possession. Two categories were considered:

a) Non-penetrative action: any technical action towards any direction that does not past opponent player (s) performed during the first three seconds of the ball possession.

b) Penetrative action: passes or dribbles towards the opponent‘s goal past opponent player (s) performed during the first three seconds of the ball possession.

3. Defensive pressure: distance between the player with the ball (first attackers) and an immediate pressing opponent player(s) (first defender(s)) during the first three seconds of the ball possession^22,26. Two categories were considered.

a) Pressure: one or several opponent players press the attackers within the first 3 s of the possession (the pressing defender(s) are always located within 1.5 meters from the first attackers).

b) No pressure: there is not any player that pressures the attackers during the first 3 s of the possession

4. Type of attack: degree of offensive directness during the offensive process. Four categories were considered:

a) Counterattack: a) the possession starts by winning the ball in play, b) the progression towards the goal attempts to utilize a degree of imbalance right from start to the end with high tempo²², c) the circulation of the ball takes place more in depth than in width, as well as the intention of the team is to exploit the space left by the opponent when they were attacking and d) the opposing team does not have the opportunity to minimize surprise, reorganize his system and be prepared defensively.

b) Combinative attack: a) the possession starts by winning the ball in play or restarting the game, b) the progression towards the goal has a high percentage of non-penetrative (any pass towards any direction that does not past opponent player (s)) and short passes, c) the circulation of the ball takes place more in width than in depth⁹, d) the intention of the team is to disorder the opponent using high number of passes and relatively slow tempo (evaluated qualitatively) and e) the opposing team has the opportunity to minimize surprise, reorganize his system and be prepared defensively

c) Fast attack: a) the possession starts by winning the ball in play or restarting the game, b) the progression towards the goal has high percentage of penetrative passes (passes towards the opponent‘s goal past opponent player (s)) and short passes, c) the circulation of the ball takes place in width and depth⁹ but the intention of the team is to disorder the opponent with a reduced number of passes and high tempo (evaluated qualitatively) and d) the opposing team has the opportunity to minimize surprise, reorganize his system and be prepared defensively.

d) Direct attack: a) the possession starts by winning the ball in play or restarting the game, b) the progression towards the goal is based on one long pass from the defensive players to the forward players (evaluated qualitatively), c) the circulation of the ball takes place more in depth than in width and the intention of the team is to take the ball directly near the goal area to have opportunities of finishing by using reduced number or passes and high tempo and d) the opposing team has the opportunity to minimize surprise, reorganize his system and be prepared defensively.

5. Possession width: use of the four longitudinal lanes of the field space during the team possession. Four categories were considered:

a) One lane: during the possession, the ball moves through one of the four longitudinal lanes.

b) Two lanes: during the possession, the ball moves through two of the four longitudinal lanes.

c) Three lanes: during the possession, the ball moves through three of the four longitudinal lanes.

d) Four lanes: during the possession, the ball moves through four of the four longitudinal lanes.

6. Dribbles per possession: number of dribbles performed by the offensive team during the possession.

Two categories were considered: a) no dribbles, b) 1+ dribbles.

7. Passes per possession: number of passes performed by the offensive team during the possession.

Three categories were considered: a) 0–3 passes, b) 4–6 passes; c) +7 passes.

8. Offensive performance: degree of offensive success of the possession, based on the degree of penetration over the opposing team. Two categories were considered:

a) Goal-scoring opportunity: the offensive team achieves to shoot at goal during the team possession.

b) No goal-scoring opportunity: the offensive team does not achieve to shoot at goal during the team possession.

Match analysis

The study is based on the principles of observational methodology^21,28 and its design is nomothetic (several games) and multidimensional (the dimensions correspond with the criteria of the REOFUT observation tool). For the analysis, the LINCE PLUS software was used to register, code and save the data.²⁹

Two observers evaluated the team possessions to maximize the analysis efficiency and inter-observer reliability in line with previous studies that analyzed team possessions in soccer.^30,31 The two observers had playing and coaching experience in youth soccer and were trained in the use of the REOFUT tool for four weeks by the principal researcher of this study, who holds wide experience in performance analysis in soccer and is one of the lead authors of the used tool. This training included theoretical lessons and practical instruction, following the guidelines and suggestions of specialized studies about observational methodology.^32,33 When the training was completed, each observer analyzed 12 matches separately. The analysis was made post-event, and each team possession was analyzed as many times as necessary for the observers.

To check the reliability of the data, the two observers in addition to the principal researcher (used as reference) analyzed 92 team possessions correspondent to a match for the analysis of inter-observer reliability. Subsequently, the principal researcher re-observed the game three weeks later for the intra-observer concordance. Kappa correlation coefficients (κ) were calculated for inter-observer and intra-observer reliability. In this sense, this analysis showed an appropriate level of reliability according to Altman criteria³⁴ (inter-observer kappa coefficient = 0.87–0.95; intra-observer kappa coefficient = 0.81–0.88).

Statistical analysis

All the analyses were performed using SPSS software (IBM SPSS, Version 27.0). An analysis of frequencies was carried out to describe the characteristics of the sample and the occurrence of each tactical dimension according to the offensive performance both for restarts and recoveries.

Due to the hierarchal structure of team possessions in soccer (each team has its specific tactical style and different ways to create GSOs), multilevel modeling³⁵ was carried out to cluster the team possessions (Level 2) within teams (Level 1) as previous studies performed⁴ (Figure 1). A mixed model was created to analyze the effect of the contextual and tactical independent dimensions (fixed effects), on the creation of offensive performance, considering the team identity (random effects). Thus, the ‘team effects’ represent unobserved team characteristics that influence the team possession outcome.

Figure 1.

Study design with a hierarchical data structure, in which team possessions are nested in teams (Adapted from González-Rodenas et al.⁴).

With this organization of the data, binary logistic regression models were constructed to predict the creation of shooting opportunities (0 = no shooting opportunity, 1 = shooting opportunity).

For the presentation of the results, unadjusted models (univariate analysis) were carried out to determine the association of each independent dimension with the dependent dimension. Secondly, adjusted logistic multilevel models (multivariate analysis) were constructed with all independent dimensions to check the interactive and interdependent effect of both contextual and tactical dimensions.

Finally, graphic charts with the predicted means were displayed for the variables that presented significant effects on the dependent variables. Pairwise comparisons of the estimated means (EMs) were performed through Fisher’s least significant difference test. The significance level was set to p < 0.050.

Results

Descriptive analysis

Table 4 shows the main characteristics of the sample. In general, recoveries produced a greater percentage of GSO than restarts (18.1% vs. 9.8%). As for contextual dimensions, the U12 teams had a greater percentage of GSO in recoveries than the U13 teams (22.3% vs. 14.4%), while winning teams had higher effectiveness to create GSO than drawing and losing teams.

Table 4.

Frequencies of the different contextual and tactical dimensions according to the creation of GSO for recoveries and restarts.

Contextual and tactical dimensions	Restarts		Recoveries
Contextual and tactical dimensions	n	GSO (%)	n	GSO (%)
Age
U12	322	9.6	377	22.3
U13	390	10.0	464	14.4
Match status
Losing	192	8.3	150	11.3
Drawing	386	9.6	477	17.8
Winning	134	12.7	214	22.9
Match half
First half	336	7.1	417	19.4
Second half	376	12.2	424	16.5
Initial zone
Defensive zoneª	300	6.3	323	10.8
Pre defensive	239	10.9	328	16.2
Pre-offensive	127	15.0	159	32.1
Offensive	46	13.0	31	38.7
Initial action
Non penetrative actionª	520	8.8	494	13.0
Penetrative action	192	12.5	347	25.1
Defensive pressure
Defensive pressure ª	458	9.4	650	18.9
Non defensive pressure	254	10.6	191	14.7
Type of attack
Combinative attackª	461	8.9	350	14.0
Fast attack	164	14.6	297	23.6
Direct attack	87	5.7	76	9.2
Counterattack	-	-	118	21.2
Possession width
One lane	134	0.7	219	15.1
Two lanes	232	11.6	295	20.3
Three lanes	138	11.6	166	16.3
Four lanes	208	12.5	161	19.3
Dribbles per possession
0	590	6.9	616	14.3
1 +	121	24.0	220	28.6
Passes per possession
0–3	405	8.9	595	17.8
4–6	192	10.9	154	22.1
7 +	115	11.3	91	12.1
Total	712	9.8	836	18.1

Regarding the type of attack, combinative attacks were the most frequent, followed by fast attacks, counterattacks and direct attacks, although the ones that achieved a greater percentage of GSO were the fast attack (14.6% in restarts and 23.6% in recoveries) and counterattacks (21.2% in recoveries).

Multilevel regression analysis

Regarding the random effects, Table 5 shows that the effect of ‘team identity’ did not present a significant variance for the creation of GSOs neither for recoveries nor restarts (p = 0.481 and 0.542, respectively).

Table 5.

Random effects of team identity on achieving goal-scoring opportunity vs. no scoring opportunity according to the possession type.

Possession type	Estimate	Standard error	Z	Sig	95% CI
Recoveries	0.102	0.145	0.750	0.481	0.006–1.642
Restarts	0.140	0.230	0.610	0.542	0.006–3.482

Table 6 shows the univariate and multivariate effects of contextual and tactical dimensions on the creation of GSO in recoveries. The multivariate model fits 83.3% of the data and shows that 6 out of 10 dimensions presented a significant effect in predicting the creation of GSO.

Table 6.

Multilevel mixed model to predict the creation of GSO vs. no GSO (reference category) in recoveries.

Tactical and contextual dimensions	No GSO vs. GSO		No GSO vs. GSO
	Univariate analysis		Multivariate analysis
	B	OR (95% CI)	B	OR (95% CI)
Age
U13
U12	0.590	1.804 (1.134–2.870)*	0.268	1.307 (0.777–2.200)
Match status
Losing
Drawing	0.486	1.626 (0.901–2.937)	0.685	1.985 (1.021–3.858)*
Winning	0.764	2.146 (1.135–4.057)*	0.994	2.703 (1.344–5.435)**
Match half
First half
Second half	−0.172	0.842 (0.589–1.203)	−0.184	0.832 (0.547–1.265)
Initial zone
Defensive zoneª
Pre defensive	0.399	1.490 (0.936–2.373)***	0.210	1.233 (0.740–2.055)
Pre-offensive	1.346	3.842 (2.348–6.286)***	1.456	4.291 (2.459–7.488)***
Offensive	1.613	5.016 (2.214–11.364)***	1.936	6.929 (2.714–17.691)***
Initial action
Non penetrative actionª
Penetrative action	0.804	2.234 (1.546–3.228)***	0.841	2.319 (1.445–3.721)**
Defensive pressure
Defensive pressure ª
Non defensive pressure	0.301	1.351 (0.855–2.133)	0.315	1.370 (0.799–2.351)
Type of attack
Combinative attackª
Fast attack	0.646	1.907 (1.907–2.881)**	0.907	2.477 (1.302–4.712)**
Direct attack	−0.422	0.656 (0.277–1.551)	0.364	1.440 (0.507–4.087)
Counterattack	0.602	0.602 (1.826–3.179)*	1.115	3.049 (1.375–6.762)**
Possession width
One lane^a
Two lanes	0.311	1.251 (0.715–2.189)	0.378	1.460 (0.835–2.552)
Three lanes	−0.019	0.981 (0.551–1.745)	0.322	1.380 (0.669–2.847)
Four lanes	0.224	1.365 (0.839–2.221)	0.714	2.041 (0.848–4.916)
Dribbles per possession
0^a
1 +	0.872	2.392 (1.633–3.504)***	0.683	1.979 (1.301–3.010)**
Passes per possession
0–3^a
4–6	0.223	1.250 (0.802–1.947)	0.924	2.519 (1.242–5.109)*
7*	−0.540	0.583 (0.293–1.158)	0.179	1.196 (0.421–3.401)
Intercept	−1.626	0.197 (0.153–0.253)***	−4.578	0.010 (0.003–0.032)

B: regression coefficient; OR: odds ratio; CI: confidence interval for odds ratio; *p ≥ 0.05; **p ≥ 0.01; ***p ≥ 0.001.

Reference category.

In this sense, contextual dimensions such as age or match status registered significant univariate effects on the creation of GSO, although only the match status effect was significant when the model was adjusted for the rest of the dimensions. Specifically, winning (OR = 2.703; p < 0.01) and drawing (OR = 1.985; p < 0.05) teams obtained higher probabilities of creating a GSO than losing teams.

For tactical dimensions, the initial zone, the initial action, the type of attack and the dribbles per possession showed univariate and multivariate effects on the creation of GSO. Particularly, the multivariate model showed that initiating in the pre-offensive (OR = 4.291; p < 0.001) and offensive (OR = 6.929; p < 0.001) sectors, performing a penetrative action (OR = 2.319; p < 0.01), progressing with a fast attack (OR = 2.477; p < 0.01) or counterattack (OR = 3.049; p < 0.01) and performing at least one dribble during the possession (OR = 1.979; p < 0.01) obtained higher probabilities to create a GSO than initiating in the defensive sector, performing a non-penetrative action, progressing with combinative attack and not performing any dribble during the team possession, respectively. Also, the number of passes per possession only had a significant effect in the multivariate analysis, showing that making between 4 and 6 passes had higher probabilities (OR = 2.519; p < 0.05) to develop GSO in comparison with making 0–3 passes, while no effect was observed for possessions that involved more than 7 passes.

Figure 2 shows the EMs with regards to the creation of GSO for each significant dimension in the multivariate analysis. It can be observed that tactical aspects such as initiating the team possession in pre-offensive and offensive zones, performing initial penetrative actions, dribbling, making between 4 and 6 passes, as well as progressing by counterattack or fast attack showed approximately 40% or higher odds of achieving a GSO. Specifically, winning teams (EM = 0.40; 95% CI: 0.26–0.55) and drawing teams (EM = 0.32; 95% CI: 0.22–0.46) presented significantly higher means than losing teams (p < 0.01 and p < 0.05, respectively). Regarding tactical dimensions, initiating in offensive (EM = 0.55; 95% CI: 0.31–0.76) and pre-offensive zones (EM = 0.43; 95% CI: 0.29–0.58) registered significantly higher EMs that initiating the team possession in pre-defensive and defensives zones (p < 0.01).

Figure 2.

Predicted means and confidence intervals related to the creation of GSO in recoveries according to different contextual and tactical dimensions after adjusting for the dimensions included in the multivariate analysis. *≤0.05; **≤0.01; ***≤0.001.

As for passing, performing between 4 and 6 passes during team possession (EM = 0.43; 95% CI: 0.28–0.59) registered higher odds of creating GSO than performing between 0 and 3 passes (p < 0.05), although no statistical differences were found in comparison with performing 7 or more passes. Additionally, performing an initial penetrative action (EM = 0.40: 95% CI: 0.27–5.41) and dribbling during the possession (EM = 0.38; 95% CI: 0.25–0.52) registered significant differences in comparison with non-performing an initial penetrative action (EM = 0.22; 95% CI: 0.14–0.33; p < 0.01) or non-dribbling during the possession (EM = 0.23; 95% CI: 0.15–0.34; p < 0.01). Finally, progressing by counterattacks (EM = 0.42; 95% CI = 0.25–0.60) and fast attacks (EM = 0.37; 95% CI: 0.25–0.51) presented differences in comparison with progressing by combinative attacks (EM = 0.19; 95% CI: 0.12–0.284; p < 0.05).

Regarding restarts, the model was able to predict 90.4% of the data. Table 7 shows that the dimensions such as match half, initial zone, type of attack, possession width and dribbles per possession had a univariate effect on the creation of GSO. However, only three dimensions such as initial action, possession width and dribbles per possession registered a significant effect when the multivariate analysis was undertaken. Specifically, performing a penetrative action increased the odds of creating a GSO (OR = 2.126; p < 0.05). Also, the fact of occupying two (OR = 4.199; p < 0.01), three (OR = 5.648; p < 0.01) and four lanes of the field (OR = 6.620; p < 0.01) during the team possession gradually increased the odds of creating a GSO in comparison with occupying only one lane. Finally, dribbling during the possession increased the probability of achieving a GSO (OR = 3.067; p < 0.001).

Table 7.

Multilevel mixed model to predict the creation of GSO vs. no GSO (reference category) in restart.

Tactical and contextual dimensions	No GSO vs. GSO		No GSO vs. GSO
	Univariate analysis		Multivariate analysis
	B	OR (95% CI)	B	OR (95% CI)
Age
U13
U12	0.039	1.040 (0.542–1.997)	−0.411	0.663 (0.377–1.165)
Match status
Losing^a
Drawing	0.045	1.046 (0.540–2.029)	0.223	1.250 (0.657–2.379)
Winning	0.411	1.509 (0.689–3.305)	0.116	1.123 (0.534–2.363)
Match half
First half^a
Second half	0.626	1.871 (1.110–3.154)*	0.502	1.652 (0.972–2.807)
Initial zone
Defensive zoneª
Pre defensive	0.568	1.765 (0.945–3.295)	0.400	1.493 (0.800–2.785)
Pre-offensive	0.935	2.547 (1.287–5.042)**	0.636	1.890 (0.936–3.815)
Offensive	0.707	2.029 (0.750–5.487)	0.767	2.153 (0.757–6.127)
Initial action
Non penetrative actionª
Penetrative action	0.428	1.534 (0.890–2.644)	0.754	2.126 (1052–4.297)*
Defensive pressure
Defensive pressure ª
Non defensive pressure	−0.080	0.923 (0.544–1.567)	0.122	1.130 (0.625–2.044)
Type of attack
Combinative attackª
Fast attack	0.608	0.671 (0.250–1.797)	0.462	1.587 (0.741–3.399)
Direct attack	−0.400	1.836 (1.050–3.212)*	0.231	1.260 (0.432–3.673)
Possession width
One lane^a
Two lanes	2.919	18.521 (2.431–138.996)**	1.435	4.199 (1.469–12.006)**
Three lanes	2.943	18.976 (2.443–147–424)**	1.731	5.648 (1.664–19.168)**
Four lanes	2.911	18.383 (2.431–138.996)**	1.890	6.620 (1.859–23.578)**
Dribbles per possession
0^a
1 +	1.439	4.216 (2.460–7.226)***	1.121	3.067 (1.765–5.328)***
Passes per possession
0–3^a
4–6	0.160	1.199 (0.599–2.401)	0.098	1.103 (0.504–2.414)
7*	0.182	1.174 (0.657–2.098)	−0.041	0.960 (0.347–2.658)
Intercept	−2.259	0.104 (0.076–0.143)***	−4.864	0.008 (0.002–0.030)***

B: regression coefficient; OR: odds ratio; CI: confidence interval for odds ratio; *p ≥ 0.05; **p ≥ 0.01; ***p ≥ 0.001.

Reference category.

The EM and confidence intervals for the creation of GSO from the dimensions with significant effects are shown in Figure 3. It can be observed that tactical aspects such as performing an initial penetrative action (EM = 0.20; 95% CI: 0.12–0.33), dribbling (EM = 0.23; 95% CI: 0.14–0.37) during the possession registered statistical differences in comparison with non-performing an initial penetrative action (EM = 0.10; 95% CI = 0.6–0.19) (p < 0.05) and non-dribbling (EM = 0.9; 95% CI: 0.13–0.37) (p < 0.01) during the team possession. Finally, the fact of occupying two (EM = 0.17; 95% CI: 0.03–0.9) three (EM = 0.22; 95% CI = 0.12–0.370) or four lanes (EM = 0.25; 95% CI: 0.15–0.39) had higher odds of achieving a GSO than occupying one lane (EM = 0.04; 95% CI: 0.01–0.13) (p < 0.01).

Figure 3.

Predicted means and confidence intervals related to the creation of GSO in restarts according to different contextual and tactical dimensions after adjusting for the dimensions included in the multivariate analysis. *≤0.05; **≤0.01; ***≤0.001.

Discussion

The aim of this paper was to describe the interactive effects of contextual and tactical dimensions on the creation of GSO in youth elite Spanish soccer, according to the type of possession. Our study found significant interactive effects of contextual and tactical dimensions on the creation of GSO, what provides new insights about key performance indicators in youth soccer teams. Besides, our findings indicated that the effects of specific tactical dimensions were different depending on the type of possession (restart vs. recovery).

Regarding the start of team sequences, initiating the possession in offensive zones and performing initial penetration was more effective to create GSO in recoveries. Existing literature in professional soccer revealed that performing penetrative actions in the first seconds after gaining the ball increased the probability to produce GSO.^11,36 In this sense, ball transition moments present the opportunity for the offensive team to exploit the space left by the opposing team while attacking, so offensive speed and outnumbering opponents are key objectives for success when recovering the ball.³⁷ Our study also found this positive effect in restarts, showing that the capacity of teams to break lines as early as possible during the possession may increase the production of GSO, even if there is not a ball transition.

As for the development of the team possessions, it is interesting to highlight that the type of attack as well as the number of passes registered a significant effect on the production of GSO for recoveries. In fact, counterattacks and fast attacks registered higher odds to create GSO than combinative attacks. These findings show the key importance of fast-attacking sequences to achieve offensive performance in youth teams during transition moments. Previous studies in professional soccer had observed the higher effectiveness of fast attacks and counterattacks.^4,9,11 Also, it is interesting to highlight that those possessions that included four to six passes were more effective than the shorter (0–3) and longer ones (+7). This seems to be different from professional soccer, where it seems that longer passing sequences are more effective than shorter ones according to multiple studies.^11,26,38,39 This result may be due to the different spatial constraints of soccer-7, where the goals are closer to each other and the relative space per player is smaller than in soccer-11. These characteristics could create a tactical context where youth players could get closer to the opposing goal after a few passes, while in soccer-11 the progression towards the goal may require more elaboration and connectivity between players.

For restarts, the key tactical dimension that showed a significant effect on the offensive performance was the possession width, so that the wider the team passing sequences were, the higher the odds of creating GSO. This finding can be expected considering that restarts, differently from recoveries, were initiated against an organized defensive team. In this scenario, it would be crucial for the offensive players to position themselves as wide as possible to increase their team´s effective playing space and move the ball from areas with more defensive density to others with more free space to progress towards the goal.⁴⁰

One of the key findings of our study was the relevance of the number of dribbles during the team possessions. For both recoveries and restarts, performing a higher number of dribbles increased the odds of creating a GSO, regardless of the influence of other tactical and contextual dimensions. It is worth mentioning that according to our descriptive data; only 17% of team sequences in restarts and 26% in recoveries included at least one dribble, which reflects the low frequency of this type of action despite its great offensive contribution. This data points out how dribbling is a disrupting action that can help teams penetrate and create GSO, so that may offer a strong tactical advantage. For example, a recent study⁴¹ observed that goal-scoring success in youth soccer players was positively associated with their dribbling speed and technical dribbling. Thus, the improvement of dribbling should be a pivotal attribute in the development of young players.^42,43

As for contextual dimensions, only the match status presented a multivariate significant effect on the creation of GSO in recoveries, showing that winning and drawing teams obtained higher odds of offensive success. In this regard, previous studies in professional soccer have observed that losing teams increase the offensive production^44,45 to try to equalize the score, which could make them assume more defensive risks. The fact that the duration of the matches in this tournament was very short could have increased the urgency for losing teams to score and for winning teams to protect their goal. In this tactical situation, the recovery of the ball by the winning teams would create an ideal scenario to counterattack and create GSO, what would explain why this effect was only observed in recoveries and not in restarts.

This investigation presents interesting practical applications. In fact, this study is the first to analyze the interactive effects of playing tactics and contextual dimensions on offensive performance in U12 and U13 teams. Considering our findings, attacking in transition moments seems to require fast attacks and counterattacks to increase the odds of creating scoring chances, whereas attacking in restarts requires to move the ball from side to side to display wide passing sequences. This information is crucial for coaches and practitioners. On one hand, coaches of youth teams should create representative learning exercises for players to practice tactical situations both recovering and restarting the game. In these scenarios, youth players should make decisions about when to be penetrative and counterattack on recoveries, as well as when to be more patient and move the ball from side to side in restarts. On the other hand, dribbling seems to be a key action that contributes to increase the odds of achieving GSO in youth soccer, suggesting coaches to promote this type of actions during their training sessions and collective game model. Finally, the findings of our study can be used by youth soccer academy directors to design long-term technical and tactical periodization in youth U12 and U13 soccer players.

However, our study has several limitations. First, our methodology is based on systematic observation by coding and registering tactical behaviors during the matches. This method, despite having multiple advantages, may not capture entirely the complex and interactive nature of collective tactical actions during the game as some authors claimed.^46–48 Secondly, our study was focused exclusively on the analysis of offensive process, so that no information of defensive dimensions was obtained. Additionally, the current study has been carried out with data from elite male youth players from Spain, and the results of the investigation should not be extrapolated to other competitions, categories or to women players. Regarding these limitations, future studies should complement our findings with other methodologies (network analysis, positional data, etc.) and include the analysis of defensive behaviors to offer a more complete evaluation of the tactical performance in youth soccer. Also, it would be interesting to analyze other competitions and contexts such as women youth soccer.

In conclusion, the effects of different contextual and tactical dimensions on the creation of GSO in youth soccer teams seem to depend on whether the start of the team possession is a ball recovery or a restart, regardless of the age category. For recoveries, our findings showed that penetrating in the initial actions, starting in the opposing half, progressing by counterattacks and fast attacks, performing four to six passes and dribbling during the team possessions increased the odds of achieving GSO. For restarts, tactical dimensions such as penetrating in the initial actions, performing width passing sequences and dribbling during the team possessions increased the odds of achieving GSO. These findings are useful for soccer coaches and analysts of youth teams, which can help them to understand the decisive dimensions that influence the offensive performance in U12 and U13 soccer teams.

Footnotes

Data availability statement

The data that support the findings of this study are available upon request to the authors.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Joaquín González-Rodenas

References

Sarmento

Clemente

Araújo

, et al. What performance analysts need to know about research trends in association football (2012–2016): a systematic review. Sports Med 2018; 48: 799–836.

Sarmento

Clemente

Afonso

, et al. Match analysis in team ball sports: an umbrella review of systematic reviews and meta-analyses. Sports Med Open 2022; 8: 1–22.

Kempe

Memmert

. “Good, better, creative”: the influence of creativity on goal scoring in elite soccer. J Sports Sci 2018; 36: 2419–2423.

González-Rodenas

Aranda-Malaves

Tudela-Desantes

, et al. Playing tactics, contextual variables and offensive effectiveness in English Premier League soccer matches. A multilevel analysis. PLoS One 2020; e0226978. DOI: 10.1371/JOURNAL.PONE.0226978

Aguado-Méndez

González-Jurado

Callejas-Jerónimo

, et al. Analysis of the goal-scoring opportunities conceded in football: a study case in the Spanish La Liga. Qual Quant 2021; 55: 1477–1496.

González-Rodenas

Malavés

Desantes

, et al. Past, present and future of goal scoring analysis in professional soccer. Retos 2020; 37: 785.

James

Mellalieu

Hollely

. Analysis of strategies in soccer as a function of European and domestic competition. Int J Perform Anal Sport 2002; 2: 85–103.

Aranda

González-Ródenas

López-Bondia

, et al. ‘REOFUT‘ as an observation tool for tactical analysis on offensive performance in soccer: mixed method perspective. Front Psychol 2019; 10: 1476.

Sarmento

Figueiredo

Lago-Peñas

, et al. Influence of tactical and situational variables on offensive sequences during elite football matches. J Strength Cond Res 2018; 32: 2331–2339.

10.

Schulze

Julian

Meyer

. Exploring factors related to goal scoring opportunities in professional football. Sci Med Football 2021; 6: 181–188.

11.

González-Rodenas

Aranda-Malavés

Tudela-Desantes

, et al. Effect of match location, team ranking, match status and tactical dimensions on the offensive performance in Spanish ‘La Liga’ soccer matches. Front Psychol 2019; 10: 2089.

12.

Folgado

Lemmink

Frencken

. Length, width and centroid distance as measures of teams tactical performance in youth football. Eur J Sport Sci 2012; 14: 487–492.

13.

González-Víllora

García-López

Contreras-Jordan

. Evolución de la toma de decisiones y la habilidad técnica en fútbol. Rev Int Med Cienc Act Fis Deporte 2015; 15: 467–487.

14.

Serra-Olivares

García-López

Goncalves

. Effects of the players’ level and age group category on positional tactical behaviour during 7-and 8-a-side football youth games. Int J Perform Anal Sport 2019; 19: 236–247.

15.

Lapresa

Río del

Arana

, et al. Use of effective play-space by U12 FC Barcelona players: an observational study combining lag sequential analysis and T-pattern detection. Int J Perform Anal Sport 2018; 18: 293–309.

16.

Olthof

Frencken

Lemmink

. The older, the wider: on-field tactical behavior of elite-standard youth soccer players in small-sided games. Hum Mov Sci 2015; 41: 92–102.

17.

O’brien-Smith

Fransen

Sheehan

, et al. Collective behaviour in high and low-level youth soccer teams. Med Sci Football 2021; 6: 164–171.

18.

Sanchez

Ramirez-Campillo

Hernandez

, et al. External loads in under-12 players during soccer-7, soccer-8, and soccer-11 official matches. Int J Environ Res Public Health 2021; 18: 4581.

19.

Hernandez-Martin

Sanchez-Sanchez

Felipe

, et al. Physical demands of U10 players in a 7-a-Side soccer tournament depending on the playing position and level of opponents in consecutive matches using global. Sensors 2020; 20: 6968.

20.

González-Víllora

Serra-Olivares

Pastor-Vicedo

, et al. Review of the tactical evaluation tools for youth players, assessing the tactics in team sports: football. Springerplus 2015; 4: 1–17.

21.

Anguera

Mendo

. La metodología observacional en el ámbito del deporte. E-balonmano.com Rev Cienc Deporte 2013; 9: 135–160.

22.

Tenga

Kanstad

Ronglan

, et al. Developing a new method for team match performance analysis in professional soccer and testing its reliability. Int J Perform Anal Sport 2009; 9: 8–25.

23.

Pollard

Reep

. Measuring the effectiveness of playing strategies at soccer. J R Stat S Series D: Stat 1997; 46: 541–550.

24.

Anguera

Magnusson

Jonsson

. Instrumentos no estándar: planteamiento, desarrollo y posibilidades. Avances en Medición 2007; 5: 63–82.

25.

Mitrotasios

Gonzalez-Rodenas

Armatas

, et al. The creation of goal scoring opportunities in professional soccer. Tactical differences between Spanish La Liga, English premier league, German bundesliga and Italian serie A. Int J Perform Anal Sport 2019; 19: 452–465.

26.

Lago-Ballesteros

Lago-Peñas

Rey

. The effect of playing tactics and situational variables on achieving score-box possessions in a professional soccer team. J Sports Sci 2012; 30: 1455–1461.

27.

Tenga

Holme

Ronglan

, et al. Effect of playing tactics on achieving score-box possessions in a random series of team possessions from Norwegian professional soccer matches. J Sports Sci 2010; 28: 245–255.

28.

Anguera

Blanco Villaseñor

Mendo

, et al. Diseños observacionales: ajuste y aplicación en psicología del deporte. Rev Psicol Deporte 2011; 11: 63–76.

29.

Soto-Fernández

Camerino

Iglesias

, et al. LINCE PLUS software for systematic observational studies in sports and health. Behav Res Methods 2022; 54: 1263–1271.

30.

Amatria

Maneiro

Teresa Anguera

. Analysis of successful offensive play patterns by the Spanish soccer team. J Hum Kinet 2019; 69: 191–200.

31.

Sgrò

Aiello

Casella

, et al. The effects of match-playing aspects and situational variables on achieving score-box possessions in Euro 2012 football championship. J Hum Sport Exerc 2017; 12: 58–72.

32.

Losada

Manolov

. The process of basic training, applied training, maintaining the performance of an observer. Qual Quant 2015; 49: 339–347.

33.

Chacón-Moscoso

Anguera

Sanduvete-Chaves

, et al. Methodological quality checklist for studies based on observational methodology (MQCOM). Psicothema 2019; 31: 458–464.

34.

Altman

. Some common problems in medical research. In: Altman

(ed) Practical statistics for medical research. London: Chapman & Hall, 1991, pp. 403–409.

35.

Sommet

Morselli

. Keep calm and learn multilevel logistic modeling: a simplified three-step procedure using Stata, R, Mplus, and SPSS. Int Rev Social Psychol 2017; 30: 229–230.

36.

Hughes

Lovell

. Transition to attack in elite soccer. J Hum Sport Exerc 2019; 236–253. DOI: 10.14198/jhse.2019.141.20

37.

Hewitt

Greenham

Norton

. Game style in soccer: what is it and can we quantify it? Int J Perform Anal Sport 2017; 16: 355–372.

38.

Hughes

Franks

. Analysis of passing sequences, shots and goals in soccer. J Sport Sci 2005; 23: 509–514.

39.

González-Rodenas

Lopez-Bondia

Calabuig

, et al. The effects of playing tactics on creating scoring opportunities in random matches from US Major league soccer. Int J Perform Anal Sport 2015; 15: 851–872.

40.

Merlin

Cunha

Moura

, et al. Exploring the determinants of success in different clusters of ball possession sequences in soccer. Res Sports Med 2020; 28: 339–350.

41.

Wilson

Smith

de Souza N

, et al. Dribbling speed predicts goal-scoring success in a soccer training game. Scand J Med Sci Sports 2020; 30: 2070–2077.

42.

Ali

. Measuring soccer skill performance: a review. Scand J Med Sci Sports 2011; 21: 170–183.

43.

Huijgen

BCH

Elferink-Gemser

Post

, et al. Soccer skill development in professionals. Int J Sports Med 2009; 30: 585–591.

44.

Fernandez-Navarro

Fradua

Zubillaga

, et al. Influence of contextual variables on styles of play in soccer. Int J Perform Anal Sport 2018; 18: 423–436.

45.

Bradley

Lago-Peñas

Rey

, et al. The influence of situational variables on ball possession in the English premier league. J Sports Sci 2014; 32: 1867–1873.

46.

Glazier

. Game, set and match? Substantive issues and future directions in performance analysis. Sports Med 2010; 40: 625–634.

47.

Vilar

Araújo

Davids

, et al. The role of ecological dynamics in analysing performance in team sports. Sports Med 2012; 42: 1–10.

48.

Duarte

Araújo

Correia

, et al. Sports teams as superorganisms. Sports Med 2012; 42: 633–642.