Identifying Soccer Players’ Playing Styles: A Systematic Review

Identifying playing styles in football is highly valuable for achieving effective performance analysis. While there is extensive research on team styles, studies on individual player styles are still in their early stages. Thus, the aim of this systematic review was to provide a comprehensive overview of the existing literature on player styles and identify research areas required for further development, offering new directions for future research. Following the PRISMA guidelines for systematic reviews, we conducted a search using a specific strategy across four databases (PubMed, Scopus, Web of Science, and SPORTDiscus). Inclusion and exclusion criteria were applied to the initial search results, ultimately identifying twelve studies suitable for inclusion in this review. Through thematic analysis and qualitative evaluation of these studies, several key findings emerged: (a) a lack of a structured theoretical framework for player styles based on their positions within the team formation, (b) absence of studies investigating the influence of contextual variables on player styles, (c) methodological deficiencies observed in the reviewed studies, and (d) disparity in the objectives of sports science and data science studies. By identifying these gaps in the literature and presenting a structured framework for player styles (based on the compilation of all reported styles from the reviewed studies), this review aims to assist team stakeholders and provide guidance for future research endeavors.


Introduction
Performance analysis in football is not only a crucial tool for coaches [1,2], but also a subject of extensive research in recent years [3,4]. The advancement of technology has provided team analysts and researchers with a vast array of data [5]. While this facilitates the work of analysts and researchers, it also raises the question of expediting analysis without compromising its effectiveness [6].
Traditionally, performance analysis has emphasized the use of isolated performance indicators that provide valuable information to coaching staffs of teams [7]. However, in recent years, there has been a growing emphasis on playing styles among scientists, as they believe that playing styles can better explain football, particularly in terms of tactical aspects [8,9]. Playing style is generally defined as the characteristic playing patterns adopted by a team [10]. The playing styles of teams have been researched for a long time [11]. In contrast, research on the playing styles of individual players has emerged more recently, as previous years primarily focused on evaluating and rating players based on performance indicators [12][13][14]. However, the performance of soccer players is multifactorial in nature,

Reporting
The review was conducted based on the PRISMA guidelines [31][32][33]. It was not approved for registering because "PROSPERO does not accept reviews of sporting performance and the specific review does not include a direct health-related outcome". Ethical approval is not required because this review will use already-published data that are not related to either human or animal use [34][35][36].

Literature Search
Two authors (S.P. and C.K.) conducted independent searches on 21 April 2023, using four databases (PubMed, Scopus, Web of Science, and SPORTDiscus). In cases where discrepancies arose between the two authors, a third author (S.M.), who has experience in conducting systematic reviews, made the final decision regarding the inclusion or exclusion of the articles in question. The search strategy applied for all databases was as follows: (("soccer player") OR (footballer)) AND (("playing style") OR ("play style") OR ("game style") OR ("playing role")). There were no restrictions on the publication date of the studies.
The selection criteria for the articles involved two stages. In the first stage (screening), the titles and abstracts were considered, while in the second stage (eligibility), the full text of the articles was reviewed. The inclusion and exclusion criteria used during these stages are presented in Table 1. Furthermore, additional papers were identified through the references of the already-selected studies, using a technique known as the "snowball effect". Articles sampled were about developmental ages and not high-level professional soccer Articles that the data were taken from small-side-games and not from the regular 11 vs. 11 game.
Articles that dealt with the playing styles of teams rather than individual players.

Data Extraction
A Microsoft Excel spreadsheet was used for recording the retrieved data. Title, author name, publication year, aim, type and size of sample, method of data analysis, and results were recorded for each retrieved paper.

Quality Assessment
All of the included studies were evaluated using the answers to ten questions (Q1-10). For this purpose, a two-point scale was adopted (where "yes" = 1 point; "no" = 0). An 11th question (Q11) concerned the publication of the article in a journal or conference ( Table 2). The checklist developed ad hoc [37,38], taking into account questions that had been used in previous systematic reviews in the field of sports [37,39] and the aims of our review. The agreement between the two authors was measured by Cohen's kappa coefficient (k = 0.98). Furthermore, the points from all ten questions were added up for each study. The rating varied from 0 to 10 points. Quality assessment was not applied for the purpose of inclusion of the studies in the review. Finally, the averages were calculated for the studies that had been published in a journal and those that had been presented at conferences. Table 2. Questions for the quality assessment of the included studies.

Q1
The study objective(s) is/are clearly set out Q2 Relevance of background literature Q3 The characteristics of the sample are clearly defined (number of players, number of matches, number of observations) Q4 Variables apply to all phases of the game Q5 The reliability/validity of the data provider is stated, is mentioned, or is measured Q6 Certain contextual variables (e.g., match status, match location, type of competition, or quality of the opponent) are taken into account Q7 The results are clearly presented Q8 Specific player styles were recognized with clearly refined characteristics for each of them Q9 A distinction is made according to player positions Q10 Conclusion supported by results Q11 Journal/Conference

Search, Selection, and Inclusion of Publications
The initial search of the four databases yielded 646 titles. After removing duplicates, 391 articles were screened for relevance on the basis of their title and abstract. As a result, 29 full texts were assessed for eligibility; 19 of them were considered unsuitable for the review, but from the references of the 10 articles that were considered suitable, another 2 investigations were found. Therefore, a total of 12 articles were included in the review.

Quality Assessment
As can be seen from Table 3, one study was evaluated on the 10-point scale with 4, three studies with 5, two studies with 6, three with 7, one with 8, and two with 9. Four studies were published in journals, while eight in conferences. The average score for all studies was 6.33, for journal studies was 7.25, while for conference papers it was 5.88.

Quality Assessment
As can be seen from Table 3, one study was evaluated on the 10-point scale with 4, three studies with 5, two studies with 6, three with 7, one with 8, and two with 9. Four studies were published in journals, while eight in conferences. The average score for all studies was 6.33, for journal studies was 7.25, while for conference papers it was 5.88.  Table 4 presents the breakdown of the 12 studies, indicating that 9 of them employed artificial intelligence techniques such as machine learning and deep learning. The remaining 3 studies utilized alternative methods, including spatiotemporal qualitative calculus, statistics, and computational approaches. It is noteworthy that all the studies were published within the past 4 years, except for one study published in 2004. Figure 2 illustrates the distribution of studies per year, categorized based on their utilization of artificial intelligence.   Regarding data sources, websites and analysis platforms played a prominent role Specifically, the Whoscored website was utilized four times, the FBref website once, the Wyscout platform twice, and the Statsbomb platform once. Furthermore, GPS devices camera-based optical tracking data, Pappalardo's public dataset, and the electronic game "EA Sports FIFA" were each employed in one study. In one instance, the events were manually coded by the first author, while in another study, the data source was no explicitly mentioned.

Descriptive Analysis
The data employed in the studies encompassed various types. Tracking data were utilized in 3 studies, event stream data were utilized in 9 out of the 12 studies, 2 studies Regarding data sources, websites and analysis platforms played a prominent role. Specifically, the Whoscored website was utilized four times, the FBref website once, the Wyscout platform twice, and the Statsbomb platform once. Furthermore, GPS devices, camera-based optical tracking data, Pappalardo's public dataset, and the electronic game "EA Sports FIFA" were each employed in one study. In one instance, the events were manually coded by the first author, while in another study, the data source was not explicitly mentioned.
The data employed in the studies encompassed various types. Tracking data were utilized in 3 studies, event stream data were utilized in 9 out of the 12 studies, 2 studies reported using match sheet data, and 1 study relied on player evaluation attributes extracted from a video game. Among the studies, 8 had access to data from over 750 matches, 2 studies used a limited number of matches (1 and 22, respectively), and 2 studies did not specify the number of matches from which their data were collected.

Quality Assessment
The qualitative evaluation of the studies reveals that researchers have yet to address (0%) the impact of contextual variables (e.g., match status, match location, type of competition, or quality of the opponent) on players' playing style. It has been established that contextual variables influence team playing styles [48][49][50][51][52][53][54]. Therefore, it would be highly valuable to investigate how these variables also affect the style of individual players.
Moreover, deficiencies were identified in other aspects of research quality. Firstly, the mention or measurement of reliability/validity of the data provider was very low (25%) among the studies. In particular, in the studies of Beernaerts, De Baets [41] and Li, Zong [12], the validity is stated, while in the study of Taylor, Mellalieu [45], the validity is measured. Secondly, in three studies [40][41][42], their purpose was not clearly stated. Finally, in three studies [45][46][47], the characteristics of the sample (and the data) are not completely clear.
On the other hand, other quality criteria yielded high percentages. Specifically, all studies had relevant background literature, and the conclusions drawn were supported by the results. Additionally, only one study [21] lacked clear presentation of the results.

Theoretical Framework for Players' Playing Styles
According to Goes, Meerhoff [6], many computer science studies neglect to provide theoretical context to explain tactical behavior. This issue is also prevalent in most of the studies included in this review. Particularly, five studies [20,21,29,40,42] compare players based on vectors without describing any specific playing styles associated with different positions in the team formation. Similarly, two studies [41,47] that employed other computational methods did not provide descriptions of player styles. This does not necessarily imply that these investigations lack quality, but it suggests that they may have different objectives or purposes.
At the same time, two studies that utilized supervised machine learning methods employed pre-existing frameworks for categorizing players' playing styles. Specifically, Aalbers and Van Haaren [43] proposed a framework consisting of 21 player styles after conducting searches in sports media and the video game Football Manager 2018. On the other hand, Ghar, Patil [46] introduced a framework comprising 15 player styles but did not specify the source(s) they relied on for its development.
In contrast, three papers not only aimed to identify styles of play but also sought to describe the playing styles adopted by players in different positions within the team formation. Firstly, Taylor, Mellalieu [45] recognized variations in playing styles among players in the same position but could not generalize the results due to the small sample size of only 22 soccer players. Nonetheless, their study provided valuable insights for future researchers to explore distinct playing styles adopted by soccer players. Subsequently, Lee, Li [44] discovered 18 play styles, while Li, Zong [12] discovered 14 player styles (18 when including styles that were similar for both the right and left sides of the field). It is worth noting that the player styles identified in the latter two studies may be limited by their reliance solely on match event data. The inclusion of fitness performance indicators or more detailed tracking data could provide additional information on players' styles.
To address the gap in the existing literature, we synthesized all player styles mentioned in the 12 studies included in this review and developed a framework consisting of 26 player styles (33 when including styles that were similar for both sides of the field) ( Figure 3). Furthermore, Table 5 Figure 3.

Style Definition Examples
Sweeper $$$$keeper Is a goalkeeper who frequently comes off their goal line to act as an additional defender. They are known for their ability to read the game well, anticipate attacks, and make sweeping clearances outside of their penalty area [46,55].

Manuel Neuer, Ederson Moraes
Ball-playing $$$$keeper Is comfortable with the ball at their feet and is actively involved in the team's build-up play. They possess good passing, dribbling, and decision-making skills, often contributing to the team's possession-based style of play [43,44].

Sweeper keeper
Is a goalkeeper who frequently comes off their goal line to act as an additional defender. They are known for their ability to read the game well, anticipate attacks, and make sweeping clearances outside of their penalty area [46,55].

Manuel Neuer, Ederson Moraes
Ball-playing keeper Is comfortable with the ball at their feet and is actively involved in the team's build-up play. They possess good passing, dribbling, and decision-making skills, often contributing to the team's possession-based style of play [43,44].

Line keeper
Also known as shot-stopper or classic goalkeeper, focuses primarily on making saves and protecting their goal. They tend to stay closer to their goal line and excel in shot-stopping abilities, including reflexes, positioning, and aerial dominance [44,46].

Gianluigi Buffon, Keylor Navas
Ball-playing defender Is a central defender who excels in distributing the ball, initiating attacks from the back, and participating in the team's build-up play. They are comfortable on the ball, have good passing ability, and often contribute to the team's possession-based style of play [12,46].

Sergio Ramos, Gerard Piqué
Stopper Is a central defender who primarily focuses on defensive duties and is known for their physicality, strong tackling, and aerial prowess. They excel in winning duels, intercepting passes, and providing a solid defensive presence to protect the goal [12,46].

Kalidou Koulibaly, Giorgio Chiellini
Sweeper Or libero is a central defender who operates as the last line of defense, positioned behind the main defensive line. They have excellent reading of the game, anticipation skills, and are responsible for clearing any potential threats that breach the defensive line [56,57].

Franco Baresi, Traianos Dellas
Full back (defensive) Is a player positioned on the defensive line whose primary responsibilities include defensive duties, marking opposing wingers, and providing support in both defensive and offensive phases of play [12,46].

César Azpilicueta, Dani Carvajal
Wing back (offensive) Is a player who operates as a hybrid between a full back and a winger. They have defensive responsibilities but are also encouraged to push forward and provide width in the attacking third. Wing backs often contribute to the team's attacking play and are known for their pace and crossing ability [12,46].

Achraf Hakimi, Joshua Kimmich
Inverted wing back Is a player who operates as a full back or wing back but prefers to cut inside and contribute to the central areas of the pitch rather than hugging the touchline. They provide a different dynamic to the team's attack by creating overloads and offering passing options in central areas [58,59].

Style Definition Examples
Deep lying playmaker Is a midfielder who operates in a deeper position on the field, typically just in front of the defensive line. They are responsible for dictating the team's play from a deeper position, distributing accurate long-range passes, and initiating attacks from the defensive third [43,44].

Holding midfielder
Or anchor, is a player whose primary role is to provide defensive stability and shield the backline.
They excel in breaking up opposition attacks, intercepting passes, and disrupting the opponent's play in the midfield area [43,44].

Sergio Busquets, Fabinho
Ball winning midfielder Or destroyer or enforcer, is a player who specializes in regaining possession for their team by winning tackles, intercepting passes, and applying pressure on the opposition. They play a vital role in breaking down the opponent's attacks and disrupting their rhythm [43,46].

Casemiro, Kanté
Box to box Is a versatile player who covers a large area of the pitch, contributing both defensively and offensively. They are known for their stamina, work rate, and ability to make significant contributions in both defensive and attacking phases of play [43,46].

Jordan Henderson, Arturo Vidal
Advanced playmaker Is a central attacking midfielder who operates in an advanced position, typically just behind the forward(s). They are creative and influential playmakers who excel in their vision, passing ability, and ability to unlock defenses with through balls and key passes [43,46].

Kevin De Bruyne, Mesut Özil
Trequartista Is a central attacking midfielder who primarily focuses on creating goal-scoring opportunities and making incisive runs into the opposition's penalty area. They are typically the most advanced midfielder and are known for their dribbling skills, close control, and ability to score goals [44,60].

Lionel Messi, Paulo Dybala
False attacking midfielder Is a player who operates in a central position but often drops deeper into midfield or moves out wide to create space for other attackers. They aim to confuse the opposition's defense and create openings by dragging defenders out of position with their movement and positional versatility [44,61].

Wide midfielder
Or mezzala or half-winger is a central attacking midfielder who tends to move laterally, overloading the wings and creating outnumbers for his team. They show a preference for flank dribble, flank pass, flank long pass, and ball recovery in the mid-front area [12,62].

Bernardo Silva, Mathieu Valbuena
Inside forward Or inverted winger, is a wide midfielder who operates in a more central position, cutting inside towards goal rather than staying wide. They often aim to create goal-scoring opportunities by making diagonal runs, dribbling towards the center, and shooting from inside the box [12,63].

Style Definition Examples
Winger Is a wide midfielder whose primary role is to provide width and deliver crosses into the box from the flanks. They are known for their speed, dribbling ability, and ability to take on defenders in one-on-one situations. Wingers often look to create scoring opportunities for teammates with accurate crosses or by cutting inside to shoot themselves [12,44].

Wide playmaker
Is a creative midfielder who operates in wider areas of the pitch. They possess excellent vision, passing ability, and decision-making skills, and are responsible for creating goal-scoring opportunities by delivering accurate crosses, through balls, or making incisive passes from the flanks [64,65].

Franck Ribéry, David Beckham
Defensive wing forward Is a wide midfielder who combines offensive and defensive responsibilities. They not only contribute to the team's attacking play but also actively track back, press opponents, and provide defensive support to their team. They play a crucial role in maintaining defensive solidity while also providing an attacking threat [44,66].

Raheem Sterling, Dirk Kuyt
Poacher Is a central forward who specializes in capitalizing on goal-scoring opportunities inside the penalty box.
They have a natural ability to find space in the box, time their runs, and pounce on loose balls to score goals [12,46].

Robert Lewandowski, Erling Haaland
Mobile striker Is a forward who possesses pace, agility, and good movement off the ball. They are known for their ability to make intelligent runs, exploit spaces, and contribute to the build-up play by linking with teammates [12,44].
Pierre Aubameyang, Timo Werner Second striker Or "supporting " or "shadow" striker is a forward who operates just behind the main central striker.
They have creative abilities and excel in providing assists, making key passes, and scoring goals by arriving late into the box [12,44].

Target man
Is typically tall, strong, and adept at holding up the ball with their back to the goal. They often act as a focal point for the team's attacks, using their physical presence to win aerial duels and bring teammates into play [12,44].

Olivier Giroud, Edin Džeko
False nine Is a forward who operates in a deeper position than traditional central forwards. They drop deep into midfield to create space and disrupt the opposition's defensive structure. They are technically skilled, good at dribbling, and have excellent vision to create scoring opportunities for themselves and teammates [47,67].

Roberto Firmino, Antoine Griezmann
Although we have established a useful framework for future researchers and soccer practitioners, it is essential to acknowledge certain limitations. Firstly, the review's scope was quite narrow, as it solely focused on research articles published in scientific journals or conferences, excluding other valuable sources such as reviews and books. Additionally, the investigations were limited to 11 × 11 soccer matches, specifically involving adult professional male soccer players, omitting women's and development soccer. Nonetheless, these criteria were necessary to ensure the conclusions drawn align with the reality of the professional competitive level. On the other hand, while we conducted a qualitative evaluation of the included articles, it did not influence their inclusion in the review. Our decision was intentional, aiming to avoid further reducing the already limited number of studies and to encompass well-established player styles utilized in everyday soccer practice. Despite some articles merely mentioning player styles without in-depth explanations of their role on the field, we incorporated styles in developing the framework to ensure its comprehensiveness.

Conclusions
Through this review, we aimed to summarize the existing scientific knowledge regarding the playing styles of soccer players. Despite the limitations of our study, we have found that recognizing different player styles can have practical implications in various areas, including player scouting, player development, team selection and tactics, and opponent analysis. Additionally, we addressed a gap in the current literature by developing a comprehensive framework that encompasses all identified player styles for each position. This framework can be valuable for future researchers and soccer professionals, including coaches, scouts, analysts, and technical directors. Moreover, our qualitative assessment of the included studies revealed several methodological deficiencies, and we also observed a lack of research examining the influence of contextual variables on players' styles. Lastly, we emphasized the need for collaboration between sports science and data science to bridge the gap between theoretical knowledge and practical applications. Based on our findings and considering the absence of previous reviews in this rapidly growing field of research, we believe that this systematic review will open new avenues in soccer performance analysis.