Introduction
The global fascination with soccer (association football) is reflected in its widespread viewership and the marked expansion of performance analysis research over the past decade (Sarmento et al., 2022). Advanced automated and semi-automated video analysis and tracking systems have revolutionised the collection of vast, multidimensional information, making traditional notational methods less prominent but still crucial. While physical performance metrics continue to dominate research (Errekagorri et al., 2022; Otero-Saborido et al., 2024) providing valuable benchmarks for coaching and training (Bradley, 2024), there is an urgent need to focus on technical-tactical aspects that are pivotal in understanding match dynamics and outcomes (Barthelemy et al., 2024; Lago-Peñas et al., 2023; Konefał et al., 2019; Yan et al., 2024).
The use of match-aggregated statistics from data providers to examine emerging technical and tactical trends in soccer has gained prominence in sports science. Although whole- match data can grant a clear overview of team behaviour (Barthelemy et al., 2024; Lepschy et al., 2021), it may overlook critical, often invisible patterns, such as the mechanisms behind goal-scoring events (Almeida, 2019; Praça et al., 2024). Goals in soccer are rare and decisive episodes with a minimal probability of reoccurring in the same way (Kubayi and Larkin, 2022); yet, goal-scoring behaviours are not entirely random, suggesting the existence of patterns that drive goal creation (Anzer et al., 2021). Therefore, moving beyond the popular analysis of game-related statistics, notational analysis greatly contributes to a better understanding of the team patterns that lead to success in modern soccer (Almeida, 2019; Sarmento et al., 2022).
Some studies highlight that goal-scoring opportunities may better predict overall team performance than goals themselves (Aguado-Méndez et al., 2021; Schulze et al., 2022). Nevertheless, completed goal-scoring sequences offer essential insights into the technical-tactical and situational dynamics that culminate in goals and demonstrate the value of notational analysis for capturing these quality-driven actions. In this sense, recent literature underscores a shift towards attacking efficiency, where top teams prioritise quality over quantity in their offensive endeavours. These teams often leverage higher passing tempos and faster ball movements to effectively penetrate well-organised defences (Schulze et al., 2022; Taha and Ali, 2023; Wallace and Norton, 2014). This trend towards more dynamic and tactically astute playing styles favours effective possession over prolonged control, as groundbreakingly evidenced by Collet (2013). The emphasis on quality is further corroborated by studies linking competitive success (e.g., advancing to the FIFA World Cup knockout phase) to shooting accuracy, rather than shot volume (Degrenne and Carling, 2024; Errekagorri et al., 2022; Kubayi and Larkin, 2022).
Efficiency in the attacking phase extends beyond simply scoring goals, including how teams regain and maintain possession. Successful teams often disrupt opponents early by regaining possession higher up the pitch (Almeida et al., 2014; Santos et al., 2017), maintaining longer possessions near the opponent’s goal within constrained spaces (Iván-Baragaño et al., 2024). Evolving strategies also involve building up from the back, utilising goalkeepers and central defenders in playmaking roles (Pan et al., 2024). However, achieving positive outcomes in top-level tournaments requires not only efficiency and tight defensive coordination, but also the ability to adapt strategically and tactically to various match contexts (Aguado-Méndez et al., 2021; Yan et al., 2024). This insight underlines the imperative for soccer strategies to centre around improving the quality of team interactions in offensive play without being confined to a single approach (positional attack, fast attack or counterattack) and disregarding the competitive circumstances.
In this vein, a compelling body of evidence supports situational variables’ impact on attacking and defending performances in professional soccer (Almeida et al., 2014; Santos et al., 2017; Sarmento et al., 2018, 2022). Notwithstanding their relevance, much of the research conducted on FIFA Men’s World Cups has neglected critical variables such as team/opponent quality, match period, and match status (Kubayi and Larkin, 2022; Praça et al., 2024; Yan et al., 2024; Yi et al., 2019). This oversight discloses a notable gap in the literature, stressing the need for further studies that integrate these situational variables to better grasp their influence on elite-level soccer match performance.
Ongoing research has pointed out the central role of a team’s relative strength in differentiating among team performances at the highest levels of soccer. At the latest FIFA World Cup, better-ranked teams outperformed their lower-ranked counterparts in shooting and passing, displaying not only more attempts at the goal and shots on target, but also superior passing accuracy and possession control (Braquinho et al., 2024; de França et al., 2024; Pan et al., 2024). These disparities are often attributed to the stronger teams’ tendency to adopt a possession-based strategy, in contrast to the faster, counterattack-focused methods preferred by lower-ranked teams (Castellano et al., 2013; de França et al., 2024; Pan et al., 2024). Defensively, higher-ranked teams applied more aggressive and coordinated pressure in advanced field positions than lower-ranked ones (de França et al., 2024; Iván-Baragaño et al., 2024; Praça et al., 2024).
Despite the extensive data on the timing of goals, with a noteworthy increase during the second half and especially in the final period (76 min–full-time) (Degrenne and Carling, 2024; Kubayi and Toriola, 2019; Mićović et al., 2023), comprehensive perspectives into the mechanisms of goal production throughout the match remain limited (Sarmento et al., 2022). The existing research also indicates that different match statuses—reflected in the goal differences between teams—prompt distinct technical-tactical behaviours. However, the available data are still too sparse and heterogeneous to draw generalisable conclusions (Sarmento et al., 2022). Therefore, it is paramount to explore how various situational variables impact key performance indicators (KPIs) that characterise goal-scoring offensive sequences in elite soccer.
This study aimed to examine the impact of four situational variables, i.e., competition stage, match period, match status, and team quality, on offensive sequences leading to goals during the 2022 FIFA Men’s World Cup in Qatar. By integrating a comprehensive set of KPIs alongside these situational factors, this research sought to provide a finer-granular analysis of goal origins in this globally celebrated tournament (Branquinho et al., 2024; Iván-Baragaño et al., 2024; Yan et al., 2024). The findings were expected to enrich strategic and tactical preparations for the World Cup, offering valuable insights for refining team strategies and developing practical guidelines tailored specifically to the demands of this premier sporting event.
Methods
Sample
The sample consisted of all offensive sequences resulting in goals during the regular time (n = 168) in the 2022 FIFA World Cup held in Qatar. This prestigious tournament comprised two distinct stages: (1) the 32-team group stage, where national teams were divided into eight groups of four teams, each playing three matches against their pool opponents, in a total of 48 matches; and (2) the knockout phase that followed the group stage and consisted of single-elimination matches leading to the crowning of the world champion. The knockout phase included a round of 16, quarter-finals, semi-finals, play-off for the third place, and the final (totalling 16 matches).
The offensive sequences analysed in this study encompass goals scored during both the group stage and the knockout phase of the tournament. Due to the low number of instances, goals scored during extra-time—specifically, four goals from two matches (Croatia vs. Brazil in the quarter-final and Argentina vs. France in the final)—were excluded from the analysis. Match recordings were obtained from FIFA-authorised TV broadcasts and converted to MP4 format for further analysis. The study adhered to the Declaration of Helsinki’s recommendations and was approved by the Scientific Board of the Department of Psychology and Physical Education at the Manuel Teixeira Gomes Higher Education Institute (ISMAT), Portimão, Portugal (approval code: 004-23; approval date: 22 November 2023).
Design
We employed a retrospective observational methodology to conduct a comparative analysis of goal-scoring offensive sequences. Following the observational methodology framework proposed by Portell et al. (2015), this study was classified as the sixth of eight possible types of observational study designs: multidimensional (considering multiple response levels, including quantitative and categorical variables), nomothetic (analysing multiple teams and matches), point/single (focusing on a single tournament edition), and extensive (using static performance indicators).
Variables and Procedures
An offensive sequence, defined as a series of individual and/or collective actions leading to a goal, was delineated based on Almeida’s criteria (2019). To characterise the offensive sequences, we utilised the Offensive Sequences Characterisation System (OSCS), a notational analysis system encompassing both simple and composite performance indicators (Almeida, 2019). Composite indicators, combining two simple variables, offered broader insights into offensive play. The OSCS underwent validation by an external panel of experts, comprising two UEFA Pro coaches with top-tier European soccer experience and two independent sports science researchers with over a decade of expertise in performance analysis. For face and content validity, the experts reviewed the variables and operational definitions, provided feedback, and suggested refinements. After incorporating these changes, full consensus was reached and the system was validated. Detailed operational definitions for these quantitative dependent indicators can be found in Table 1.
Table 1
Performance-related dependent indicators and operational definitions in the Offensive Sequences Characterisation System (OSCS) (Almeida, 2019).
In addition to the performance indicators derived from the OSCS, three categorical variables were included as dependent variables: ball recovery type, ball recovery location (Almeida et al., 2014), and team possession type (Sarmento et al., 2018). The offensive sequences were further analysed based on four independent variables: competition stage, match status, match period, and team quality. Table 2 provides an overview of the type, categories, and operational definitions of each categorical variable included in the study.
Table 2
Categories, operational definitions, and collection procedures of categorical variables in the study.
| Variable | Categories | Operational definition/collection procedures |
|---|---|---|
| Independent | ||
| Competition stage | 1) Group stage 2) Knockout phase | Indicates the stage of the competition during which the goal-scoring event occurred. |
| Match status | 1) Losing 2) Tied 3) Winning | Represents the evolving score of a match immediately before the goal-scoring event. Categories were defined in relation to the number of goals scored and conceded by the scoring team at the time of data entry. |
| Match period | 1) 1–30 min 2) 31–60 min 3) 61 min–FT | Recorded as 1, 2 or 3 depending on the period in which the goal was scored during the match (Almeida et al., 2014). Note: FT – full-time. |
| Team quality | 1) Worse-ranked 2) Similarly-ranked 3) Better-ranked | Represents the quality difference between the scoring team and its opponent. Considering the FIFA men’s world ranking just before the competition (October 6, 2022), a k-means cluster analysis was conducted on total points for grouping teams into three quality categories. For example, if the scoring team was playing against an opponent from a lower quality group, it was categorised as “better-ranked”. |
| Dependent | ||
| Ball recovery type (Almeida et al., 2014) | 1) Interception | When the defender prevents a ball passed by an opponent from reaching its intended receiver by contacting the ball and keeping his own team in possession of the ball. |
| 2) Tackle | When the defender dispossesses the opponent of the ball through a physical challenge or defensive pressure. | |
| 3) Goalkeeper save | When the goalkeeper prevents the opposing team from scoring a goal after any kind of shot, i.e., a kick, a header or any intended deflection of the ball toward a goal. | |
| 4) Set piece | Static situations deriving from opponents’ misses or fouls (goal kicks, thrown-ins, off-sides, and free kicks), and opponents’ goals. | |
| 5) Turnover won | When the defender collects, somewhere in the pitch, a ball lost (clearances or missed passes) by the opposing team. | |
| Ball recovery location (Almeida et al., 2014) | 1) Defensive 2) Defensive midfield 3) Offensive midfield 4) Offensive | Determined by dividing the pitch into 4 transverse zones with the same size. |
| Team possession type (Sarmento et al., 2018) | 1) Counterattack | Rapid progression of the ball using a degree of imbalance from the ball recovery zone to the finishing zone. Ball circulation occurs more in depth than in width. Reduced number of passes (≤5). Reduced number of players intervening directly on the ball (≤4). Reduced time of the offensive sequence (<12s). |
| 2) Fast attack | Ball circulation is performed in width and depth with short and quick passes. Reduced number of passes (≤7). A maximum of 6 players with direct intervention on the ball. The sequence time has a maximum of 18 s. | |
| 3) Positional attack | Opposing team displays a balanced defensive organisation. Ball circulation is performed more in width than in depth, predominantly with short passes. High number of passes (>7) and players involved (>6). Long offensive sequence duration (>18 s). | |
The notational analysis was conducted using VLC Media Player (VideoLAN® Organisation, France) to review video footage of goal-scoring offensive sequences. Microsoft® Excel 365 (Microsoft® Corporation, USA) was used to manually code the quantitative and categorical performance variables selected to characterise each offensive sequence. Afterwards, the Excel database was exported to Statistical Package for the Social Sciences (SPSS®), version 28.0 (IBM® Corporation, USA), for data analysis.
Reliability
Reliability in using the OSCS and the categorical variables was assessed through intra- and inter-observer testing procedures. Three observers were involved in the reliability assessment: the author (Ob1) and two other previously trained observers (Ob2 and Ob3), also co-authors of the study. All observers had more than ten years of experience in performance analysis techniques. Prior to reliability assessments, all observers completed three pilot sessions, each coding ten randomly selected goal-scoring offensive sequences from the 2021/2022 UEFA Champions League. Discrepancies were reviewed after each session to harmonise interpretations and maintain uniform application of the OSCS criteria.
To assess intra-observer reliability, Ob1 completed a test-retest protocol with a six-week period separating both sessions to prevent potential learning effects. Twenty offensive sequences were randomly selected from the total sample (~12%) for notation in each session. For inter-observer reliability assessment, the three observers participated in a coding session where they independently notated the same set of offensive sequences. The intraclass correlation coefficient (ICC) and Weighted kappa (κw) were calculated to evaluate intra- and inter-observer agreements. The ICC was primarily used for discrete performance variables, while κw was applied for categorical variables and discrete indicators with low counts. Table 3 displays the intra- and inter-observer reliability results.
Table 3
Intra- and inter-observer reliability for quantitative and categorical performance-related variables.
The results for intra-observer reliability testing demonstrated an overall excellent level of agreement for both the ICC (>0.99) and κw (>0.91). For inter-observer reliability, ICC values were consistently high (>0.98), indicating excellent agreement among observers. The strength of agreement measured by κw ranged from “good” (team possession type, Ob1 and Ob2: κw = 0.77) to “very good” (κw > 0.8) (O’Donoghue, 2010).
Statistical Analysis
The effects of competition stage, match status, match period, and team quality on offensive performance indicators were initially explored using descriptive statistics, including medians, interquartile ranges, and relative frequencies. Since the assumptions for applying parametric tests were violated, specifically the assumption of normality, non-parametric tests were implemented. Mann-Whitney U tests were conducted for the competition stage and Kruskal-Wallis tests for match status, match period, and team quality.
The effect sizes (ES) for Mann-Whitney tests and post-hoc pairwise comparisons of Kruskal-Wallis tests were calculated employing the following equation (1), as indicated by Field (2018):
where r was the ES estimate for the Mann-Whitney test or the pairwise comparison of Kruskal Wallis test, Z the z-score produced by SPSS and N the number of total observations on which Z was based.
The interpretation of ES followed the benchmarks proposed by Cohen (1992): small (r ≥ 0.1), medium (r ≥ 0.3), and large (r ≥ 0.5). For categorical variables, we conducted Chi-square tests of independence to evaluate their association with ball recovery type, ball recovery location, and team possession type. Adjusted standardised residuals (AR) were calculated to further investigate the relationship between independent and dependent categorical variables. Residuals were deemed significant if they fell beyond the ±2.0 range, indicating frequencies higher or lower than expected with 95% confidence. ES were calculated using Cramer’s V statistic and assessed based on Cohen’s benchmarks (1992) for different degrees of freedom. The degrees of freedom for Cramer’s V were determined as the smaller of (R−1) or (C−1), with R representing rows and C representing columns (Gravetter and Wallnau, 2013). Subsequently, multinomial logistic regression was employed to explore how factors influenced categorical performance-related variables (ball recovery type, ball recovery location, and team possession type), with “interception”, “defensive zone”, and “positional attack” serving as reference categories. A significance level (α) of 0.05 was established for all statistical analyses.
Results
This section is divided into three subsections to enhance reader comprehension: simple performance indicators, composite performance indicators, and categorical performance variables.
Simple Performance Indicators
Key metrics such as duration, players involved, touches, passes, crosses, shots, and set pieces were examined across competition stage, match status, match period, and team quality. Table 4 summarises the median (Md) and interquartile range (IQR) for these indicators, highlighting significant results.
Table 4
Descriptive statistics for simple performance indicators of goal-scoring offensive sequences in the 2022 FIFA World Cup, presented as Medians and Interquartile Ranges (Md (IQR)).
Mann-Whitney U tests showed no significant differences between the group stage and the knockout phase across performance indicators. ESs were generally trivial, except for set pieces (r = 0.14; small).
The Kruskal-Wallis tests did not reveal significant differences across different match statuses (losing, tied, and winning) and match periods (1–30 min, 31–60 min, 61 min–FT) for any performance indicators. Minimal ES further endorsed the lack of significant differences. However, there were significant differences between worse-, similarly, and better-ranked teams for several performance indicators. Worse-ranked teams had significantly shorter offensive sequence duration than better-ranked teams (p = 0.045; r = −0.22, small ES). Lower-ranked teams involved fewer players (p = 0.044; r = −0.22, small ES) and performed fewer ball touches (p = 0.002; r = −0.31, medium ES) and passes (p = 0.023; r = −0.23, small ES) in their successful offensive sequences compared with better-ranked teams. Also, when building up to score, similarly ranked teams performed fewer touches on the ball (p = 0.014; r = −0.25, small ES) than better-ranked teams.
Composite Performance Indicators
Table 5 depicts composite performance indicators by competition stage, match status, match period, and team quality, outlining central tendencies, variability, and significance across different situational circumstances.
Table 5
Descriptive statistics for composite performance indicators of goal-scoring offensive sequences in the 2022 FIFA World Cup, presented as Medians and Interquartile Ranges (Md (IQR)).
Mann-Whitney U tests revealed no significant disparities between competition stages across composite performance indicators. No statistically significant differences were observed, though ES varied across composite performance indicators. Effect sizes ranged from trivial (Touches/Players, r = 0.002) to large (Touches/Duration, r = 0.079; Passes/Duration, r = 0.056; Passes/Players, r = 0.051), with medium-sized effects observed for Players/Duration (r = 0.033), Passes/Touches (r = 0.041), and Goals/Shots (r = 0.041). These variations may be attributed to random variability within groups.
Match status and match period did not produce differences across composite performance indicators. However, Kruskal-Wallis’ procedures uncovered important distinctions between worse- and better-ranked teams. As a means to achieve a goal, low-ranking teams employed offensive sequences with a slower tempo of intervention on the ball (Touches/Duration: p = 0.012; r = −0.26, small ES), and with fewer ball touches (Touches/Players: p = 0.006; r = −0.28, small ES) and passes per player involved (Passes/Players: p = 0.038; r = −0.23, small ES) compared with better-ranked teams.
Categorical Performance Variables
Tables 6–8 provide the frequencies, Chi-square test details, and adjusted residuals for each dependent categorical variable.
Table 6
Ball recovery type distribution (absolute and relative frequencies) by competition stage, match status, match period, and team quality.
[i] Note 1: no significant association between independent variables and the ball recovery type was found (p > 0.05); + higher observed frequency than expected; − lower observed frequency than expected Note 2: Cramer’s V values of 0.165 (competition stage), 0.091 (match status), 0.193 (match period), and 0.127 (team quality)
Table 7
Ball recovery location distribution (absolute and relative frequencies) by competition stage, match status, match period, and team quality.
Table 8
Team possession type distribution (absolute and relative frequencies) by competition stage, match status, match period, and team quality.
[i] Note 1: no significant association between independent variables and the team possession type was found (p > 0.05); + higher observed frequency than expected; − lower observed frequency than expected
Note 2: Cramer’s V values of 0.168 (competition stage), 0.035 (match status), 0.120 (match period), and 0.147 (team quality)
The Chi-square analysis revealed no significant associations between the independent variables and ball recovery type (p > 0.05). Despite the lack of statistical significance, small associations were generally observed. Set plays were the most common recovery type during the group stage (29.2%) and the knockout phase (41.7%). Teams in a losing position recovered the ball most frequently through interceptions (28.2%), while set plays were dominant when the match was tied (38.4%) or the team was winning (28.6%). In the first 30 min, set plays (32.3%) and turnovers won (32.3%) were prevalent. Between 31 and 60 min, tackles (26.6%) were higher than expected (AR = 2.5), whereas in the last third of the match interceptions (32.9%) were remarkably higher (AR = 2.1), and tackles (9.6%) were lower than expected (AR = −2.3). Better-ranked teams primarily recovered the ball through interceptions (29.8%) and set plays (28.6%). Worse-ranked teams had more recoveries through set plays (33.3%).
Table 9 presents the variable estimates for ball recovery type. The logistic regression model confirmed the influence of match period on the dependent variable. In short, the odds of regaining possession through a tackle instead of an interception significantly increased by 398.4% (p = 0.006) in the middle of matches relative to the last 30 min (plus additional time). The situational factors of competition stage, match status, and team quality did not have significant effects (p > 0.05).
Table 9
Variable estimates for the multinomial logistic regression of ball recovery type as a function of situational factors in FIFA World Cup 2022.
Furthermore, a significant association was found between match period and ball recovery location (p = 0.038). During the first 30 min, the defensive midfield (45.2%) was the most common recovery zone. Between 31 and 60 min, offensive recoveries (17.2%) were higher than expected (AR = 2.4). In the last 30 min, the defensive zone (47.9%) recoveries were prominent. The Chi-square analysis indicated no significant associations between the other independent variables and ball recovery location (p > 0.05); nonetheless, some patterns emerged, suggesting trivial (competition stage) to small associations (match status and team quality). The defensive zone was the most common recovery area in both the group stage (39.2%) and the knockout phase (45.8%), when matches were tied (49.3%) or the scoring team was winning (41.1%). When the score was unfavourable (losing), the ball recovery frequency in the defensive zone (25.6%) was lower than expected (AR = −2.2), occurring most regularly in the defensive midfield (38.5%).
The multinomial logistic regression model (Table 10) also indicated that match status and match period significantly affected the ball recovery location. The probability of recovering the ball preceding a goal scored in the defensive midfield, rather than the defensive zone, was 63.1% (p = 0.04) lower with an equalising score, compared to when ahead. On the contrary, when losing, teams increased by 530.6% (p = 0.045) the likelihood of regaining the ball in the offensive zone (vs. defensive one) compared to when winning. Moreover, the chances of regaining a goal-scoring possession in the defensive midfield (vs. defensive) significantly increased by 252.9% (p = 0.025) in the first third of the match compared to the final third. Similarly, there were increases of 321.3% (p = 0.009) and 301.2% (p = 0.033) in recovering the ball in the offensive midfield and offensive zones, respectively, rather than in the defensive one, between 31 and 60 min in relation to the last match period.
Table 10
Variable estimates for the multinomial logistic regression of ball recovery location as a function of situational factors in FIFA World Cup 2022.
Although no significant associations were unveiled between the independent variables and team possession type (p > 0.05), there were observable trends with practical significance. The associations were classified as trivial for match status and small for competition stage, match period, and team quality. Positional attacks were the dominant possession type across all categories, particularly for better-ranked teams, where the observed frequency exceeded the expected value (AR = 2.7). In the 31–60-min period, there was a lower frequency of positional attacks than expected (AR = −2.2). Fast attacks became more prominent during the knockout phase (AR = 2.1) compared to the group stage, where the observed frequency was lower than anticipated (AR = −2.1). Counterattacks were the least common possession type overall, but were slightly more frequent for worse-ranked teams (19.4%).
The multinomial logistic regression analysis further confirmed the absence of significant effects of situational factors on team possession type (Table 11). Notably, the models explained a small portion of the variance in team possession type and a moderate portion in ball recovery type and location.
Table 11
Variable estimates for the multinomial logistic regression of team possession type as a function of situational factors in FIFA World Cup 2022.
Discussion
Research on performance analysis, especially match analysis, has significantly advanced over the past decade, driven by emerging technologies and the proliferation of data providers (Sarmento et al., 2022). However, numerous scholars have emphasised the need for more focused studies on technical-tactical performance indicators while accounting for regularly overlooked situational variables (Barthelemy et al., 2024; Lago-Peñas et al., 2023; Praça et al., 2024; Yan et al., 2024). Addressing this gap, this study examined the impact of competition stage, match period, match status, and team quality on KPIs characterising goal-scoring offensive sequences during the 2022 FIFA World Cup.
The main findings illuminate the influence of situational variables on successful offensive sequences in elite soccer. Team quality proved pivotal, as better-ranked teams exhibited more complex and prolonged offensive actions than their lower-ranked counterparts. Surprisingly, variables such as competition stage, match period, and match status did not significantly alter simple and composite performance indicators. This suggests that teams maintained a consistent strategic-tactical approach during the build-up and progression towards scoring. Nonetheless, the analysis revealed that match period and match status affected both the manner and the location of possession regains critical to setting up goals. Though team quality was the dominant factor in determining the effectiveness of offensive strategies, there was a subtle interaction between the match context and team performance that predominantly influenced defence-to-offence transition behaviours as matches unfold.
Since the 2018 World Cup, distinguishing between teams eliminated at the group stage and those progressing to the knockout phase has become a common method for analysing team performance (Iván-Baragaño et al., 2024; Kubayi and Larkin, 2022; Yi et al., 2019). Although this approach represents a useful proxy for team strength, it neglects the varied competitive demands characteristic of each tournament stage. Our findings revealed no significant performance differences between the group and knockout stages, suggesting that elite national teams were well-prepared to manage both the pressures and elevated stakes of knockout rounds (Fernandes et al., 2020). However, the observed variation in effect sizes across different KPIs signalled complex performance dynamics not fully captured by median comparisons alone. In the knockout phase, there was a meaningful increase in fast attacks, accompanied by an increment in the tempo of ball interventions and collective involvement in goal-scoring events. The trend towards quicker, more aggressive possession styles likely represents a tactical adaptation to better exploit unusual imbalances in opponents’ defensive organisation (Almeida, 2019; Sarmento et al., 2018).
At the match level, segmenting full-time matches into smaller intervals has been widely recognised as a crucial situational factor in soccer performance analysis, with numerous studies stressing its significant impact on various performance metrics (Sarmento et al., 2022). A higher incidence of goals towards the end of matches (76 min–FT) has been frequently noted, likely due to fatigue accumulation (Aguado-Méndez et al., 2020; Degrenne and Carling, 2024; Kubayi and Toriola, 2019; Mićović et al., 2023). This observation suggests that teams might shift their strategic-tactical approaches as matches draw to the final whistle.
Despite the absence of significant changes in simple and composite performance indicators with match progression, categorical variables remained sensitive to timing, pointing to intricate dynamics between match period and performance. During the mid-game (31–60 min), teams markedly adjusted their defensive tactics to initiate successful offensive sequences. The increased defensive aggressiveness, demonstrated by an increase in ball recoveries through tackles, coincided with shifts in ball recovery location, moving from defensive midfield (early game) to more offensive zones (mid-game). Aguado-Méndez et al. (2021) similarly reported a threefold increase in the likelihood of conceding a goal-scoring opportunity due to an opponent’s steal, rather than a turnover, during the second half, which underscores the importance of defensive pressure as play evolves. Furthermore, this aggressive defence facilitated the employment of quicker attacking styles, such as fast attacks instead of positional play, seemingly exploiting opponents’ disarray during transitions to breach less organised defensive lines (Aguado-Méndez et al., 2020, 2021; Lago-Ballesteros et al., 2012; Sarmento et al., 2018). As matches approached the end, there was a noticeable increase in goal-scoring sequences originating from the defensive zone. This may reflect teams’ ability to maintain tight formations and exploit opportunities while arguably defending a favourable scoreline.
Transitioning from the influence of match progression, it is paramount to explore how match status (winning, tied or losing) affects the successful technical-tactical behaviours of high-level teams. Notwithstanding the multiple existing research insights, the data on this relationship remain mixed and inconclusive (Sarmento et al., 2022). Our analysis uncovered no significant differences in simple and composite performance indicators based on the scoreline, contrasting with prior studies positing that teams behind in a match tend to increase possession and passing frequency to control the game (Konefał et al., 2018, 2019; Lago, 2009; Lago-Ballesteros et al., 2012). Instead, teams in the 2022 FIFA World Cup maintained a consistent attacking approach aimed at scoring, regardless of the evolving match status. Perhaps the exclusive focus on goal-scoring offensive sequences reduces the sensitivity needed to detect fine-grained changes occurring during match-status maintenance or transition phases (Aguado-Méndez et al., 2021; Konefał et al., 2019; Schulze et al., 2022). A broader analytical approach may better capture the tactical and behavioural adjustments teams make in response to the current score. Nonetheless, we detected a significant effect of match status on ball recovery location: teams were more likely to recover possession in offensive zones when losing, a trend also corroborated in previous research (Almeida et al., 2014; Santos et al., 2017). This shift conceivably reflects tactical adjustments, such as pushing defensive lines higher and applying aggressive pressing to regain possession and exploit vulnerabilities. Heightened urgency after conceding may also lead players to take greater risks in creating scoring opportunities from these offensive recoveries (Santos et al., 2017).
Unlike Lago (2009), who emphasised the evolving scoreline as the critical variable influencing possession and play zones, our findings corroborate those of Castellano et al. (2013), which pinpoint team quality as the most influential factor in successful offensive sequences. Both quantitative performance indicators and categorical variables demonstrated a clear tendency for better-ranked teams to adopt a more controlled and structured approach when building from the back and progressing into the final third to score. Such structured play is often linked to team spatial distribution, with higher offensive width, length, and surface area observed in the offensive phase against weaker opponents (Castellano et al., 2013). In practice, this resulted in longer sequences, greater player involvement, and more touches and passes compared to lower-ranked teams (Zhou et al., 2019). These results uphold prior findings from the same tournament edition (Branquinho et al., 2024; de França et al., 2024; Pan et al., 2024), where stronger teams favoured positional attacks to score. Of note, these teams improved their efficiency by increasing the tempo of ball intervention and involving more players in passing connections (Pan et al., 2024; Taha and Ali, 2023; Yi et al., 2019). Interestingly, during defence-to-attack transitions, and contrasting previous studies (Iván-Baragaño et al., 2024; Praça et al., 2024; Santos et al., 2017), better-ranked teams did not recover possession more often in advanced pitch zones to create goal-scoring events. Rather, their attacking success relied less on where possession was regained and more on how they moved the ball once in control.
Limitations and Future Research
Some limitations must be considered when interpreting the findings. First, the analysis focused solely on goal-scoring sequences, which represent the least frequent type of attacking plays in a match. Consequently, the data may not fully capture the broader spectrum of offensive dynamics (Aguado-Méndez et al., 2021; Schulze et al., 2022). Future research should expand the sample to include non-successful offensive sequences thereby providing a more holistic understanding of team performance. Additionally, due to the relatively small sample size, the match period and match status variables were consolidated into fewer categories, theoretically limiting the depth of the analysis. More extensive studies could mitigate this limitation by using a more granular segmentation of both variables. Second, whereas team quality was assessed using FIFA’s rankings before the tournament via K-means clustering, this may not entirely reflect a team’s strength or consistency throughout the World Cup. Future research could explore models that account for fluctuations in team performance over the competition. Third, the tool used in this study relied on static variables rather than sequential ones, restricting the ability to examine the complex and dynamic nature of soccer. Upcoming studies should incorporate sequential analyses, such as time-series, lag sequential, or t-pattern approaches, to more accurately depict behavioural changes throughout matches. Lastly, improving the multinomial logistic regression models with larger datasets from multiple World Cup editions would enhance the robustness and generalisability of the findings, yielding deeper insights into the evolving trends of elite soccer.
Practical Implications
From the study’s main findings, we propose four practical applications to help coaching staff enhance performance in high-level tournaments:
Teams can analyse performance metrics in finer detail to identify strategic shifts, such as increasing dynamic ball movement and player rotations, to improve outcomes against stronger defences in later stages of the match.
Coaches may adopt an ecological approach to foster adaptive defensive strategies, encouraging increased assertiveness as the match progresses. Training should focus on recovering the ball in advanced areas to facilitate faster attacks by exploiting weaknesses in the opponent’s defensive transitions.
Manipulating match status during training, along with other task constraints (e.g., numerical relation), can simulate real-game scenarios, promoting higher defensive lines and aggressive pressing to regain possession near the opponent’s goal.
Adaptive attacking strategies based on relative team strength are paramount. Lower-ranked teams might focus on increasing the speed and efficiency of ball movement. At the same time, higher-ranked sides can optimise controlled, structured play, balancing positional and faster styles to suit different match phases.
Conclusions
This study stresses the crucial role situational variables play in shaping goal-scoring sequences in elite soccer, with team quality emerging as the most decisive factor. Better-ranked teams demonstrated longer, more structured attacks with a greater tempo of ball intervention. Although competition stage, match period, and match status did not substantially impact KPIs related to build-up and progression, they influenced defensive-to-offensive transitions, especially in mid-game and when teams were behind. Ultimately, elite soccer’s success is driven by both team quality and situational adaptability. To optimise performance under varying match circumstances, coaches must account for these variables in training and match preparations, as fostering a team’s adaptability to the ever-changing contextual dynamics is fundamental to succeed in modern soccer.
