Skip to main content
NCBI home page
European Journal of Sport Science logoLink to European Journal of Sport Science
. 2024 Sep 22;24(10):1526–1536. doi: 10.1002/ejsc.12177

Retirements of professional tennis players in ATP and WTA tour events

Lia Oliver 1, Ernest Baiget 2, Jordi Cortés 3, Joan Martínez 4, Miguel Crespo 5, Martí Casals 2,6,7,✉
PMCID: PMC11451576  PMID: 39308146

Abstract

Professional tennis competition at the highest levels requires high physical, technical, tactical, and mental demands. Player retirement is a scenario that often occurs at the highest echelons of the game. The objective of this study was to descriptively determine which factors influence retirement of matches in tennis. Matches from the Women's Tennis Association (WTA) and Association of Tennis Professionals (ATP) tours played over 44 and 46 years were included in the study, respectively. The results showed an increasing trend in the incidence of retirements in both ATP and WTA events, especially in recent years. Factors associated with the match characteristics, such as the type of surface, the type of tournament, and the round of the draw, were shown to influence retirement. Variables associated with player characteristics, such as the age difference between players, did not show to influence retirement, whereas the ranking difference did. When comparing ATP and WTA matches, similar results were observed in which surfaces and tournaments had the highest or lowest incidence of retirements. On the other hand, as per the rounds, the results are not conclusive. The findings highlight the growing trend of match retirements in professional tennis and emphasize the relevant influence of match characteristics such as surface type, tournament type, and round of the draw. These insights can guide coaches, players, and tournament organizers in developing strategies to mitigate retirements and inform future research on enhancing player longevity and performance in tennis.

Keywords: epidemiology, injury prevention, professional players, retirements, tennis

Highlights

  • An increasing trend in match retirements has been observed in both Association of Tennis Professionals (ATP) and Women's Tennis Association (WTA) events, especially in recent years, indicating a growing issue in professional tennis.

  • Factors related to match characteristics, such as playing surface, tournament type, and round of the draw, have been identified as key influencers of match retirements.

  • Player characteristics, specifically the age difference between players, do not significantly affect retirement, but the ranking difference does, highlighting its relevance.

  • Similar patterns were found in ATP and WTA matches regarding the impact of playing surfaces and tournaments on retirement rates, while the influence of match rounds remains inconclusive.

1. INTRODUCTION

Tennis is one of the most popular sports in the world in which the match duration is not defined by the rules of the game as it has no time limit. Male and female professional tennis matches involve high‐intensity intermittent efforts, including forceful strokes, fast accelerations, and decelerations and changes of directions, played across different indoor and outdoor surfaces (Pluim et al., 2023). Thus, matches can last for several hours, causing on the player a high‐level of physical and mental exhaustion that can end with injuries and/or retirement from the match (Dines et al., 2015). Moreover, professional tennis players face a continuous increase in competition density, resulting in increased physical, mental, technical, and tactical demands and consequently injury rates (Fu et al., 2018).

To face the high‐demanding competition loads required to participate in the ATP and WTA tours, depending on their rankings and their preseason commitments, professional players usually begin their training seasons around the last two months of the year (i.e., mid‐November to late December), with a general and/or specific preparatory phase that usually comprises a variable number of microcycles (Poignard et al., 2020) (lasting between 1–2 and 7 weeks). Then, players alternate between precompetitive and competitive phases with a training versus competition ratio which may be around 40%–60% (Poignard et al., 2020). Heavy training loads combined with short or incomplete preparation phases, together with a congested tournament schedule, can lead to diverse states of fatigue. The need for players to travel extensively and constantly adapt to different playing conditions exacerbates this issue. As a result, coaches and players must implement high‐quality techniques to prevent injuries and possible retirements (Poignard et al., 2020).

From a general sports perspective, injury and illness surveillance, as well as epidemiological studies, are fundamental elements of the concerted efforts to protect the health of athletes and should be the first step in any injury prevention program (Steven et al., 2013). Furthermore, long‐term injury surveillance enables possible implemented changes to be monitored over time, and when relating them to the development of the game, modifications in playing style, rule and/or tournament changes, or introduction of new equipment or playing surfaces (Bahr et al., 2020). To promote uniformity in definitions and methods and to facilitate the comparison of data from different studies, several consensus statements on sports injury epidemiology have been published. Most of these statements are designed to address specific sport disciplines. A tennis‐specific consensus statement was published in 2009, with a subsequent update in 2021. These documents were designed to enable standardized documentation and analysis of tennis‐related injuries by establishing consistent definitions, methods, and reporting procedures specific to tennis (Pluim et al., 2009; Wickham, 2016). The main point of difference between the two documents is the update and refinement of these standards in 2021 to reflect new research and advancements in the field. This manuscript adheres to the guidelines established in these consensus statements, ensuring that the data and analysis are consistent with widely accepted methods in the field. By following these established guidelines, the study provides reliable and comparable insights into the factors influencing player retirement in professional tennis.

In professional tennis, players may retire at any time after a match has started and they are unable to complete it due to injury or illness. Consequently, the match is awarded to the opponent regardless of its score. Retrospective studies have shown that retirements in professional tennis matches have increased in number and frequency over the past decade (McCurdie et al., 2017). In fact, as stated by Kryger et al., since the mid‐1980s, a 3‐fold increase in player retirements from ATP matches has been noticed, with an acceleration since 2002 (Okholm Kryger et al., 2015). This trend may have been influenced by several factors. One of them is the higher number of intense and explosive actions during the game due to the increased intensity of the matches (Gale‐Watts et al., 2016). In addition, the type of tournament and the round played could influence the likelihood of retirement. It can be hypothesized that retirements may be less frequent in later stages of professional tennis tournaments due to the higher stakes and prestige associated with advancing further, as well as increased player motivation to complete matches despite physical challenges (Hartwell et al., 2017; Maquirriain et al., 2016; Pluim et al., 2016). The players' ATP and WTA rankings could also influence the retirements from tennis matches. For example, research has shown that higher‐ranked players could be less likely to lose a match against lower‐ranked players due to a variety of factors, including better medical conditions and control, as well as the fact that age has been found a retirement risk factor (Rice et al., 2022).

The results of previous studies regarding the existence of differences in the incidence of retirements between male and female professional players are inconclusive. The analysis of withdrawal and retirement rates has shown mixed outcomes. Hartwell et al. (2017) examined USTA Pro Circuit tournaments and found that males are more likely to retire than females, highlighting a potential gender disparity in retirement incidence (Hartwell et al., 2017). Conversely, McCurdie et al. (2017) analyzed data from The Championships, Wimbledon, and found the opposite trend, with females exhibiting a higher likelihood of retirement compared to males (McCurdie et al., 2017). Other research has provided no significant differences between genders in terms of retirement rates. Pluim et al. (2006) reviewed tennis injuries and their prevention but did not identify relevant differences between male and female players concerning retirements (Pluim et al., 2006). Despite these studies, there remains a paucity of research specifically comparing WTA and ATP databases to elucidate gender‐specific trends in retirement rates. This gap indicates a need for more focused investigations to clarify the role of gender in retirement incidences among professional tennis players.

The aims of this study were to analyze the epidemiological patterns of retirements suffered by professional tennis players in ATP and WTA event matches during the period 1973–2019 and to relate the different match factors that influence retirement. We hypothesize that the incidence of player retirements has increased over the years, and we analyze the possible factors contributing to this trend in a large‐scale study with a comparison between ATP and WTA tour events.

2. MATERIALS AND METHODS

2.1. Study design

An observational, retrospective cohort study was conducted based on tennis matches from both WTA and ATP professional events.

3. DATA COLLECTION

The dataset on tennis matches was downloaded from GitHub (https://github.com/skoval/deuce/tree/master) or the R package “Deuce” (Stephanie Kovalchik, 2020). Both GitHub repository and Deuce package provide easy access to a rich set of data on professional tennis (WTA and ATP events). The information in these databases and package is taken from the http://www.tennisabstract.com/ and https://github.com/JeffSackmann websites (Sackmann, 2024a, 2024b). The Open Era of tennis began in 1968 when the Grand Slam tournaments agreed to allow professional players to compete against amateurs. After ATP was created, ATP rankings began in August 1973 (Association of Tennis Professionals, 2024). The WTA was founded in June 1973, and WTA rankings began in November 1975 (Women's Tennis Association, 2024). Therefore, the data used in this study include the year‐range among 1973–2019 for ATP and 1975–2019 for WTA. In the ATP tour, the tournament categories included were "Grand Slams," “Masters,” "Tour Finals," and "ATP250 or ATP500" were selected. In the WTA tour, following the same criteria, the tournaments selected were "Premier," "Premier Mandatory," and "Grand Slams”. Once the inclusion and exclusion criteria were applied, the ATP database contained 168,137 matches over 46 years, while WTA database had 46,452 matches over 44 years.

3.1. Variables

Seventy‐two variables were identified in both ATP and WTA original datasets (named “atp_matches.RData” and “wta_matches.RData”, found in https://github.com/skoval/deuce/tree/master/data). However, 48 variables were eliminated because they did not provide relevant information for the current study. Therefore, a total of 16 variables were considered in the final dataset. The outcome variable was Retirement with two possible categories: “yes” or “no”. Retirement was defined as “yes” when the match ended incomplete due to retirement of at least one of the two players involved in the match (Stephanie Kovalchik, 2020). Several covariables of interest that define match characteristics (the main objective of the study) were included in the dataset as follows:

  • best_of: maximum number of sets for the match. That number can be three or five in the ATP tour, whereas it is always three in the WTA tour.

  • Round: the round in which the match was played. The rounds were classified into 3 levels: (1) Qualifying rounds, including Q1, Q2, and Q3; (2) Preliminary rounds, including R128, R64, R32, R16, and Round Robin; and (3) Final rounds, comprising quarter final, semifinal (SF), and final rounds (F).

  • Tour: type of competition, being ATP for men and WTA for women.

  • Tournament_level: level of the tournament (tournaments included are specified in section Data Collection).

  • Surface: type of the court surface (i.e., carpet, clay, grass, or hard).

  • Games: total number of games played by the two opponents.

  • Year: year the match was played.

On the other side, covariables regarding players' characteristics were as follows:

  • Winner_hand, loser_hand: dominant arm for each player of the match.

  • Winner_age, loser_age: winner and loser player age in years at the time of the match.

  • Winner_rank, loser_rank: match players ranking at the time of the match.

  • Dif_age: age difference in years between winner and loser.

  • Dif_rank: ranking difference between winner and loser.

We investigated the annual number of matches played in both ATP and WTA tours to evaluate the potential influence of match load as a factor affecting the retirement outcome variable. The R workspace which contains the data frames used for the data analysis is available at: https://github.com/marticasals/Retirements_epidemiology_ATP_WTA_Tour.

3.2. Statistical analysis

In the descriptive analysis, absolute (n) and relative (%) frequencies were computed for categorical variables, while measures of central tendency and dispersion were calculated for the continuous ones. A bivariate analysis was performed to describe the characteristics of the match and the players when a retirement occurred. Unadjusted retirements incidence rates (IRs), which estimate the speed with which retirements develop by observing an athletic population over a specified period (Nielsen et al., 2019), were calculated by dividing new retirements by person‐time exposure. The consensus statement on epidemiological studies of medical conditions in tennis suggested that such exposure should be calculated using the duration of the time played (Nielsen et al., 2019). In this study, the IR was calculated as the number of retirements per 1000 played games as a surrogate of the game time. Played games was the sum of all games played during a match. The number of retirements, the exposure as number of games, the IR, and its 95% confidence interval (95% CI) were provided for each category of relevant variables. In addition, following the STROBE statement for observational studies (Vandenbroucke et al., 2007) and the CONSORT statement for randomized controlled trials (Moher et al., 2010), relative and absolute measures of association between covariates and the presence of retirements were given. They were expressed as incidence risk ratios (IRRs) (and their 95% CI) and risk differences (with their 95% CI). The incidence rate ratio (IRR) was estimated as the ratio of IRs between the two specified studied groups (i.e., male vs. female). Furthermore, with the aim to identify how many more retirements were sustained in one group as compared with another one, the absolute measure of risk difference was calculated by subtracting rates from the 2 exposures group (Nielsen et al., 2017). The incidence of retirements was plotted over the range of years included in this study, and it was compared between male (ATP) and female (WTA) tennis players.

At the multivariable level, a generalized linear Poisson model was fitted. The model expression for i‐th match is as follows:

logλi=loggi+α+βXi+ui

where Y i ∼ Poisson(λ i ); λ i is the expected number of retirements for the i‐th match; gi is the number of games of the i‐th match, which is the offset of this model; X i includes all independent variables of interest for the i‐th match; α is the intercept of the model; β represents the vector of coefficients associated to covariates; and u i is the error term.

Model selection was performed using a stepwise method based on the Bayesian information criterion. The fitted models were tested and used the test for overdispersion proposed by Kleiber and Zeileis (Kleiber, 2008). The interaction between tour types (ATP and WTA) and match characteristics was analyzed. This analysis was conducted both graphically, using interaction plots, and statistically, by including the interaction term in the model. Since relevant interactions were observed with most of the covariates included in the model, and interpreting results with interactions among more than two variables can be exceedingly complex (Harrell, 2012), Negative Binomial (before observing overdispersion) and Poisson regression models for ATP and WTA tours were separately fitted.

The incidence of retirements was explored as a function of year, revealing different behaviors in two groups (≤1990 and >1990 years). In the ATP model, the two slopes for both groups were considered. This approach allowed for a better understanding of the factors associated with retirements in each tour. Measures of association were calculated using the IRR with 95% CI. The significance level was set at α = 0.05.

All analyses were performed using version 4.1.3 of the R statistical software (R Core Team, 2013). The R package compareGroups (Subirana et al., 2014) was used to describe characteristics according to the presence of retirement. The epi.2by2 function of the R package epiR (Stevenson et al., 2024), setting method as cohort time, was mainly used to calculate the IRs. Most of the graphics were obtained using the ggplot2 (Wickham, 2016) package. The overdispersion was tested with the dispersion.test function in the R package AER (Kleiber, 2008). The ANOVA function from the package stats (Team RC, 2022) was used to compare between models. The R code is available at: https://github.com/marticasals/Retirements_epidemiology_ATP_WTA_Tour.

4. RESULTS

4.1. Descriptive characteristics of ATP and WTA matches

A total of 3539 (2.11%) retirements from 167,211 matches were found in the ATP tour matches, while 801 (1.73%) retirements occurred in 46,268 matches of the WTA tour ones (Table 1). Hard surface was by far the most frequently used surface for both tours. The ATP and WTA tours showed differences in terms of retirement percentages according to the tournament level: Female Grand Slams matches had the lowest frequency of retirements (1.01%) as opposed to the male ones, which had the highest rate (2.69%). While no relevant differences are apparent between rounds and percentage of retirements in the ATP tour matches, frequencies vary in WTA tour matches, ranging from 1.52% in the qualifying rounds to 2.13% in the final rounds.

TABLE 1.

Descriptive analysis of all variables according to the occurrence of retirement in Association of Tennis Professionals and Women's Tennis Association professional tours.

Retirement
ATP WTA

YES

N = 3539 (2.11%)

NO

N = 163672 (97.89%)

YES

N = 801 (1.73%)

NO

N = 45,467 (98.27%)
Tournament level ATP/WTA
ATP250 or ATP500/Premier 2235 (1.94%) 113,133 (98.06%) 360 (3.51%) 9833 (96.49%)
Grand Slams/Grand Slams 758 (2.69%) 27,471 (97.31%) 325 (1.01%) 31,876 (98.99%)
Masters/Premier Mandatory 544 (2.35%) 22,629 (97.65%) 116 (3.03%) 3708 (96.97%)
Tours finals/‐ 2 (0.45%) 439 (99.55%) ‐ ‐
Surface
Carpet 235 (1.36%) 17,095 (98.64%) 9 (2.89%) 302 (97.1%)
Clay 1258 (2.13%) 57,831 (97.87%) 179 (1.55%) 11,341 (98.45%)
Grass 365 (1.92%) 18,670 (98.08%) 107 (1.11%) 9531 (98.89%)
Hard 1681 (2.34%) 70,076 (97.66%) 506 (2.04%) 24,293 (97.96%)
Round
Final 445 (1.79%) 24,405 (98.21%) 55 (2.13%) 2532 (97.87%)
Preliminary 2581 (2.11%) 119,870 (97.89%) 500 (1.82%) 26,978 (98.18%)
Qualifying 513 (2.58%) 19,397 (97.42%) 246 (1.52%) 15,957 (98.48%)
Winner hand
Left 502 (2.01%) 24,517 (97.99%) 60 (1.73%) 3403 (98.27%)
Right 3007 (2.13%) 138,161 (97.87%) 648 (1.91%) 33,190 (98.09%)
Unknown 30 (2.93%) 994 (97.07%) 95 (1.06%) 8874 (98.94%)
Loser hand
Left 484 (2.07%) 22,868 (97.93%) 40 (1.34%) 2924 (98.66%)
Right 3023 (2.14%) 138,194 (97.86%) 653 (2.08%) 30,720 (97.92%)
Unknown 32 (1.21%) 2610 (98.79%) 108 (0.91%) 11,823 (99.09%)
Diff. age a 4.28 (3.25) 4.25 (3.30) 4.21 (3.39) 4.38 (3.38)
Diff. ranking a 84.4 (138) 95.6 (150) 61.6 (88.7) 70.2 (100)
Open in a new tab
a

Reported as mean (standard deviation).

4.2. Descriptive characteristics and match load of ATP and WTA players

The median age of the ATP and WTA players participating in the matches were 25.36 and 23.67 years, respectively. In both tours, most of the players were right‐handed (84.4% and 70.5%).

An important factor to consider when studying the etiology of retirements in tennis is the assessment of the external load to which players are exposed, specifically the match load is a key factor in the retirements. The match load in the ATP tour has a median of 3455 (Min: 2619 – Max: 4698) matches per year, while the median for the WTA tour is 732 (Min: 253 – Max: 2466) matches per year. Regarding the number of games played since the inception of both tours, in the ATP tour, this number increased on average 0.02 games per person‐year, and in the WTA tour, this increase was of 0.04 games played per person‐year.

4.3. Epidemiology measures of retirement

In the ATP tour, the “exposure time” to retirement was 4,058,882 games through the 46 years analyzed. The overall IR of retirement was 0.87 per 1000 games (95% CI: 0.84–0.90) played in the ATP tour. In the WTA tour, the “exposure time” to retirement was 980,504 games through the 44 years analyzed. The overall IR of retirement was 0.81 per 1000 games (95% CI: 0.7–0.88).

Figure 1 shows the IR per 1000 games over time stratified by ATP and WTA tours. The IR was mainly lower in women than in men until approximately 2011, when it was similar in both genders. From this year onward, a marked increase in the IR was observed in both tours, but it was much higher for women than for men. By 2017, while there was a tendency for men to stabilize and return to the usual pre‐2007 period levels, women continued to have a high IR until at least 2019, the final year data of this study.

FIGURE 1.

FIGURE 1

Open in a new tab

Retirement incidence over time per 1000 games for Women's Tennis Association (light blue) and Association of Tennis Professionals (blue) tours.

In the ATP tour (Table 2), the Masters event was the tournament level that presented a greater IR of retirement, equal to 1.03 (95% CI: 0.94–1.12), which was 25% higher compared to Grand Slams. Tournaments of both the ATP250 and ATP500 categories had an IR of retirement 5% greater than Grand Slams. Tour Finals were the tournaments with the lowest IR (0.19 retirements/1000 game), but it should be noted that the sample size of Tour Finals was quite low and only 2 retirements were observed, an aspect that should be considered when drawing conclusions. As per the risk difference, playing Masters involves 0.21 retirements/1000 games more than playing Grand Slams and 0.04 retirements/1000 games less than Grand Slams for ATP250 or ATP500 tournaments. Regarding surfaces, matches played on hard courts had the highest IR, with 0.96 retirements per 1000 games (CI 95%: 0.91–1.01), while those played on carpet were the ones showing the lowest IR, with 0.59 per 1000 games. This was approximately a 63% more than in carpet. When comparing matches played on grass with those on clay, clay court matches had an IR approximately 37% higher than those played on grass. As per the round matches, qualifying rounds were those with the highest IR in the ATP tour, with 1.15 (95% CI: 1.06–1.26) per 1000 games. The lowest IR was found on the final rounds with 0.75 retirements per 1000 games; on the contrary, the first rounds show an IR 53% higher than in the final rounds; the risk difference was 0.41 retirements/1000 games between them. No significant differences were found between preliminary and final rounds.

TABLE 2.

Association of Tennis Professionals and Women's Tennis Association tours incidence rates, incidence rate ratio (IRR), risk difference, and all their corresponding 95% confidence intervals (95% CIs) by games.

Tournament characteristics Retirements Games Incidence rate (95% IC) IRR (95% CI) Risk difference (95% CI)
ATP Tour
Tournament level
Grand Slams 758 927,785 0.82 (0.76, 0.88) 1 (Ref.) 0 (Ref.)
Masters 544 528,411 1.03 (0.94, 1.12) 1.26 (1.13, 1.41) 0.21 (0.11, 0.32)
ATP250 or ATP500 2235 2,592,110 0.86 (0.83, 0.90) 1.05 (0.97, 1.15) 0.04 (−0.02, 0.11)
Tours finals 2 10,576 0.19 (0.02, 0.68) 0.23 (0.03, 0.84) −0.63 (−0.90, −0.36)
Surface
Grass 365 544,654 0.67 (0.60, 0.74) 1 (Ref.) 0 (Ref.)
Clay 1258 1,371,036 0.92 (0.87, 0.97) 1.37 (1.22, 1.54) 0.25 (0.16, 0.33)
Hard 1681 1,744,963 0.96 (0.92, 1.01) 1.44 (1.28, 1.61) 0.29 (0.21, 0.38)
Carpet 235 398,229 0.59 (0.52, 0.67) 0.88 (0.74, 1.04) −0.08 (−0.18, 0.02)
Round
Qualifying 513 444,509 1.15 (1.06 1.26) 1 (Ref.) 0 (Ref.)
Final 445 596,269 0.75 (0.68, 0.82) 0.65 (0.57, 0.74) −0.41 (−0.53, −0.29)
Preliminary 2581 3,018,104 0.86 (0.82,0.89) 0.74 (0.67, 0.82) −0.30 (−0.40, −0.19)
WTA tour
Tournament level
Grand Slams 325 680,126 0.48 (0.43, 0.53) 1 (Ref.) 0 (Ref.)
Premier 360 218,308 1.65 (1.48, 1.83) 3.45 (2.96, 4.02) 1.17 (0.99, 1.35)
Premier Mandatory 116 82,070 1.41 (1.17, 1.70) 2.96 (2.37, 3.67) 0.94 (0.67, 1.20)
Surface
Grass 107 206,121 0.52 (0.43, 0.63) 1 (Ref.) 0 (Ref.)
Clay 179 242,569 0.74 (0.63, 0.85) 1.42 (1.11, 1.82) 0.22 (0.07, 0.36)
Hard 506 525,296 0.96 (0.88, 1.05) 1.86 (1.50, 2.31) 0.44 (0.31, 0.57)
Carpet 9 6518 1.38 (0.63, 2.62) 2.66 (1.18, 5.24) 0.86 (−0.05, 1.77)
Round
Qualifying 246 346,964 0.71 (0.62, 0.80) 1 (Ref.) 0 (Ref.)
Final 55 55,775 0.99 (0.74, 1.28) 1.39 (1.02, 1.87) 0.28 (0.00, 0.55)
Preliminary 500 577,765 0.87 (0.79, 0.94) 1.22 (1.05, 1.43) 0.16 (0.04, 0.27)
Open in a new tab

For the WTA tour matches (Table 2), the tournament level that presented a greater IR was Premier (1.65 and 95% CI: 1.48–1.83), followed by Premier Mandatory (1.41 retirements/1000 games and 95% CI: 1.17–1.70). That is approximately an IR 245% and 196% higher than in Grand Slams all in both sets and games, respectively. As per the risk difference, data on Premier matches showed 1.17 retirements/1000 games more than in Grand Slams and 0.94 retirements/1000 games more than in Premier Mandatory. Regarding surfaces, grass court matches showed the lowest IR, with 0.52 retirements per 1000 games. Matches played on carpet had an IR approximately 166% higher than those on grass. The difference risk showed 0.86 retirements/1000 games less on grass than on carpet. Nevertheless, it should be noted that for the latter, only 9 retirements were found. Matches played on hard courts were the second ones with higher IR (0.96 retirements/1000 games), 85% higher than on grass, and with a risk difference of 0.44 retirements/1000 games on grass courts. When comparing matches played on clay and grass, those on clay had an IR 42% higher than on grass with a risk difference equal to 0.22 retirements/1000 games.

4.4. Multivariable analysis of ATP and WTA matches

The multivariable results using the Poisson regression with the offset (games) for risk factors of retirement in ATP and WTA tour matches are shown in Table 3, respectively. The ATP model showed that playing on clay (IRR: 1.32 and 95% CI: 1.16–1.51) and hard (IRR: 1.25 and 95% CI: 1.11–1.43) courts increases the risk of retiring as compared to playing on grass courts. Furthermore, it was found that playing Masters (IRR: 1.15 and 95% CI: 1.01–1.30) and ATP250 or ATP500 (IRR: 1.16, and 95% CI: 1.05–1.28) events increases the risk of retiring as compared to Grand Slams and to tour Finals (IRR: 0.2 and 95% CI: 0.03–0.64).

TABLE 3.

Regression coefficients for the Association of Tennis Professionals and Women's Tennis Association models.

Variables Estimate SE IRR (95% CI) p‐value
ATP model
Surface
Grass ‐ ‐ 1
Clay 0.280 0.067 1.32 (1.16–1.51) <0.001
Hard 0.227 0.065 1.25 (1.11–1.43) <0.001
Carpet 0.115 0.095 1.12 (0.93–1.35) 0.227
Round
Qualifying ‐ ‐ 1
Final −0.037 0.077 0.96 (0.83–1.12) 0.626
Preliminary 0.109 0.062 1.12 (0.99–1.26) 0.080
Tournament level
Grand Slams 1
Masters 0.138 0.065 1.15 (1.01–1.30) 0.033
ATP250 or ATP500 0.150 0.051 1.16 (1.05–1.28) 0.003
Tour finals −1.580 0.709 0.21 (0.03–0.64) 0.026
Year
≤1990 0.004 0.008 1.004 (0.990–1.020) 0.577
>1990 0.015* 0.009* 1.019 (1.014–1.025) 0.083*
WTA model
Surface
Grass ‐ ‐ 1
Clay 0.018 0.014 1.02 (0.99–1.05) 0.192
Hard 0.020 0.013 1.02 (0.99–1.05) 0.115
Carpet 0.027 0.058 1.03 (0.92–1.15) 0.640
Round
Qualifying ‐ ‐ 1
Final −0.006 0.021 0.99 (0.95–1.04) 0.788
Preliminary 0.019 0.010 1.02 (0.99–1.04) 0.058
Tourney level
Grand Slams ‐ ‐ 1
Premier 0.029 0.013 1.03 (1.00–1.06) 0.028
Premier Mandatory 0.014 0.018 1.01 (0.98–1.05) 0.460
Year −0.001 0.0005 0.999 (0.998–1.00) 0.017
Open in a new tab

The coefficients of the simplified WTA model are shown in Table 3. The estimate coefficients showed that playing Preliminary (IRR: 1.02 and 95% CI: 0.99–1.04) and Premier round matches (IRR: 1.03 and 95% CI: 1.00–1.06) imply an increase in the risk of retirements as compared to Qualifying and Grand Slams.

5. DISCUSSION

A match retirement in tennis is defined as the scenario when a player is unable to continue playing a match or resuming a suspended match once it has started. This study aimed to describe the main match factors that influence professional player retirements and examine the epidemiological pattern of tennis retirement for WTA and ATP tour players. Matches from both tours played over 44 and 46 years (WTA and ATP, respectively) were included. The absolute and relative incidence at different levels of match factors, such as surface, round, and tourney level, were analyzed.

The global incidence of retirements has increased by approximately 50% in the ATP tour matches from 2006 to 2014, with a stabilization at pre‐2006 values as of 2015. In the WTA tour, the incidence has been increased sixfold the incidence from 2006 to 2012, with highly fluctuating values since 2006 but with no tendency to return to values prior to 2006. The observed increase in the incidence of retirements may be attributed to several factors including the rise in training and competitions loads witnessed over in the last decades (Fleming et al., 2023; Maquirriain et al., 2016; Okholm Kryger et al., 2015). With advancements in sports science and technology, athletes now undergo more rigorous training regimens and participate in a greater number of tournaments throughout the year, potentially leading to increased physical and mental fatigue. Additionally, the higher intensity of intermittent bouts of anaerobic exercise during matches, characterized by rapid and intense movements followed by brief periods of rest, can place significant strain on players' muscles and cardiovascular systems, predisposing them to fatigue‐related injuries and exhaustion (Gale‐Watts et al., 2016). These factors collectively contribute to a greater likelihood of retirements during matches, reflecting the evolving demands of professional tennis competition. This study has found an increase of 1.5 times in ATP and up to 2.8 times in WTA on the number of matches played. This increase in the matches load coincides in time with the increase in the incidence of retirements. Furthermore, since the increase in load is much higher in the WTA tour, it would also explain why the incidence of retirements in WTA increases to a greater extent than ATP.

As per the age of male and female players, it has been found that the median age of the players has been increasing over the years, from 23 years in ATP in 1985 to 27 years in 2019. In WTA, it has gone from approximately 21 years in 1985 to 25 years in 2019. This progressive increase highlights the maintenance of an almost constant difference of 2 years between WTA and ATP players. The age difference between men and women has been explained by differences in anatomic development as well as hormonal and physiological differences between sexes (Okholm Kryger et al., 2015). Nevertheless, in our study, the age difference between players does not significantly influence the retirement risk. Hence, the escalation of competitive burden, characterized by intensified training regimes, higher tournament frequency, and elevated expectations for performance emerges as a more precise explanation for the observed increase in retirement incidence.

The fact that tennis is played on a variety of court surfaces is unique as compared to other sports. On each of the surfaces, the bounce of the tennis ball is different which may cause a change in game style from the players, and therefore, may affect the retirement or injury incidence (Alexander et al., 2022). Moreover, the playing surface influences the external physical demands of the match, such as the force exerted on joints and muscles, the distance covered, and the intensity of movement. These factors subsequently affect the player's physiological responses during the match (Fernandez‐Fernandez et al., 2010). The results over the 44 and 46 years show that the playing surface influence retirement incidence in tennis. Specifically, grass courts are the ones with the lowest retirement IRs in both ATP and WTA matches, with hard courts showing the highest. On the same line, previous studies showed higher injury rates in hard courts compared to slow surfaces such as clay or grass courts (Pluim et al., 2017). In this sense, the grass surface is considered slow surface due to increased shock absorption and loss of ball speed, which might entail and entirely different set of stresses on the body (Fu et al., 2018) and the rally duration in grass courts has been noted shorter than hard and clay courts (Pluim et al., 2023). On the contrary, hard surfaces permit rapid changes in movement and direction and high rates of acceleration and deceleration becoming more likely to put more strain on muscles and tendons and increment injury and retirement incidences (Alexander et al., 2022).

The findings shown in Table 3 for the adjusted models suggest that the tournament level is related to the number of retired matches. Grand Slams tournaments are the ones with the lowest retirement incidence. Most of the previous studies that have analyzed a single type of tournament, such as the Davis Cup (Maquirriain et al., 2016), Wimbledon (McCurdie et al., 2017), and those that have included more than one type of tournament (Okholm Kryger et al., 2015), have not considered the type of tournament as a possible factor influencing the retirement by itself. The results of the present study suggest a strong positive relationship between the type of tournament and the incidence of retirement, especially in WTA tournaments, where in Premier and Premier Mandatory competitions both the absolute and relative incidence of retirement increased very significantly as compared to the Grand Slams (approximately 350% higher). According to the data used in this study, it is challenging to provide an explanation of the reasons why this occurs. It could be hypothesized that the Grand Slams, being the competitions that provide more ATP and WTA points and a greater prize money, in most cases represent the main season competitive goal of professional players. Consequently, players may exert greater effort to avoid and minimize the retirements in these tournaments, despite the potentially longer and more physically demanding matches, especially in the case of males who typically play best‐of‐five sets.

Few studies have analyzed other match factors that may influence retirement apart from the surface type. Those that have studied the effect of round type have found no significant differences, although only ATP matches were included (Breznik et al., 2012). Our results suggest that the round at which the match is played seems to influence retirement significantly and differently for ATP and WTA players. Qualifying rounds are those with the highest retirement incidence in ATP, whereas the opposite was found in WTA. A study on retirements and walkovers between 2013 and 2017 in Grand Slams pointed out that these occur in the first rounds (Lundy, 2017). In this study, it is suggested that the first round has the most qualifiers, who have already played matches just to make the main draw. As a result, they may be more drained or injured. Additionally, the first round tends to include more players on the fringe of making a living at professional tennis. Regarding ranking, previous studies have found that the retirement incidence increased proportionally with a bigger difference in both players' rankings (Breznik et al., 2012). Our results are in this line, the difference in ranking position between players significantly influence retirement, the bigger the difference, the higher the retirement risk. A recent study on second‐ and third‐tier tournaments in the ATP and WTA tours found similar trends, with surface type and match round influencing retirement rates, and noted an increase in retirements over the years (Palau et al., 2024). This reinforces our findings and suggests that these patterns are consistent across different levels of professional tennis competition.

This study has some limitations. Firstly, the incidences of retirements could have been calculated more accurately by having a minutes variable. However, the incidences have been calculated with two different denominators, sets, and games, which already indicate the exposure time of the players and which, moreover, with both denominators, the results are practically the same. Secondly, the WTA database does not contain as many data as the ATP does, which may cause the results to be somewhat biased. The variables that are most affected are age and ranking; therefore, although they have been included in the descriptive and regression analyses, these have not been considered in the analysis of incidence, since the match's own variables are the ones that have had the most weight in this study. Thirdly, due to the wide range of years covered by this study, some records may not be uniform over time. In these 44–46 years of registration, there have been changes in tennis regulations. Although the results may be altered by some factors changing over time, it was felt to be important to conduct an epidemiological study of tennis retirement of these dimensions and that, overall, these small biases have been mitigated. Fourthly, the lack of detailed analysis of retirements by specific tournaments or dates. The close scheduling of major events, such as the Roland‐Garros and Wimbledon, played on different surfaces within a short period, could potentially influence retirement rates due to the insufficient adaptation time. However, our data show fewer retirements on grass compared to clay and hard courts, suggesting that this effect may not be significant. Future research should explore the impact of tournament scheduling and surface transitions on player retirements to provide a more comprehensive understanding of these factors.

6. CONCLUSIONS

This study is the first to describe which factors affect retirement in both WTA and ATP tour matches through a period of 44 and 46 years, respectively. The study highlights the increased retirement incidence trend in both ATP and WTA tours. It demonstrates that the type of surface, the round, and the tournament affect the risk of retirement. When comparing ATP and WTA tour matches, similar results were observed as per the surfaces that have the highest or lowest incidences of retirements. It was also found that the Grand Slam tournaments have significantly less incidence of retirements in both databases as compared to the other events. As per the rounds, the results are also different between them. This descriptive analysis of retirements in female and male professional tennis and the identification of the factors that positively and negatively affect the retirement can help coaches, players, doctors, technical staff, or tournament organizing teams to deliver strategies to improve interventions aimed at preventing injuries that can assist in the reduction of retirements at the highest echelons of the game.

AUTHOR CONTRIBUTIONS

All authors wrote the article, critically read it. All authors have read and approved the final version of the manuscript and agree with the order of presentation of the authors.

CONFLICT OF INTEREST STATEMENT

The authors declare that they have no conflicts of interest.

ACKNOWLEDGMENT

This research was funded by the Ministerio de Ciencia e Innovación (Spain) (PID2019‐ 352 104830RB‐I00) and the Departament de Recerca i Universitats de la Generalitat de Catalunya (Spain) [2021 SGR 01421 (GRBIO)].

REFERENCES

  1. Alexander, Sneha , Naaz N., and Fernandes S.. 2022. “The Incidence of Injuries across Various Tennis Surfaces: A Systematic Review.” ITF Coaching & Sport Science Review 30(88): 39–44. 10.52383/itfcoaching.v30i88.353. [DOI] [Google Scholar]
  2. Association of Tennis Professionals . 2024. “ATP Tour [Internet].”: Available from: https://www.atptour.com/en/corporate/history. Accessed July 25, 2024.
  3. Bahr, Roald , Clarsen Ben, Derman Wayne, Dvorak Jiri, Emery Carolyn A., Finch Caroline F., Hägglund Martin, et al. 2020. “International Olympic Committee Consensus Statement: Methods for Recording and Reporting of Epidemiological Data on Injury and Illness in Sports 2020 (Including the STROBE Extension for Sports Injury and Illness Surveillance (STROBE‐SIIS)).” Orthop J Sport Med 8(2): 1–33. 10.1177/2325967120902908. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Breznik, K. , and Batagelj V.. 2012. “Retired Matches Among Male Professional Tennis Players.” Journal of Sports Science and Medicine 11(2): 270–278. [PMC free article] [PubMed] [Google Scholar]
  5. Dines, Joshua S. , Bedi Asheesh, Williams Phillip N., Dodson Christopher C., Ellenbecker Todd S., Altchek David W., Windler Gary, and Dines David M.. 2015. “Tennis Injuries: Epidemiology.” Journal of the American Academy of Orthopaedic Surgeons 23(3): 181–189. 10.5435/jaaos-d-13-00148. [DOI] [PubMed] [Google Scholar]
  6. Fernandez‐Fernandez, Jaime , Kinner Vanessa, and Ferrauti Alexander. 2010. “The Physiological Demands of Hitting and Running in Tennis on Different Surfaces.” The Journal of Strength & Conditioning Research 24(12): 3255–3264. 10.1519/jsc.0b013e3181e8745f. [DOI] [PubMed] [Google Scholar]
  7. Fleming, James Andrew , Field Adam, Lui Steve, Naughton Robert Joseph, and Harper Liam David. 2023. “The Demands of Training and Match‐Play on Elite and Highly Trained Junior Tennis Players: A Systematic Review.” International Journal of Sports Science & Coaching 18(4): 1365–1376. 10.1177/17479541221102556. [DOI] [Google Scholar]
  8. Fu, Michael C. , Ellenbecker Todd S., Renstrom Per A., Windler Gary S., and Dines David M.. 2018. “Epidemiology of Injuries in Tennis Players.” Curr Rev Musculoskelet Med 11(1): 7–11. 10.1007/s12178-018-9452-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Gale‐ Watts, Adam S. , and Nevill Alan M.. 2012. “From Endurance to Power Athletes: The Changing Shape of Successful Male Professional Tennis Players.” European Journal of Sport Science 16(8): 948–954. 10.1080/17461391.2016.1192690 [DOI] [PubMed] [Google Scholar]
  10. Harrell, F. E. 2012. “Regression Modeling Strategies.” R package version: 6–12. [Google Scholar]
  11. Hartwell, Matthew J. , Fong Sandra M., and Colvin Alexis C.. 2017. “Withdrawals and Retirements in Professional Tennis Players: An Analysis of 2013 United States Tennis Association Pro Circuit Tournaments.” Sport Health 9(2): 154–161. 10.1177/1941738116680335. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Kleiber, C. Z. A. 2008. Applied Econometrics with R. Springer. [Google Scholar]
  13. Lundy, A. 2017. “Failure to Complete: Numbers behind the Rate of Withdrawals on ATP Tour.”: Available from: https://www.espn.com/tennis/story/_/id/21586057/tennis‐failure‐complete‐numbers‐rate‐withdrawals‐retirements‐atp‐tour. Accessed July 25, 2024.
  14. Maquirriain, Javier , and Baglione Roberto. 2016. “Epidemiology of Tennis Injuries: An Eight‐Year Review of Davis Cup Retirements.” Eur J Sport Sci [Internet] 16(2): 266–270: Available from:. 10.1080/17461391.2015.1009493. [DOI] [PubMed] [Google Scholar]
  15. McCurdie, I. , Smith S., Bell P. H., and Batt M. E.. 2017. “Tennis Injury Data from the Championships, Wimbledon, from 2003 to 2012.” British Journal of Sports Medicine 51(7): 607–611. 10.1136/bjsports-2015-095552. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Moher, David , Hopewell Sally, Schulz Kenneth F., Montori Victor, Gøtzsche Peter C., Devereaux P. J., Elbourne Diana, Egger Matthias, and Altman Douglas G.. 2010. “Consolidated Standards of Reporting Trials Group. CONSORT 2010 Explanation and Elaboration: Updated Guidelines for Reporting Parallel Group Randomised Trials.” Journal of Clinical Epidemiology 63(8): e1–e37. 10.1016/j.jclinepi.2010.03.004. [DOI] [PubMed] [Google Scholar]
  17. Nielsen, Rasmus Oestergaard , Bertelsen Michael Lejbach, Verhagen Evert, Mansournia Mohammad Ali, Hulme Adam, Møller Merete, and Casals Martí. 2017. “When Is a Study Result Important for Athletes, Clinicians and Team Coaches/staff?” British Journal of Sports Medicine 51(20): 1454–1455. 10.1136/bjsports-2017-097759. [DOI] [PubMed] [Google Scholar]
  18. Nielsen, Rasmus Oestergaard , Debes‐Kristensen Katrin, Hulme Adam, Bertelsen Michael Lejbach, Møller Merete, Parner Erik Thorlund, and Mansournia Mohammad Ali. 2019. “Are Prevalence Measures Better Than Incidence Measures in Sports Injury Research?” British Journal of Sports Medicine 53(7): 396–397. 10.1136/bjsports-2017-098205. [DOI] [PubMed] [Google Scholar]
  19. Okholm Kryger, Katrine , Dor Frédéric, Guillaume Marion, Haida Amal, Noirez Philippe, Montalvan Bernard, and Toussaint J.‐François. 2015. “Medical Reasons behind Player Departures from Male and Female Professional Tennis Competitions.” The American Journal of Sports Medicine 43(1): 34–40. 10.1177/0363546514552996. [DOI] [PubMed] [Google Scholar]
  20. Palau, Maria , Baiget Ernest, Cortés Jordi, Martínez Joan, Crespo Miguel, and Casals Martí. 2024. “Retirements of Professional Tennis Players in Second‐And Third‐Tier Tournaments on the ATP and WTA Tours.” PLoS One 19(6): e0304638. 10.1371/journal.pone.0304638. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Pluim, Babette M. , Clarsen Benjamin, and Verhagen Evert. 2017. “Injury Rates in Recreational Tennis Players Do Not Differ between Different Playing Surfaces.” British Journal of Sports Medicine 3(9): 1–5. 10.1136/bjsports-2016-097050. [DOI] [PubMed] [Google Scholar]
  22. Pluim, Babette M. , Fuller Colin W., Batt Mark E., Chase Lisa, Hainline Brian, Miller Stuart, Montalvan Bernard, et al. 2009. “Consensus Statement on Epidemiological Studies of Medical Conditions in Tennis, April 2009.” Clinical Journal of Sport Medicine 19(6): 445–450. 10.1097/jsm.0b013e3181be35e5. [DOI] [PubMed] [Google Scholar]
  23. Pluim, Babette M. , Jansen Marleen G. T., Williamson Samuel, Berry Cain, Camporesi Silvia, Fagher Kristina, Heron Neil, et al. 2023. “Physical Demands of Tennis across the Different Court Surfaces, Performance Levels and Sexes: a Systematic Review with Meta‐Analysis.” Sports Medicine 53(4): 807–836. 10.1007/s40279-022-01807-8. [DOI] [PubMed] [Google Scholar]
  24. Pluim, B. M. , Loeffen F. G. J., Clarsen B., Bahr R., and Verhagen E. A. L. M.. 2016. “A One‐Season Prospective Study of Injuries and Illness in Elite Junior Tennis.” British Journal of Sports Medicine 50(7): 585–590. [DOI] [PubMed] [Google Scholar]
  25. Pluim, B. M. , Staal J. B., Windler G. E., and Jayanthi N.. 2006. “Tennis Injuries: Occurrence, Aetiology, and Prevention.” British Journal of Sports Medicine 40(5): 415–423. 10.1136/bjsm.2005.023184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Poignard, Mathilde , Guilhem Gaël, de Larochelambert Quentin, Montalvan Bernard, and Bieuzen François. 2020. “The Impact of Recovery Practices Adopted by Professional Tennis Players on Fatigue Markers According to Training Type Clusters.” Front. Sport Act Living 2: 109. 10.3389/fspor.2020.00109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. R Core Team . 2013. R: A Language and Environment for Statistical Computing. Vienna, Austria: R Foundation for Statistical Computing: Available from: http://www.r‐project.org/. [Google Scholar]
  28. Rice, Robyn Porter , Roach Kathryn, Kirk‐Sanchez Neva, Waltz Bret, Ellenbecker Todd S., Jayanthi Neeru, and Raya Michele. 2022. “Age and Gender Differences in Injuries and Risk Factors in Elite Junior and Professional Tennis Players.” Sport Health 14(4): 466–477. 10.1177/19417381211062834. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Sackmann, J. 2024a. “JeffSackmann/tennis_atp.”: Available from: https://github.com/JeffSackmann. Accessed July 25, 2024.
  30. Sackmann, J. 2024b. Tennisabstract [Internet].Available from: tennisabstract.com. Accessed July 25, 2024.
  31. Stephanie Kovalchik. Skoval/deuce: Resources for Analysis of Professional Tennis Data. 2020. Available from: https://github.com/skoval/deuce/
  32. Steven, Mark , Sergeant Evan, Nunes Telmo, Cord Heuer, and Marshall Jonathon, and Javier Sanchez R hornton . 2013. “EpiR: Tools for the Analysis of Epidemiological Data.”: Available from: https://fvas.unimelb.edu.au/research/groups/veterinary‐epidemiology‐melbourne.9–43. [Google Scholar]
  33. Stevenson, M. , SergeantE Heuer C., Nunes T., Heuer C., Marshall J., et al. 2024. “epiR: Tools for the Analysis of Epidemiological Data.”: Available from.https://cran.r‐project.org/web/packages/epiR/index.html.
  34. Subirana Isaac, Sanz Héctor, Vila, Joan . Building Bivariate Tables: ThecompareGroupsPackage forR. Journal of Statistical Software, 57, (12), 2014. Available from: 10.18637/jss.v057.i12 [DOI] [Google Scholar]
  35. Team RC . 2022. R: A Language and Environment for Statistical Computing. Vienna, Austria: Available from: https://www.r‐project.org/. [Google Scholar]
  36. Vandenbroucke, Jan P. , von Elm Erik, Altman Douglas G., Gøtzsche Peter C., Mulrow Cynthia D., Pocock Stuart J., Poole Charles, Schlesselman James J., and Egger Matthias. 2007. “Strengthening the Reporting of Observational Studies in Epidemiology (STROBE): Explanation and Elaboration.” Epidemiology 18(6): 805–835. 10.1097/ede.0b013e3181577511. [DOI] [PubMed] [Google Scholar]
  37. Wickham, Hadley . 2016. ggplot2: Elegant Graphics for Data Analysis [Internet]. New York: Springer‐Verlag: Available from: https://ggplot2.tidyverse.org. [Google Scholar]
  38. Women’s Tennis Association . 2024. “WTA Tour.”: Available from: https://www.wtatennis.com/about. Accessed July 25, 2024.

Articles from European Journal of Sport Science are provided here courtesy of Wiley

RESOURCES