Performance profiles and match-to-match variability of the most demanding passages during the FIFA World Cup Qatar 2022: the effect of playing positions and match contextual factors

Abstract

Understanding how player performance profiles and match-to-match variability differ across playing positions in short competitions can provide insight into how players adapt to dynamic, match-related environments. This study investigated the most demanding passages of kinematic (running distance; MDPk) activity during the 2022 FIFA World Cup, focusing on player profile and match-to-match variability for five different positions and considering various contextual factors: tournament phase, team ranking, match status, and match outcome. The MDPk (> 15 km/h) was identified across three rolling average durations (1-, 3-, and 5-minutes(min)) and analyzed according to its constituent variables, moderate-speed running (MSR), high-speed running (HSR), and sprinting (SPR), as well as the total distance covered (TD) within each epoch. Effect sizes were used to assess differences in player performance profiles and match-to-match variability across positions according to all contextual factors. Regarding players’ performance profiles, trivial to moderate effects were observed across all contextual factors, with a stronger influence in match status. Where higher distances were recorded, when the team was winning, for full-backs in the 5-min window TD (d = -0.41 [-0.61; -0.21]), and MSR (d = -0.42 [-0.62; -0.22]) and for center forward in the 3-min window SPR (d = -0.58 [-0.87; -0.28]). All positions showed small to large match-to-match variability effects across different contexts. This was particularly evident for top tier teams, where higher variability was noted in the 1-min window for central defenders in TD (d = -0.51 [-1.12; -0.01]), MDPk (d = -0.74 [-1.32; -0.17]) and SPR (d = -0.71 [-1.17; -0.12]). Similar achieved by central midfielder, with higher variability in MDPk in 3-min window (d = -0.55 [-1.07; -0.07]) and 5-min window (d = -0.79 [-1.35; -0.24]). These findings provide deeper insight into MDP profile and variability for each position based on different rolling average durations and specific contextual factors during an international tournament, thereby offering valuable context for player performance analysis.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13102-026-01578-z.

Introduction

Several studies on soccer have analyzed the match performance and its complexities, revealing the diverse challenges faced by players [1–4]. From that, the importance of understanding the physical and tactical demands placed on players during competition have been highlighted. Though, a deeper comprehension of these demands is essential for enhancing individual and collective performance and advancing the overall quality of the competition [5].

Advancements in sports science, particularly the development of tracking technologies, have made it increasingly practical to monitor players’ performance during both training and competition [6]. For example, to further understand physical demands placed on players in training and competition, data collection on players’ external loads have been recommended and considered both by researchers and practitioners [7].

Previous methodologies analyze the average demands experienced by players during match or training session, as recent research has shifted towards analyzing the most demanding passages (MDP) of match play to capture the most intense moments during competition. These high-intense moments, although brief, involve minimal recovery time and are thus crucial for optimizing performance and preventing injuries [8]. Therefore, it is essential not only to understand the overall MDP, but also to examine the profiles in the different playing positions, as they reflect the unique demands associated with each playing role.

Several studies suggest that the most accurate approach to analyze MDP is by applying a rolling average method to predefined time windows (e.g., 1, 3, 5 min) to identify the most demanding moments of the match [9–11]. This method has been applied to identify the MDP in various team sports, including rugby [12], Australian Football [13], and soccer [11].

Recent studies have explored a novel approach to analyze MDP from a multifactorial perspective (both mechanical and kinematical) [14, 15]. For the kinematic approach, the authors proposed the use of the rolling average for distance performed above 15 km/h and divided these moments into their constituent variables, moderate-speed running (MSR), high-speed running (HSR) and sprinting (SPR). However, the extent to which match-specific contextual factors and the variability of MDPk constituent variables (MSR, HSR, and SPR) influence performance require further investigation.

To understand the variability between the constituent variables of the MDP, it is important to consider a range of contextual factors. These include match location, quality of opposition, and match outcome [16–19]. Additionally, players’ different positional roles influence performance characteristics, as different demands are placed on players depending on their tactical responsibilities [20]. For instance, previous research has reported differences between central defenders and midfielders, between midfielders/full-backs and forwards/wide midfielders [20], and even between full-backs and central defenders [21].

Considering these findings, an evaluation of different contextual factors is crucial to further understand the performance profile and variability in each position. Characterizing this positional variability is critical for evidence-based decisions related to player selection, tactical planning, load management, and recovery strategies, as it captures the distinct demands imposed by each role within condensed tournament schedules. In fact, understanding how contextual factors influence the MDP demands in the context of tournaments, such as the World Cup, is extremely important for national teams. Previous research has shown that players tend to exhibit higher physical outputs during critical stages of key tournaments, such as semi-finals and finals [22], and there is a higher total distance covered when a team is winning against higher-ranked opponents [23].

Therefore, the main objective of this study was to investigate the performance profile and match-to-match variability of the MDPk for specific positions during the FIFA World Cup Qatar 2022. While prior research has identified differences in various univariate metrics for the MDPs in soccer [24–26]. , the present study specifically focuses on analyzing the profile and quantifying the variability of the MDPk and its constituent variables.

Materials and methods

Data source

Access to data was provided through the official FIFA Data Plaform. Positional data were collected for all player across 64 matches of the FIFA World Cup Qatar 2022 using a multicamera computerized tracking system (TRACAB, Chyron Hego). The system operates with high-definition cameras at 25 Hz. The validity and reliability of this system have been previously established [27]. Goalkeepers were excluded due to their differential position demands. Only outfield players who started and played for at least 70 min of the match were included (399 players) [26], to eliminate the effects of substitutes, whose physical output may reflect distinct strategies [28]. Previous research has demonstrated that substitute players display distinct peak running demands during match play compared with starting players [25], with the majority of substitutions typically occurring after approximately 70 min of play [29]. Furthermore, evidence from the literature indicates that most maximal demand periods (MDPs) occur during the first half of the match and within the initial 15–20 min of the second half [14, 30].

Data treatment

Raw data from each match were downloaded from the FIFA Data Platform. Raw positional data for each match of the tournament was collected and processed using Python 3.8. For all matches, data were analyzed for the full match duration, including any stoppage time but excluding extra-time [31]. The player’s speed was calculated from the raw positional data, and outliers of unphysical human speeds (> 11 m/s) were replaced with the mean of adjacent data.

A multifactorial kinematic variable (MDPk) was identified based on the total distance covered at speeds greater than 15 km/h, consisting of the maximum sum of the rolling average across the defined durations and composed by moderate speed running (MSR; 15–19.8 km/h), high-speed running (HSR; 19.8–25.2 km/h), and sprinting (SPR; >25.2 km/h) [14, 15]. The total distance performed within each time window (TD) was also considered. MDPk values were computed using a rolling average approach for each player in every match across 1-, 3- and 5-minute (min) durations. The maximum value of distance covered at speeds greater than 15 km/h was identified for each period and recorded as the MDP, generating a total of 3069 observations for 399 players, divided into the following five positional categories: full-back (FB, n = 89), central defender (CD, n = 80), wide midfielder (WM, n = 93), central midfielder (CM, n = 93) and center forward(CF, n = 44) [32].

The influence of four contextual factors on the MDPk variables was investigated: (1) tournament phase - group stage vs. knockout phases; (2) team ranking, with the top 11 national ranked teams based on the FIFA ranking (as of October 6, 2022) categorized as “Top Tier”, and the other as “Non-Top Tier”; (3) match status, defined as winning vs. not winning (drawing or losing); and (4) match outcome, defined as win vs. not win (draw or lost).

Statistical analysis

Match performance and variability were analyzed according to the following contextual factors: (a) tournament phase, (b) team ranking, (c) match status, and (d) match outcome, across all positional groups.

Two complementary analytical approaches were employed. First, to examine the match profiles, linear mixed-effects models were used, with players included as random effects to account for repeated measures. This approach enabled the evaluation of the effects of contextual factors on the MDPk components while accounting for the hierarchical structure of the data (i.e., multiple match observations nested within players). Effect sizes (Cohen’s d, ± 95% CI) for these analyses were calculated using the estimated marginal means derived from the mixed models. Second, to quantify the match-to-match variability, the coefficient of variation (%CV) was calculated for each player as the standard deviation divided by the mean and multiplied by 100 [19, 32]. Variability analyses included only players with at least two match observations (range: 2–7 matches) for each contextual factor. Because each player contributed a single CV value per condition, independent t-tests were used to compare between-group differences (e.g., between tournament phases, team ranking, match status, and match outcome). For these comparisons, Cohen’s d (± 95% CI) was computed using group means.

Thresholds for interpreting effect sizes were defined as: trivial (0–0.19), small (0.20–0.49), moderate (0.50–0.79), and large (≥ 0.80). All statistical analyses were performed using Python (version 3.8) in the Spyder Scientific Development Environment (https://www.spyder-ide.org/) with the DABEST package (version 0.3.1) [33]. Statistical significance was accepted at p ≤ 0.05. Data are presented as mean ± standard deviation (SD).

Results

Performance profiles and variability were analyzed across all contextual factors and for the different playing positions.

Tournament phase

Regarding the tournament phase, results for both performance profiles (see complementary Table 1) and variability (see complementary Table 2) were analyzed. Effect sizes for profile and variability by position, as well as MDPk components across tournament phases, are shown in Fig. 1. Positional differences varied between tournament phases, with effects ranging from small to moderate. During the group stage, CD showed higher distances, particularly in the 5-min HSR window (d = -0.41 [-0.63; -0.20]; p < 0.05), and WM in the 5-min TD (d = -0.42 [-0.61; -0.23]; p < 0.05). Conversely, in the knockout phase, CF covered greater distances in 3-min TD window (d = 0.69 [0.39; 0.99]; p < 0.05).

Fig. 1.

Effect sizes for match-to-match variability and player profiles between tournament phases across positions. Error bars indicate uncertainty with 95% confidence intervals

In terms of variability, moderate to large effects were observed during the group stage, with greater variability observed for CM in 5-min MSR window (d = 0.84 [0.25; 1.49]; p < 0.05). During the knockout phase, CD exhibited higher variability in 3-min TD window (d = -0.70 [-1.13; -0.07).

Team ranking

With respect to team ranking, results for both performance profiles (Complementary Table 3) and variability (Complementary Table 4) were analyzed. Regarding effect-size (Fig. 2), the profile analysis displayed, in top-tier teams, only CM in the 1-min MSR window (d = -0.32 [-0.51; -0.12]; p < 0.05) demonstrated higher distances. Regarding non-top-tier teams, FB greater distances in 3-min TD window (d = 0.45 [0.24; 0.65]; p < 0.05), MDPk (d = 0.53 [0.33; 0.74]; p < 0.05) and 5-min SPR window (d = 0.45 [0.25; 0.65]; p < 0.05.

Fig. 2.

Effect sizes for match-to-match variability and player profiles between ranking across positions. Error bars indicate uncertainty with 95% confidence intervals

Additionally, there was a small to moderate effect on variability, with top-tier teams showing greater variability for CD in 1-min TD window (d = -0.51 [-1.12; -0.01]; p = 0.05), MDPk window (d = -0.74 [-1.32; -0.17]; p < 0.05) and SPR window (d = -0.71 [-1.17; -0.12]; p < 0.05), 3-min MDPk window (d = -0.76 [-1.29; -0.28]; p < 0.05), 5-min MDPk window (d = -0.67 [-1.21; -0.11]; p < 0.05) and MSR window (d = -0.61 [-1.21; -0.06]; p < 0.05). Similar to CD, CM showed higher variability in MDPk in 3-min window (d = -0.55 [-1.07; -0.07]; p < 0.05) and 5-min window (d = -0.79 [-1.35; -0.24]; p < 0.05), 5-min MSR window (d = -0.64 [-1.21; -0.10]; p < 0.05) and 1-min SPR window (d = -0.77 [-1.32; -0.21]; p < 0.05). Regarding not top-tier teams, no significant results were achieved.

Match status

Concerning match status, results for both performance profiles (see Complementary Table 5) and variability (see Complementary Table 6) were evaluated. The analysis of match status effects on MDP (Fig. 3), showed several significant results when comparing winning and not-winning conditions across positions and time windows.

Fig. 3.

Effect sizes for match-to-match variability and player profiles between match status across positions. Error bars indicate uncertainty with 95% confidence intervals. Values in red do not present interval data, as they include fewer than three observations (n < 3)

In terms of profile, when the team was not-winning, the most substantial effect was observed for CF in the 3-min MSR window (d = 0.30 [0.03; 0.59]; p = 0.24) and 5-min SPR window (d = 0.30 [0.03; 0.58]; p = 0.26). In the opposite direction, when the team was winning, higher distances for FB in the 5-min TD window (d = -0.41 [-0.61; -0.21]; p < 0.05), and MSR window (d = -0.42 [-0.62; -0.22]; p < 0.05) and for CF in the 3-min SPR window (d = -0.58 [-0.87; -0.28]; p < 0.05).

In variability, notable differences were observed depending on whether the team was winning or not. The values regarding variability that had not enough data (less than 3 entries) were excluded from the analysis. When the team was not winning, higher variability was observed in CD in the 3-min TD window (d = 0.68 [-0.05; 1.23]; p < 0.05) and 5 min TD window (d = 0.64 [0.11; 1.06]; p < 0.05). The same for the FB position in the 3-min SPR window (d = 0.54 [0.09; 0.94]; p < 0.05). Additionally, similar to results of CD and FB, for the CM in the 5-min TD window (d = 0.91 [0.43; 1.35]; p < 0.05) and 3-min MDPk window (d = 0.63 [-0.04; 1.12]; p < 0.05). In contrast when the team was winning, higher variability was observed in the WM 1-min SPR window (d = -0.70 [-1.37; -0.11]; p < 0.05).

Match outcome

Based on the match outcome, results for both performance profiles (Complementary Table 7) and variability (Complementary Table 8) were examined. Regarding effect-size (Fig. 4), the most significant results showed small to moderate effects in profile and small to moderate effect on variability. Regarding the profile, higher distances were performed when the team won the match, for CM in the 1-min MDPk window (d = -0.35 [-0.55; -0.16]; p < 0.05). In the opposite way, when the team did not win, no significant differences were found.

Fig. 4.

Effect sizes for match-to-match variability and player profiles between match outcome across positions. Error bars indicate uncertainty with 95% confidence interval

In variability, when teams win the match, it was higher in CM in the 3-min MDPk window (d = -0.61 [-1.13; -0.07]; p < 0.05). Conversely. in teams that didn’t win the match, higher variability was observed for CM in the 3-min SPR window (d = 0.50 [0.01; 1.01]; p < 0.05).

Discussion

The present study aimed to investigate the effect of contextual factors on the performance profiles (distances covered) and match-to-match variability of the multifactorial kinematic MDP (MDPk) and its constituent variables across three durations. The primary findings highlight the progression of MDP values across the three time-windows and are consistent with previous research demonstrating a decline in relative intensity as the duration increases [10, 20]. These results provide important context for analyzing MDPk and emphasize the influence of positional roles under varying match contextual factors, which is essential for prescribing training loads during match-specific sessions.

The use of multifactorial kinematic MDP variables, comprising key univariate metrics such as distance covered across multiple velocity thresholds [14, 15]. Allows for better monitoring and prescription of training sessions. This approach allows practitioners to assess simultaneous demands across various speed zones, providing a more comprehensive understanding of players’ kinematic load during the most demanding passages. To our knowledge, no study has analyzed multifactorial kinematic MDP distances and their variability across different contextual factors while considering player positions. The analysis of within-player, match-to-match variability in MDPk under different match contextual factors, as showed in Fig. 1 (phase), Fig. 2 (ranking), Fig. 3 (match status), and Fig. 4 (match outcome). The results showed small-to-trivial differences across all contextual factors examined. This suggests that players’ performance remained relatively stable regardless of contextual factors analyzed.

Our findings are consistent with previous studies, that MDP demands are position-dependent [21, 30, 34, 35]. Specifically, WM and FB covered the greatest SPR distances, while CD covered the least [20, 34]. This suggests that winger roles, such as WM and FB, are more frequently engaged in high-speed actions than CD, who typically perform shorter runs [36, 37]. This pattern reflects tactical role differences, where wide players often offer support both offensive and defensive phases over larger areas, while central defenders control more constrained areas. Regarding TD, CM players covered the greatest distance, whereas CD reported the lowest values [20]. These positional variations are critical to consider, as high-intensity actions are closely linked to an increased risk of injury [38]. These positional differences are not merely observational as they have physiological and injury-prevention implications. Consequently, it is recommended that training drills based on MDP durations are tailored to meet the specific physical demands of each playing position. Additionally, in a tournament context, it is important to emphasize key aspects such as player fatigue and recovery monitoring, tactical adaptability, and substitution strategies.

During the tournament phase, the profiles in the MDP showed differences depending on the phase of the competition. During group stage, CD covered more distance HSR, corroborating that this position is more involved in high-intensity activities during less decisive phases of a tournament [39]. In the knockout phase, CF exhibited higher TD, particularly in the 3-min windows, suggesting greater physical demands as the tournament advanced, as it has been suggested by [36]. Understanding the positional differences in these demands during short-duration tournaments enhances the effectiveness of player load management and tactical planning, particularly when considering the specific stage of the competition. For variability, during the group stage, it is higher for CM, especially in the 5-min window for MSR. In knockout CD showed higher variability, especially for in 3 min TD. Although the results were not significant regarding the higher variability in the knockout phase, previous research indicate that players in highly competitive scenarios, such as elimination matches or high-stakes tournaments, showed greater variability and elevated performance metrics [40].

Regarding ranking, FB positions from not-top tier teams, performed greater distances in TD, MDPk and SPR during longer windows (3- and 5-min), these results may be explained by the existence of intense sprint efforts as a tactical approach to take part in both attacking and defending duties [20]. Likely reflect a tactical emphasis on sustained high-intensity involvement in both offensive and defensive phases. This supports the idea that these teams utilize FB as dynamic, dual-role players who frequently engage in repeated sprint efforts to support transitions. In contrast, for 1-min windows, for CM had higher distances from top-tier teams, in MSR, this could be explained by the top teams, that have higher levels of collective synchronization allow the ability to explore space better and interpret the information [41]. This suggests a compensatory physical workload, possibly to better tactical cohesion. These teams may rely more on individual physical effort, that could lead to greater distances covered in positioning and spacing. Concerning variability, non-top tier teams showed no significant results, which could be an indicator of a bigger fluctuation in moderate activities, likely by the influenced by tactical strategies [40]. In contrast, CM from top tier teams have greater variability in MDPk for the 3- and 5-min window, that could be influenced by their duties, both defensively and offensively [42]. Also, in this, CD had higher variability in MDPk for 1-, 3- and 5-min windows, suggesting inconsistent patterns above moderate speed.

For match status, when the team is winning FB exhibits higher distances in TD and MSR. The increase in distances in the TD and MSR suggests that players require higher levels of effort to maintain defensive stability and transitions during winning status [43]. This reflects a tactical emphasis on controlled aggression and pressing even when in a favorable scoreline. In contrast, CF have higher MSR and SPR in the longer windows when not winning, which reflects a more aggressive and proactive approach to overturn the score [44]. On other hand, the significant variability was shown by WM when the team is winning, with higher variability in SPR during the 1-min window, indicating unpredictability, which aligns with wide midfielders required higher tactical flexibility in different match scenarios [11]. On contrary, CD and CM showed higher variability when the team is not winning, particularly in TD during 5-min windows, indicating that the players in these positions had higher adaptations to tactical changes in losing scenarios [39]. Overall, the data indicates that match status significantly influences both the profiles and variability, and that players adjust their behavior not only based on their role but also in response to the evolving tactical demands of the game.

In match outcome, when the result was a win, only CM experience higher distances, especially in MDPk. This may reflect tactical strategies focused on exploiting space and maintaining attacking pressure to preserve or extend the lead. When the team did not win, the analysis yielded non-significant results, indicating that the team’s final outcome did not have a statistically significant effect on individual player performance. Regarding variability, it was higher when the team wins for the CM in the MDPk in 3-min, potentially indicating their role in adjusting tempo, distributing play, and responding to shifting match dynamics to control the game. In contrast, when the team does not win, CM had higher variability in SPR in the 3-min window. Importantly, the observed profile and variability in both win and not-win conditions reinforce the notion that running performance alone is not a definitive predictor of match outcome. These findings suggest that players’ dynamics are more variable in both successful and unsuccessful matches, due to the dynamic nature of performance, and from where defends that running metrics are unrelated to final match outcome [43, 45].

It is also important to understand position-specific variability in physical demands for designing effective training periodization, especially during congested tournament schedules. Evidence shows that players in different positions experience distinct external load patterns, which fluctuate considerably match-to-match due to tactical roles, opponent strategies, and game context [11]. Such variability necessitates a flexible and individualized approach to load monitoring, ensuring that training loads are adjusted dynamically to accommodate fluctuating demands and optimize recovery [46]. Integrating these insights into periodization frameworks enables practitioners to prescribe targeted training stimuli, balance fatigue and mitigate injury risk while maintaining performance across consecutive matches [47]. Ultimately, adapting training to the nuanced variability inherent in tactical and positional demands enhances the relevance of training and supports consistent performance during competitive play.

Despite these findings, this study aimed to highlight the relationship between different components of the MDPk, positions on the field and selected contextual factors. While the present study provides valuable insights into the most demanding passages using multifactorial kinematic variables, certain limitations should be acknowledged. The MDP selected was based on transformed positional data, and it only focuses on kinematic MDP, data that may not capture all high-intensity actions, particularly short, explosive efforts such as changes of direction, contact situations, or brief accelerations in congested areas, that could potentially underestimate true physical demands. Also, factors such as individual player characteristics, opposition team strategies, and match location may influence these demands [48]. In addition, there’s no integration of any tactical or technical indicators that could bring a better contextual understanding [49]. While the current study identifies significant variability in the MDPs, it is important to consider that such variability may be partially explained by tactical formations and team strategies. The present study did not control the influence of team formations and tactical systems employed. This constitutes a major limitation for interpreting the results, especially considering our aim to quantify and compare performance profiles across different playing positions. Incorporating tactical data and video analysis in future research would enhance the ecological validity and practical relevance of such findings.

Conclusions

For match preparation, practitioners regularly use the average match data to understand the different demands. However, we suggest that incorporating MDP analysis can provide a more nuanced understanding of position-specific demands and the influence of contextual factors. Our research indicates that the relative intensities and variability in MDPk performance across three time-windows are position-specific, with differences observed according to the various components of the MDPk distances and contextual factors. Center forward, when not winning, covered greater distances in MSR and SPR over longer periods. In contrast, full-backs from non-top-tier teams covered higher distance in TD and MDPk across three minute windows. These findings provide deeper insight into position-specific magnitude of MDPk variables across different rolling average durations and contextual conditions. Furthermore, the results highlight the importance of considering position-specific variability in MDPk as a novel and relevant metric for evaluating physical performance during match play. Measuring variability in MDP is essential for contextualizing performance data practical application. Based on these findings, practitioners are advised to incorporate multifactorial MDPk into training design and load monitoring by replicating the intensity and complexity of the most demanding passages during drills, ensuring that players are adequately prepared for peak match demands. Such contextualization allows coaches and practitioners to detect meaningful changes at the individual level under different contextual factors, supporting more accurate comparisons of physical performance across matches.

Authors’ contributions

AC, BG; Responsible for conception and design of the project. AC, FY; Data collection. AC, BG; Performed data analysis, interpreted the data. AC, FY, HF, JB, EA, BT, BG reviewed and approved the final version of the manuscript to be published.

Funding

The work was supported by an Fundação para a Ciênca e a Tecnologia scholarship funded by the Portuguese Ministry of Science, Technology and Higher Education [FCT grant: 2022.13804.BD].

Data availability

The data that support the findings of this study are available on the FIFA Data Platform, but restrictions apply to the availability of these data, which were used under license for the current study, and so are not publicly available. Data is available from the authors upon reasonable request and with permission of Portuguese Football Federation and FIFA. If someone wants to study further with the data, please contact the corresponding author.

Declarations

Ethics approval and consent to participate

The study involves FIFA World Cup Qatar 2022 matches to investigate the most demanding passages of kinematic activity, focusing on player profile and match-to-match variability for which authors have mentioned—The research project did not involve human participants, human experiments, and human data. The match data are permitted to make scientific research by the official FIFA Data Plaform. Authors confirm that all methods were carried out in accordance with relevant guidelines and regulations. This study was conducted according to the ethical principles of the World Medical Association Declaration of Helsinki and approved by Departamento de Desporto e Saúde, Escola de Saúde e Desenvolvimento Humano, Universidade de Évora.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.