More substitutions changed team substitution strategy? An analysis of the FIFA World Cup 2002–2022

Abstract

Background

The increase in the number of substitutions allowed in football from three to five has profoundly influenced the game. The impact of the rule change on the FIFA World Cup needs further verification.

Methods

A total of 2,151 team substitution opportunities and 2,410 substitutions in 384 matches from six FIFA World Cups (2002–2022) were analyzed to assess its impact. One-way ANOVA was used to assess differences in average substitution times among the six FIFA World Cups. Factors affecting the time of substitutions were explored using Generalized Linear Mixed Models.

Results

In each FIFA World Cup, over 92% of substitutions occurred during half-time and the second half, with a higher proportion in knockout stage matches than group stage matches. Group stage substitutions tended to happen earlier, particularly when teams were losing. As allowed substitutions increased, multiple substitutions in one opportunity rose from 4% to 38%. Of the 2,410 substitutions in the six FIFA World Cups, 21.45% were offensive, 63.65% were neutral and 14.90% were defensive. Winning teams made the highest percentage of defensive substitutions, while drawing or losing teams made the highest percentage of offensive substitutions. Match status significantly affected the time of the first three substitutions, and the presence of extra time significantly affected the time of the fourth substitution.

Conclusion

Analysis of substitutions in FIFA World Cups (2002–2022) reveals: Most substitutions occur during halftime and the second half; Substitutions are earlier in group stages, especially for losing teams; Increased allowed substitutions lead to more multiple-player substitutions; Defensive substitutions are more common when winning, while offensive ones are frequent when drawing or losing; Match status, ranking gap, extra time, game round, and substitution rules significantly influence the time of team substitutions.

Introduction

The FIFA World Cup is a widely popular global sporting event, attracting over 3.4 million spectators and an estimated 5 billion viewers worldwide during the 2022 tournament [1]. Due to the COVID-19 pandemic, the International Football Association Board (IFAB) implemented a temporary amendment in 2020, allowing teams to make up to five substitutes, with a limit of three substitution opportunities, excluding halftime and full-time substitutions [2]. This temporary measure was formalized into official regulation in 2022 [3], marking the 2022 FIFA World Cup the first to permit up to five substitutions during regular play.

There were two other notable rule changes. In 2016, an additional substitution during extra time was introduced and applied during the 2018 FIFA World Cup [4]. The second change, initiated by IFAB in December 2020, focused on concussion substitutions [5]. During the 2022 FIFA World Cup, each team was allowed one concussion substitution, which did not count towards the standard substitution limit [6]. Theoretically, each team in the 2022 FIFA World Cup could make up to six substitutions during regular time and up to seven substitutions in matches involving extra time.

Once a football match begins, the coach's ability to directly shape the game is limited, with substitutions being a primary tool for influence [7]. However, concrete evidence supporting the significance of substitutions in football matches is lacking [8]. Scholars have long been intrigued by the timing and strategy of substitutions in football. In the early twenty-first century, the Markov process model was used to investigate optimal substitution strategies for teams to maximize their points throughout a season [9, 10]. Key factors influencing the timing of substitutions include the current score and whether the team is playing at home or away [11]. Coaches tend to delay substitutions when their team is winning, while they make changes earlier when the team is drawing or losing [7].

Research has also indicated that making more substitutions does not necessarily provide a discernible advantage when considering team strength and game time. Furthermore, no consistent time frame in the second half shows a clear benefit from substitutions [12]. A survey of football professionals revealed that most believe substitute players play a crucial role in determining a team's success, as substitutions can significantly impact a match's outcome. However, some argue that hastily introduced substitute players might harm the game [13]. From a psychological perspective, the belief that substitutions can positively influence a match may be influenced by memory and confirmation biases. People tend to remember notable instances when substitutions led to crucial goals, equalizing or reversing the score, while many routine substitutions often go unnoticed [14].

During the era of three substitutions, the timing of substitutions varies across seasons, leagues, and tournaments. Substitutions are commonly made at half-time or during the second half of matches [15]. For example, in the 2004/2005 La Liga season, there were few first-half substitutions, followed by a significant increase at half-time, with a steady rise in substitutions from the 46th to the 70th minute, tapering off in the final 20 min [11]. In the same season, the timing of the first substitution showed the most significant variance among the Premier League, La Liga, and Serie A, while the second and third substitutions showed more minor differences [7]. A study on the Premier League revealed that substitutions occurred more frequently at half-time and between the 60th and 85th minutes [16]. During the 2014/2015 La Liga season, initial substitutions mostly happened between the 46th and 75th minutes, with second substitutions occurring between the 61st and 90th minutes and the third substitutions between the 76th and 90th minutes [17].

Teams tended to make more substitutions when losing compared to drawing or winning, and home teams tended to substitute players earlier than away teams [17]. In the 2013/2014 UEFA Champions League season, there were fewer first-half substitutions, a concentration of substitutions at half-time, an increase between the 57th and 78th minutes, and a decrease in substitutions in the final 10 min. The quality of the opponent and the game's specific situation were critical factors influencing the substitution decisions of top football coaches [18].

In recent years, the evolution of football tactics, an increase in the number of matches, and higher match frequency have raised demands on players, thus increasing the risk of injuries. Consequently, there have been growing calls to allow more substitutions to alleviate the players' workload [19]. Currently, the increase in substitutions provides coaches with greater flexibility and significantly impacts team performance. Several studies have indicated that more substitutions can reduce players' physical demands, thereby lowering the risk of injuries [20,21,22,23,24,25]. Additionally, research has shown that increased substitutions can enhance team running performance, particularly in metrics such as total distance covered and sprint frequency [26, 27].

Regarding the timing and frequency of substitutions, outcomes vary across different competitions due to recent rule implementation. For instance, in the 2019/2020 and 2020/2021 La Liga seasons, the increase in substitutions did not affect the timing of the first substitution, and not all coaches utilized all five substitutions. However, in the previous season, almost all coaches made use of all three substitutions [26]. In the Big Five leagues and the UEFA Champions League during the 2018/2019 and 2020/2021 seasons, the average timing of all substitutions in the Bundesliga, Serie A, La Liga, and Ligue 1 shifted earlier with the increased substitution opportunities [28].

Conversely, in the UEFA Champions League season that allowed five substitutions, the average timing of all substitutions was earlier compared to the previous three-substitution season [28]. Analysis of substitution data from the Bundesliga spanning six seasons from 2017/2018 to 2022/2023 revealed that coaches tended to replace more players with more substitutions, although not all coaches used all five substitutions. Following the rule change, while the timing of the first substitution did not significantly shift, the timing of the final substitution notably lagged [29]. It is important to note that the studies mentioned primarily analyze substitutions in club competitions, with limited research on substitutions in national team competitions.

After the 2022 Qatar FIFA World Cup, FIFA published an analytical article at the FIFA Training Centre. The article noted that substitutions during this FIFA World Cup occurred earlier, especially in the knockout stage, with a higher frequency between the 45th and 70th minute [30]. However, it is important to note that the article only presented substitution timing data for the 2018 and 2022 FIFA World Cups, omitting information about substitutions in previous tournaments. This study aims to statistically analyze substitution data from six FIFA World Cups from 2002 to 2022 to provide a more comprehensive and precise assessment of substitution timing and contribute to a more well-rounded and accurate conclusion.

Method

Samples

This study focuses on matches played in six FIFA World Cups held in 2002, 2006, 2010, 2014, 2018, and 2022, totaling 384 matches, including 288 group stage matches and 96 knockout stage matches. Across these FIFA World Cups, we documented 2,410 player substitutions involving 2,151 team substitution opportunities. Substitution data was collected by reviewing video footage from these tournaments. Any discrepancies between match reports and video data were resolved in favor of the video footage. We obtained technical reports from the 2002–2018 World Cups from the FIFA official website's Documents page (https://www.fifa.com/about-fifa/official-documents/). Match reports for the 2022 FIFA World Cup were sourced from the FIFA Training Centre website's Post Match Summary Reports page (https://www.fifatrainingcentre.com/en/fwc2022/post-match-summaries/post-match-summary-reports.php/). Since the 2002 FIFA World Cup, FIFA lists the names of the players of each participating team in the official technical report, including the players' positions and numbers. These reports also have detailed information about each match, including starters, substitutions, and goals scored. The above key information was not available in the FIFA World Cup technical reports before 2002. As the data for this study were derived from publicly available data, no ethical review was required.

Variables

The variables used in this study and the definitions of the variables are listed in the Table 1.

Table 1 The definition and abbreviation of variables

Statistical analysis

In this study, data were recorded using Microsoft Excel 2019 and analyzed with IBM SPSS Statistics 27. TPS data were grouped into 5-min intervals. Extra time substitutions were categorized into four intervals: 45 extra, 90 extra, 105 extra, and 120 extra. TTS data included the average time, standard deviation, and frequency of substitutions within each FIFA World Cup. Different match status were analyzed separately. One-way ANOVA was employed to assess differences in average substitution times among the six FIFA World Cups. For each FIFA World Cup, NTPS and NTUTSO were aggregated. The three NSOSO sub-variables were calculated and compared to the total NTUTSO within each FIFA World Cup, along with averages and standard deviations of NTPS and NTUTSO. One-way ANOVA was used to determine differences in average total player substitutions and total team substitution events among the six FIFA World Cups. ST data were collected and compared for each FIFA World Cup based on substitution frequency. Finally, a Generalized Linear Mixed Model (GLMM) was used to analyze the factors influencing the time of team substitution.

To analyze factors affecting substitution time, this study will employ GLMM. The independent variable of the model is TTS, which will be modeled and analyzed for the first, second, third, and fourth team substitutions, with 766, 747, 585, and 49 data respectively, totaling 2,147 entries. Due to the limited data (only four instances), factors affecting the fifth substitution opportunity were not analyzed using GLMM.

The model includes five dependent variables: ranking gap, game round, presence of extra time, substitution rules, and match status. The random variable in the model is the year of the FIFA World Cup.

In summary, the initial model framework is as follows:

$${Y}_{TTS}={\beta }_{0}+{\beta }_{1}{RG}_{ij}+{\beta }_{2}{GR}_{ij}+{\beta }_{3}{PET}_{ij}+{\beta }_{4}{SR}_{ij}+{\beta }_{5}{MS}_{ij}+{T}_{ij}+{\varepsilon }_{ij}$$

(1)

In the above formula, Y_TTS is the dependent variable (Time of Team Substitution), β₀ is the intercept, β₁ to β₅ are the independent variables, including the Ranking Gap, the Game Round, the Presence of Extra Time, the Substitution Rules, and the Match Status. T is the random variable (year of the FIFA World Cup), and ε is the random error.

Results

Time of player substitution

A total of 2,410 player substitutions were analyzed, with 1,800 substitutions in the group stage and 610 substitutions in the knockout stage. Figure 1 shows the distribution of player substitutions within each time interval during the group stage, represented as a ratio of the total group stage substitutions for each respective FIFA World Cup. Figure 2 shows the distribution of player substitutions within each time interval during the knockout stage, shown as a ratio of the total knockout stage substitutions for the corresponding FIFA World Cup. The results reveal both broad similarities and specific differences in player substitution timing across various FIFA World Cups.

Fig. 1

Ratio of substitutions to total substitutions by time period in the group stage of the FIFA World Cup 2002–2022

Fig. 2

Ratio of substitutions to total substitutions by time period in the knockout stage of the FIFA World Cup 2002–2022

Firstly, the majority of substitutions occur during halftime and the second half. In the group stage, over 92% of substitutions occur after the 46th minute, with the highest percentage recorded at 97.67% in 2022, followed by 97.03% in 2010. In the knockout stage, substitutions after the 46th minute exceed those in the group stage of the same tournaments, with the highest proportion observed at 99% in 2018 and 98.09% in 2022.

Secondly, during the group stage, after the peak of substitutions in the first 5 min of the second half, there is a minimal total substitution rate between the 51st and 55th minute. Substitutions gradually increase after the 56th minute, peaking between the 66th and 80th minutes. Substitutions decrease in the final 10 min, with the smallest decline observed in 2022.

Thirdly, in the knockout stage, mirroring the group stage, after the peak in substitutions within the first 5 min of the second half, there is a minimal total substitution rate between the 51st and 55th minute. In 2002, 2010, 2014, and 2018, substitutions begin to rise from the 56th minute, reaching a peak between the 66th and 85th minutes. In the last 5 min, there is a slight decrease in total substitutions, which is smaller compared to the group stage. In 2006 and 2022, total substitutions began increasing from the 61st minute, reaching their peak between the 71st and 75th minutes, followed by a slight decrease in the last 15 min.

Lastly, during extra time in the knockout stage, 2002 had the lowest total substitutions, while 2006 and 2010 had a proportion of substitutions below 10%. However, from 2014 onwards, substitutions during extra time account for more than 15% of the total knockout stage substitutions.

Time of team substitution

Table 2 presents the TTS in different match status during the FIFA World Cup group stage from 2002 to 2022. Table 3 provides insights into the TTS in various match status during the knockout stage of the same tournaments. The results show that TTS vary across FIFA World Cups in both the group and knockout stages, revealing specific patterns.

Table 2 Time of team substitutions in different match status in the group stage of the FIFA World Cup 2002–2022

Table 3 Time of team substitutions in different match status in the knockout stage of the FIFA World Cup 2002–2022

Firstly, in each FIFA World Cup, team substitutions in the group stage generally occur earlier, on average, than those in the knockout stage for each corresponding sequence.

Secondly, across all FIFA World Cups, the earliest average TTS occurs when a team is losing. In the group stage, except for the third TTS in the 2002 FIFA World Cup, substitutions when winning occur at a later average timing, while substitutions when drawing fall in between. In the knockout stage, the first TTS in each FIFA World Cup follows this pattern, with instances in subsequent sequences where substitutions when drawing have a later average timing compared to substitutions when winning.

Thirdly, there is a significant difference (P < 0.05) in the average timing of the first three substitutions in the group stage among the six FIFA World Cups. The earliest first and second TTS occurred in 2006, while the latest occurred in 2010. The earliest third TTS in the group stage was in 2002, while the latest was in 2018. In the knockout stage, the earliest first TTS was in 2002, and the latest in 2018. The earliest second TTS was in 2022 and the latest in 2014. The third TTS was latest in 2014 and earliest in 2022. The fourth TTS occurred earlier in 2022 compared to 2018.

Team substitution strategy

Table 4 displays data on NTPS, NTUTSO, NSOSO of teams across FIFA World Cups from 2002 to 2022. Table 5 provides data on NSOSO of teams under different match status during the same tournaments.

Table 4 NTPS, NTUTSO, NSOSO of teams of the FIFA World Cup 2002–2022

Table 5 NSOSO of teams under different match status of the FIFA World Cup 2002–2022

Before 2022, approximately 95% of substitutions involved replacing only one player. However, in the 2022 FIFA World Cup, this proportion dropped to around 62%. Historically, teams typically made a maximum of two substitutions in a single opportunity, with the group stage witnessing more of these than the knockout stage. In 2022, nearly one-third of substitutions involved replacing two substitute players simultaneously, and about 6% included three substitute players. However, in the knockout stage, simultaneous substitutions of three players were less common. Regarding match status, from 2002 to 2018, teams made more dual substitutions when losing compared to drawing or winning situations. In the 2022 FIFA World Cup, dual substitutions when winning were less frequent than in drawing or losing situations. The frequency of substituting two or three players simultaneously did not vary significantly across the three status.

Regarding substitution types, Table 6 shows that out of 2410 substitutions in six FIFA World Cups, the percentage of offensive substitutions was 21.45%, neutral substitutions were 63.65% and defensive substitutions were 14.90%. When teams were winning, there were more defensive substitutions, while in level-score situations, offensive substitutions slightly outnumbered defensive ones. When losing, there were significantly more offensive substitutions than defensive ones.

Table 6 Number of substation type of players under different match status of the FIFA World Cup 2002–2022

Factors affecting the time of team substitution

The results of the GLMM analysis are shown in Tables 7, 8, 9, and 10.

Table 7 Factors affecting the first TTS

Table 8 Factors affecting the second TTS

Table 9 Factors affecting the third TTS

Table 10 Factors affecting the fourth TTS

From Table 7, it is evident that both match status and ranking gap exert a significant influence on the first TTS. Match status have the most pronounced effect. In terms of coefficients, teams in a losing situation tend to make their first substitution earliest, while teams in a winning situation tend to delay their first substitution the most. Across all data from the six FIFA World Cups, the average time of the first TTS for the top-ranked team is slightly earlier than that for the bottom-ranked team.

Table 8 indicates that significant factors affecting the second TTS include match status, presence of extra time, game round, and ranking gap. Match status are the most significant influence. Regarding coefficients, teams in a losing situation typically have the earliest second TTS, while winning teams tend to delay their second substitution the most. Additionally, the second substitution occurs earlier in knockout stages without extra time.

Across different game rounds of the FIFA World Cup, the second TTS is earlier in the group stage compared to the knockout stage. The third round of the group stage having the earliest substitution time and the first round the latest. In the knockout stage, excluding the play-off for third place, the second TTS is earliest in the round of 16, with subsequent rounds witnessing progressively later second substitutions. In terms of ranking gap, teams higher in the rankings generally make their second substitution earlier than lower-ranked teams.

Table 9 indicates that significant factors influencing the third TTS include match status, presence of extra time, and substitution rules. Match status and presence of extra time are the most significant effects. Regarding coefficients, teams in a losing situation typically have the earliest third TTS, while winning teams tend to delay their third substitution the most.

In knockout matches without extra time, teams make their third substitution much earlier compared to both group stage matches and knockout matches with extra time. Under different substitution rules, such as allowing four substitutions in the 2018 FIFA World Cup, teams executed their third substitution earlier than under other circumstances.

As seen from Table 10, the presence of extra time is the sole factor significantly affecting the fourth TTS. In knockout matches without extra time, teams made their fourth substitution much earlier compared to group stage matches and knockout matches with extra time.

In summary, the factors influencing the timing of team substitutions varied across different substitution sequences. Match status and ranking gap significantly influenced the first TTS. The second TTS was notably influenced by match status, presence of extra time, game round, and ranking gap. Match status, presence of extra time, and substitution rules significantly influenced the third TTS. Finally, the presence of extra time was the significant factor affecting the fourth TTS.

Discussion

About the time of player substitution

In terms of substitution timing distribution, this study reveals a notable scarcity of substitutions during the first half, with the majority occurring at halftime and in the second half. This observation aligns with earlier findings [11, 15, 16, 32].

During the FIFA World Cup knockout stages, even fewer substitutions occur in the first half compared to the group stages. This indicates that coaches delay substitutions in these winner-takes-all matches where each substitution opportunity is precious. This trend of delaying substitutions has become increasingly prominent over the years.

Regarding specific time intervals for substitutions, it is important to note that different FIFA World Cups, match status, and match stages may yield varied results. Cultural factors such as history, traditions and values of different countries can also influence a team's game strategy, philosophy and playing style [33, 34]. Across six FIFA World Cups, the time interval with the highest number of substitutions showed variability during the group stage. In 2002, the 76–80 min range saw the most substitutions, closely followed by the 66–70 min range. In 2006, the peak of substitutions occurred in the 66–70 min range, followed by the 71–75 min range. In 2010, the 66–70 min range was the most active for substitutions, followed by the 76–80 min range. In 2014, the 76–80 min range witnessed the highest number of substitutions, followed by the 71–75 min range. In 2018, the 71–75 min range was the busiest for substitutions, followed by the 66–70 min range. In 2022, the 71–75 min range had the most substitutions, followed by the 76–80 min range. When expanding the time intervals to 15 min, during the group stage of all six FIFA World Cups, the peak for substitutions consistently falls between the 66–80 min mark.

In the knockout stages, the time interval with the most substitutions was the 76–80 min in 2002 and the 66–70 min in 2014. For 2006, 2010, 2018, and 2022, the most active time range for substitutions was the 71–75 min. According to a technical analysis article from the FIFA Training Centre, the 2022 FIFA World Cup featured a higher percentage of substitutions between the 45–70 min mark during the knockout stages compared to 2018 [30]. However, among the six FIFA World Cups, the year with the highest proportion of substitutions in this time range was 2006.

Turning to player substitutions during extra time in the knockout stages, the lowest proportion was observed in 2002. This was likely due to the Golden Goal rule, which often resulted in shorter extra periods. In such cases, coaches aimed to utilize all three substitutions before extra time. However, after the abolition of the Golden Goal rule, the number of player substitutions during extra time did not strongly correlate with the number of matches going into extra time.

In 2006, there were six matches in extra time, with substitutions accounting for nearly 10% of knockout stage substitutions. In 2010, with only four extra time matches, the proportion of substitutions stood at approximately 8%. The situation changed in 2014, with eight extra time matches and substitutions making up slightly over 18%. In 2018 and 2022, each featuring five extra time matches, the proportion of substitutions reached 18% and 16%, respectively. This evolving trend in the proportion of substitutions during extra time also supports the notion that coaches have increasingly delayed substitutions over time.

About the time of team substitution

This study on player substitution timing in the last six FIFA World Cups aligns with prior research findings. Notably, teams tend to substitute players earlier when they are losing compared to when they are drawing or winning [17]. Conversely, substitutions are typically delayed when a team holds a winning position [18].

Regarding specific timings, the average time for substitutions during the group stage in each FIFA World Cup tends to be earlier than corresponding substitutions in the knockout stage. This pattern mirrors the distribution of player substitution timings. Particularly noteworthy is the tendency for substitutions to occur later in level-score games, observed prominently in several instances: the third substitution in the group stage and knockout stage of the 2002 FIFA World Cup; the second and third substitutions in the knockout stage of the 2006 FIFA World Cup; the third substitution in the knockout stage of the 2010 FIFA World Cup; the second and third substitutions in the knockout stage of the 2014 FIFA World Cup; and the first to fourth substitutions in the knockout stage of the 2022 FIFA World Cup. This trend suggests that coaches may strategically delay substitutions to maintain field balance in knockout-stage matches with tied scores. They may wait until opposing players' stamina diminishes before introducing substitutes with superior running abilities, aiming to create more goal-scoring opportunities.

Multiple studies consistently demonstrate that substitutes cover more distance and engage in higher-intensity running after entering the field [8, 15, 16, 26, 27]. Additionally, substitutes show enhanced performance in technical aspects such as ball control, passing, shooting, and defending [35]. Analysis of player running data from the 2022 Qatar FIFA World Cup supports this, showing that substitutes cover more distance at higher intensities per minute [36]. Moreover, research indicates that substitutions made when scores are tied can significantly impact match outcomes [37]. These findings provide robust evidence supporting the strategic use of substitutions in football matches.

About team substitution strategy

This study represents the first attempt to quantify team substitutions based on actual substitution opportunities used by teams, addressing a significant limitation in previous research that focused on individual player perspectives. The challenge arose because previous methods couldn't distinguish the order of substitutions when teams replaced multiple players simultaneously in a single opportunity.

During the era of three substitutions, such issues were rare. Across the five FIFA World Cups from 2002 to 2018, fewer than 6% of substitutions involved replacing two players simultaneously. However, with the introduction of five substitutions, data solely from the 2022 FIFA World Cup reveals that nearly 40% of substitutions involved simultaneously substituting two or three players. Defining substitutions solely from the perspective of individual players under these circumstances makes it difficult to accurately record the order of substitutions, potentially compromising calculations of the average time between substitutions.

Previous studies, such as one analyzing data from the Big Five leagues and the UEFA Champions League, also relied on individual player substitution times for their analysis. They calculated average substitution times for the 2020/2021 season, considering both five substitutions in leagues like the Bundesliga, La Liga, Serie A, and Ligue 1, and 5 + 1 substitutions in the UEFA Champions League [28]. The accuracy of these times may require further validation.

This study's findings align with previous research on team substitution strategies. For example, regarding substitution types, neutral substitutions are frequently observed [8, 18, 31], while offensive substitutions tend to outnumber defensive substitutions overall [18]. In FIFA World Cup scenarios, defensive substitutions are more prevalent when a team is winning. Conversely, when a team is drawing, offensive substitutions are more common than defensive ones. In situations where a team is losing, offensive substitutions also outnumber defensive substitutions.

In addition, a team's substitution strategy may be influenced by the continent where the team is based. Research has shown that players from South America and Africa are more prone to committing errors compared to players from Europe, Asia, and North America. South American players tend to commit more fouls and consequently receive more yellow and red cards [38]. It is often observed in matches that when a key player in the team receives a yellow or red card, the coach tends to make quick decisions to make a substitution.

Regarding the number of substitutions made by teams in each substitution opportunity, as the allowed number of substitutions increased from three to five, the proportion of single-player substitutions in each FIFA World Cup decreased dramatically. In the five FIFA World Cups from 2002 to 2018, the maximum number of players substituted in a single opportunity was two, and this approach was more frequently employed in the group stage than in the knockout stage.

Furthermore, when a team was losing, coaches were more likely to substitute two players simultaneously. In the 2022 FIFA World Cup, the frequency of substituting two players in a single opportunity was similar across the three different match status. However, there were significant differences between the group stage and the knockout stage. In the group stage, the most frequent scenario for substituting two players at once was during a tied match. Conversely, in the knockout stage, the most frequent scenarios were when the team was winning or losing. Substituting three players in a single opportunity was more common in the group stage, whereas such extensive substitutions were rare during the knockout stage. This suggests that coaches tend to be more cautious and conservative in their substitution strategies during the knockout stage compared to the group stage.

About factors affecting the time of team substitution

This study found that the factors influencing team substitution times varied based on the order of the substitutions.

Match status is the most influential factor for the first three TTS. For the fourth and fifth TTS, match status is not significant, possibly due to the limited amount of data, with only a few dozen entries from the 2022 FIFA World Cup available for analysis.

The ranking gap significantly affects the first and second TTS, with higher-ranked teams making substitutions earlier. There may be several reasons for this. Firstly, the FIFA World Ranking calculation method is influenced by past performance and may not accurately reflect the team's current strength during the FIFA World Cup. Secondly, it is very common for the weaker team to beat the stronger team in FIFA World Cup, meaning the ranking gap between teams may not directly translate to on-field performance. Therefore, further research is needed to clarify the relationship between team ranking gap and substitution timing.

The game round is only related to the second TTS, though there is also a notable relationship with the third and fourth TTS during the group stage. This reflects a trend where teams become increasingly cautious with their substitutions as the FIFA World Cup progresses.

The presence of extra time significantly impacts the second and third TTS. In knockout stages, teams delay their substitutions if they anticipate a draw in regular time. Conversely, they may use substitutions earlier to prepare for extra time. The fourth TTS is significantly related only to the presence of extra time, likely due to the new substitution rules introduced in the knockout stage of the 2018 FIFA World Cup.

Substitution rules significantly affect only the third TTS. Under the three-substitution rule, the third substitution is the last, leading teams to use it cautiously, resulting in a later TTS. In contrast, under the four-substitution rule, teams may use the third substitution earlier, considering the additional substitution allowed in extra time.

Limitation

While the rule allowing five substitutions in football matches has been in place for nearly three years, it is important to acknowledge that team substitution strategies may vary across different periods, countries, leagues, competition levels, and various stages within the same competition.

This study focused solely on the FIFA World Cup, limiting the generalizability of findings to other football competitions. Additionally, this study did not consider the influence of external factors such as climatic differences between continents, stadium geography, weather conditions during the game, player injuries, players’ age, or team tactics on substitutions. The 2022 Qatar FIFA World Cup was the first ever held in the northern hemisphere during the winter months, and the impact of seasonal factors on substitutions has not been analyzed.

Therefore, broader research is needed to explore how the increased number of substitutions affects the timing and strategies employed by teams. This study does not investigate the potential influence of substitution patterns on match results. As the number of substitutions increases, examining how substitutions impact a team's overall performance becomes increasingly valuable.

Conclusion

Through the analysis of substitutions in the six FIFA World Cups from 2002 to 2022, the following findings can be summarized: 1) Substitutions primarily occurred during halftime and the second half; 2) Teams made substitutions earlier during the group stage matches compared to the knockout rounds; 3) Teams losing in a match made substitutions at earlier times; 4) An increase in the number of allowed substitutions led to a significant rise in teams substituting two or more players in a single substitution opportunity; 5) Defensive substitutions were more frequent when teams were winning, while offensive substitutions were more common when teams were drawing or losing; 6) Factors such as match status and ranking gap significantly influenced the time of the first substitution. Match status, presence of extra time, game round, and ranking gap affected the time of the second substitution. Match status, presence of extra time, and substitution rules impacted the time of the third substitution. The presence of extra time had a significant effect on the time of the fourth substitution.