Differences within Elite Female Tennis Players during an Incremental Field Test

Abstract

Purpose 

To compare technical and physiological responses between junior and professional female players during an incremental field test to exhaustion specific to tennis.

Methods 

Twenty-seven female players (n = 14 and 13 for juniors and professionals, respectively) completed an incremental field test to exhaustion specific to tennis, which consisted of hitting alternatively forehand and backhand strokes at increasing ball frequency (ball machine) every minute. Ball accuracy and ball velocity were determined by radar and video analysis for each stroke, in addition to cardiorespiratory responses (portable gas analyzer).

Results 

The stage corresponding to the second ventilatory threshold (+20.0%, P = 0.027), time to exhaustion (+18.9%, P = 0.002) and maximum oxygen uptake (+12.4%, P = 0.007) were higher in professionals than in juniors. The relative percentage of maximal HR was lower at both the first (−4.7%, P = 0.014) and the second (−1.3%, P = 0.018) ventilatory thresholds in professionals. Backhand ball velocity was the only technical parameter that displayed larger (+7.1%, P = 0.016) values in professionals.

Conclusions 

Compared with juniors, female professional tennis players possess higher exercise capacity, maximal and submaximal aerobic attributes along with faster backhand stroke velocities during an incremental field test specific to tennis.

In tennis, junior players transitioning to professional level must adapt to an increasingly complex environment to become a well-rounded athlete. Although previous studies on female players are mainly focused on physiological (¹), biomechanical (²), and/or more practical (³) aspects, there is paucity of studies that have identified what are the key areas for improvement to reach the professional level. There is a limited understanding of how to prepare female players for a smooth transition into the professional Women’s Tennis Association (WTA) tour. Hence, access to the top 100 is highly selective with a current mean age of 25.8 ± 4.9 yr, whereas the mean age of the top 100 WTA players only was 22.0 ± 4.0 yr, 10 yr ago (⁴). These reports highlight an increasingly competitive situation female players face at around 18 yr of age between the end of the Junior tour and the commencement of a sustainable professional career on the WTA circuit. In response to a 3-yr International Tennis Federation (ITF) Player Pathway review of professional and junior (between 2001 and 2013), the board of directors has recently approved the creation of a new global ITF transition tour for 2019 to provide a clearer and more effective structural organization of tennis governing bodies, in turn enabling a larger number of developing players to live sustainably from their professional tennis activities (⁵). Beyond this major restructuring, targeted propositions around key physical fitness and technical effectiveness components to be developed in priority for elite female players in their demanding tournament schedules (⁵) are largely missing. The present research is therefore conducted for providing comparative data between junior and professional players.

Recent findings from match-play tennis analysis highlight the clear need for sex-specific training and practice designs (^6,7). Beyond obvious differences in stroke production and physical qualities, it is crucial to obtain reference values in female players of different training background for talent identification or to tailor new skill acquisition to the developmental stage of the athlete (^8,9). In an environment that has become increasingly difficult to decrypt, analysis of technical and physiological characteristics under standardized conditions (unlike match play) at the different stages along the career of a female player may help building appropriate technical or physiological training adapted to this category of player.

To evaluate stroke production, rather simple ball accuracy (BA) indices have been used such as the ability to place the ball on a target located in different zones on the tennis court (^10–14). Hence, errors made by a player in relation to all shot performed in a field test (percentage of balls in the zone) are a good predictor of successful groundstrokes performance (¹⁴). For instance, almost perfect correlations between the national rank position and the ranking of stroke ratings were reported for the Smekal’s field test (¹²). However, this analysis was based on 12 males with large national ranking differences [top 30 to over 200]. In the original version of the Smekal’s field test (¹²), players began hitting balls at a ball frequency (BF) of 12 shots per minute, with subsequent increment stages of 2 shots per minute every 3 min, which in turn probably limit the involvement in stroke production and emphasis accuracy. As movement velocity depends on the accuracy requirements of the task, an increased on-court movement velocity (i.e., players under time pressure) would inevitably lead to decreased motor accuracy (^8,9). Interestingly, Rota et al. (¹⁵) illustrated a management of the speed-accuracy trade-off in favor of the ball velocity (BV) of serve and forehand drives in response to intense intermittent exercise. This implies that stroke production must be determined from simultaneous measurements of BV and BA with increasing fatigue levels. It is currently unclear if female players’ strategy would differ according to their training background to prioritize either BA or BV when physically exhausted.

We have recently developed an incremental (by adjusting BF) test to exhaustion specific to tennis (TEST) that consists of hitting alternatively forehands and backhands into target areas (¹⁶). Test to exhaustion specific to tennis simulates some of the components of actual match play (i.e., use of the tennis court dimensions, combination of specific footwork and hitting actions) and uses a reliable ball throwing machine (¹⁷), allowing standardized conditions for testing highly skilled players. Previously, we reported that technical alterations (i.e., decrease in BV, BA, and technical performance [TP]) during TEST only occurred at high intensity (>80% of maximum oxygen uptake [V˙O_2max]) in young (~18 yr old), elite male players. To date, no study reported specifically what could be the main physiological and technical differences observed between elite junior and professional female players, under standardized conditions, as encountered during TEST.

The aim of the present study was to compare technical and physiological parameters, for forehand and backhand strokes combined and separately, during TEST completion by elite junior (JUN) and professional (PRO) female tennis players. Although technical alterations were expected above typical competition intensities (i.e., above 80% V˙O_2max) in both groups, we tested the following hypotheses: (i) stroke performance (overall BV and BA) and aerobic fitness (sub-maximal and maximal indices) are lower, whereas (ii) technical alterations would occur earlier (i.e., at a lower relative intensity) in JUN compared with PRO. A second aim was to test the nature of relationships between players’ rank order and technical/physical indices.

METHODS

Ethic Statement

Both the players and their parents (for minors) provided written informed consent for the study after the procedures and potential risks associated with participation in the study were fully explained. The scientific committee of the French Tennis Federation approved the study that was performed in accordance with the ethical standards reported (¹⁸) and conformed to the recommendations of the Declaration of Helsinki.

Participants

Twenty-seven elite competitive female tennis players (mean ± SD: age, 16.7 ± 3.1 yr; stature, 169.1 ± 5.5 cm; body mass, 59.7 ± 5.3 kg) volunteered to participate in the study. Fourteen players (age, 14.7 ± 1.0 yr; stature, 168.9 ± 2.9 cm; body mass, 57.7 ± 5.0 cm) had an ITF junior ranking (229 ± 216; range, 52–770), whereas 13 players (age, 18.9 ± 3.2 yr; stature, 168.9 ± 7.4 cm; body mass, 61.8 ± 4.9 cm) were holding a WTA (665 ± 351; range, 132–1211) ranking. All rankings were established on the day of TEST execution. All current participants are or were members of the national teams of the French Tennis Federation (International Tennis Number 1 [elite]). During the 3 months before testing (November), they participated regularly to official tennis competitions (i.e., “ITF Juniors,” and “ITF Futures” or WTA tournaments) for a total of 5 to 10 matches monthly.

Players’ Rank Order

The international tennis ranking (ITF junior and WTA) was used to rank players from 1 to 27 in our population sample following ITF junior and WTA positions.

Experimental Design

All participants performed a newly developed incremental test protocol for the first time; the so-called Test to Exhaustion Specific to Tennis (⁶). For all players, TEST was conducted under similar standard environmental conditions (temperature ~20°C, relative humidity ~50%) on an indoor tennis court (i.e., GreenSet® surface, GreenSet Worldwide S.L., Barcelona, Spain). All participants were given written and verbal instructions to report for testing in a well-rested, well-hydrated and well-nourished state, and to refrain from eating at least 2 h before testing. They were told to refrain from strenuous training and to maintain their usual nutritional and hydration habits the day before the test. One week before the main experimental trial, a familiarization session was scheduled allowing TEST requirements to be explained.

Experimental Procedures

Test to exhaustion specific to tennis

The detailed description of TEST has been published elsewhere (^14,16). Briefly, TEST procedure consists of hitting balls thrown at constant velocity (mean, 86 km·h⁻¹; coefficient of variation for BV = 1.7% and 1.5% for right and left corners of the baseline, respectively), alternating forehand and backhand strokes, by a “Hightof” ball machine. Players had to hit balls cross-court in a prescribed pattern (i.e., topspin drive), whereas the landing point for thrown balls was set 3 m in front of baseline. Slice strokes were not allowed because of their potential influence on ball positioning and therefore on TEST performance and associated physiological responses.

After a standardized warm-up (i.e., a 2-min “habituation” phase where a BF of 16 shots per minute with minimal lateral displacement) and 1 min of passive rest (quiet standing), the main test procedure begun: a BF of 10 shots per minute was first selected, which was then increased by 2 shots per minute every minute until the stage corresponding to a BF of 22 shots per minute. From there, increment in BF was set at +1 shots per minute until exhaustion. After each 1-min stage, a 30-s passive recovery break (quiet standing) was implemented.

Players were asked to perform TEST as closely to what they would do during official competitions. They were told to “hit the ball with the best possible velocity/accuracy ratio.” Participants had visual of the areas in which they were aiming at. Stroke involvement was motivated by “live” (immediate) feedback. Based on previous results (¹³), reporting range of BV ball velocities were between 86 and 120 km·h⁻¹ (submaximum to maximum strokes), BV < 80 km·h⁻¹ was chosen as the criteria for unsuccessful BV, and BA (30% of balls landing outside the target zone) at the end of each stage was completed.

The TEST ended with players’ voluntary felt exhausted or failed to reach and hit the ball twice in a row or when the players were no longer able to perform strokes with an acceptable execution technique and a demise in BV/BA. Performance was measured as the time to exhaustion.

Physiological Measurements

Expired air was analyzed continuously (breath-by-breath measurements) for oxygen consumption (V˙O₂) using a portable gaz analyser (Metamax II CPX system; Cortex®, Leipzig, Germany). Gas and volume calibration of the measurement device were performed before each test according to manufacturer’ instructions. HR was recorded continuously (Suunto Ambit2®, Vantaa, Finland). Furthermore, 25-μL capillary blood samples were taken from fingertip and analyzed for blood lactate concentrations (LT-1710; Arkray®, Japan) at baseline, during TEST (i.e., during the 30-s recovery periods after every stage until a values of 4 mmol·L⁻¹ was obtained and thereafter every two stages) and 15 s after exhaustion.

V˙O_2max was determined by the observation of a “plateau” or leveling off in V˙O₂ or when the increase in two successive periods was less than 150 mL·min⁻¹ (¹⁹). Maximal HR (HR_max) was considered as the highest value reached during the final minute of the test. For final analyses, we only considered values from 60% to 100% of V˙O_2max, with step increases of 5%.

Detection of ventilatory thresholds (VT) was done by analyzing the points of change in slope (breakdown in linearity) of ventilatory parameters (^19,20). The first ventilatory threshold (VT1) was determined using the criteria of an increase in the ventilatory equivalent for oxygen (V˙_E/V˙O₂) with no increase in the ventilatory equivalent for carbon dioxide (V˙_E/V˙CO₂) and departure from the linearity of V˙_E caused by a more rapid increase in ventilation. VT2 corresponded to an increase in both V˙_E/V˙O₂ and V˙_E/V˙CO₂. All VT assessments were made by visual inspection of graphs of time plotted against each relevant respiratory variable measured during testing. All visual inspections were carried out by two experienced exercise physiologists. The results were then compared and averaged. The difference in the individual determinations of VT2 was <3%.

The maximal aerobic frequency (MAF) is considered as the last stage maintained with a plateau of V˙O_2max. It differs with maximal work rate and last stage (²¹).

Evaluation of Groundstroke Performance

During TEST, groundstroke production was assessed by the mean of two “primary” variables: BV and BA. Ball velocity (km·h⁻¹) was measured with a Solstice 2 radar (Hightof’^®, France). BA (%) was defined as the percentage of correct hits in the defined zones (³). For each stage, BV and BA data were averaged (BV_mean and BA_mean, respectively) and expressed for forehands (BV_f and BA_f, respectively) and backhands (BV_b and BA_b, respectively), separately. Finally, because BV and BA in combination better reflect the overall stroke effectiveness in tennis, an index was calculated for forehands and backhands separately (TP_f and TP_b) and averaged (TP_mean) as the product of these two other technical variables (TP = BV × BA).

Statistical Analysis

Mean (±SD) was calculated for all variables. Mean difference for change between groups (in %), and 95% confidence interval (95% CI) were reported when appropriate. A one-way ANOVA was used to compare absolute values between JUN and PRO. For each ANOVA, effect size (ES) was calculated (Cohen d) with the following criteria: an ES of <0.2 is classified as a trivial, 0.2 to 0.4 as a small, 0.5 to 0.7 as a moderate and >0.8 as a large effect. The BV, BA, and TP data for each groundstroke (separately and combined) were compared using a two-way repeated measures analysis of variance (group: [JUN vs PRO] × time [60%, 65%, …, 100% of V˙O_2max]). However, when the normality test failed, a Mann–Whitney rank sum test was performed at each time interval. Multiple comparisons were made with the Tukey post hoc test. Finally, Pearson rank order correlations were used to test associations (for forehands and backhands separately) of players’ ranking with technical and physiological parameters. The following criteria were adopted to interpret the magnitude of r: < 0.1, trivial; 0.1 to 0.3, small; 0.3 to 0.5, moderate; 0.5 to 0.7, large; 0.7 to 0.9, very large; and 0.9 to 1.0, almost perfect (²²). Statistical significance was accepted at P < 0.05. The statistical analyses were performed using SigmaStat 3.5 software.

RESULTS

TEST performance and physiological parameters

TEST duration was longer (+18.9%; 95% CI, 8%–30%; P = 0.002) and, at exhaustion, V˙O_2max was also higher (+12.4%; 95% CI, 4%–23%; P = 0.007) in PRO versus JUN. % HR_max values at VT1 and VT2 were lower in PRO (−4.7%; 95% CI, 1%–8%; P = 0.014 and −1.3%; 95% CI, 0%–2%; P = 0.018, respectively), whereas HR_max did not differ between groups. Compared with JUN, V˙_E (between 80% and 100% of V˙O_2max; range, 8.4%–27.5%) and V˙O₂ (between 90% and 100% of V˙O_2max; range, 5.3%–5.8%) values were higher in PRO (Fig. 1B and D). HR was higher (+2.1%–8.6%) between 60% and 90% of V˙O_2max in JUN versus PRO (Fig. 1A).

FIGURE 1: Changes in HR (A); ventilation (V˙ _E, B); blood lactate concentration (Lactate, C) and oxygen uptake (V˙O₂, D) as a function of exercise intensity (% of V˙O_2max) for professional (dotted line) and junior (continuous line) female players. Vertical arrow indicates the second ventilatory threshold. *P < 0.05, **P < 0.01, ***P < 0.001, significantly different between the two groups. T, G, and I respectively refer to ANOVA main effects of time, condition and interaction between these two factors with P value and ES in parentheses. T, C, and I, respectively, refer to ANOVA main effects of time, condition and interaction between these two factors with P value and ES in parentheses. #P < 0.05, significantly different vs value at 100% of V˙O_2max.

Groundstroke performance

PRO had significantly higher BA_mean values at 70% (+19.6%) and 75% (+29.1%) of V˙O_2max, but reached lower (−27.7%) values at 100% of V˙O_2max (Fig. 2B). Compared with JUN, TP was higher [+6.1%–34.3%] from 60% to 90% of V˙O_2max in PRO (Fig. 2C). Overall, BV_b (+7.2%; 95% CI, −3% to 12%; P = 0.016) and TP_b (+21.1%; 95% CI: 0%–42%, P = 0.049) were higher in PRO than in JUN, independent of time (Table 1). Compared with JUN, BA_f (+21.9%; 95% CI, −7% to 39%; P = 0.042) and TP_f (+21.9%, 95% CI: −3% to 29%) were higher for PRO at 70% of V˙O_2max, but lower in BA_f at 100% of V˙O_2max (−37.5%; 95% CI, −58% to 5%; P = 0.016 (Fig. 3C and E). No significant differences appeared in JUN for both TP_mean and BA_mean (Fig. 2B and C). Post hoc also revealed that BV_mean were significantly higher for JUN when we compared values from 60% to 85% of V˙O_2max (0.015 ≤ P ≤ 0.05) versus 100% of V˙O_2max, whereas significant differences appeared in PRO from 80% to 90% of V˙O_2max (0.007 ≤ P ≤ 0.046) (Fig. 2A).

TABLE 1: Physiological variables (A) and technical parameters (B) during TEST in Junior and Professional female players.

FIGURE 2: Changes in BV (BV_mean, A), BA (BA_mean, B), and TP (TP_mean, C) for forehand and backhand stroked combined as a function of exercise intensity (% of V˙O_2max) for professional (dotted line) and junior (continuous line) female players. Vertical arrow indicates the second ventilatory threshold. * P < 0.05, **P < 0.01, significantly different between the two groups (at specific exercise intensities). T, G, and I, respectively, refer to ANOVA main effects of time, group and interaction between these two factors with P value and ES in parentheses. #P < 0.05, significantly different vs value at 100% of V˙O_2max.

FIGURE 3: Changes in BV_f (A); BV_b (B); BA_f (C); BA_b (D); TP_f (E); TP in TP_b (F), as a function of exercise intensity (% of V˙O_2max) for the professional (dotted line) and junior (continuous line) female players. Vertical arrow indicates the second ventilatory threshold. *P < 0.05, significantly different between the two groups. T, G, and I, respectively, refer to ANOVA main effects of time, condition and interaction between these two factors with P value and ES in parentheses. #P < 0.05, significantly different vs value at 100% of V˙O_2max.

Relationships between variables

Table 2 shows Pearson correlations between players ranking and physiological and technical responses to TEST. There were negative correlations between players’ ranking and BV_b (r = −0.485, P = 0.01), BA_b (r = −0.455, P = 0.017), TP_b (r = −0.465, P = 0.015), and BV_mean (r = −0.446; P = 0.02) (Table 2).

TABLE 2: Correlation coefficients between players’ rank order (both groups combined) and the various physiological (A) and technical (B) parameters.

DISCUSSION

We compared physiological responses and stroke production (BV, BA, and TP) for forehands and backhands (separately and the two groundstrokes combined) between JUN and PRO female players undergoing TEST along with relationships between players’ rank order and these responses. Whereas technical decrement did not occur earlier in JUN, magnitudes of change were larger in PRO above VT2 intensity. Our novel results are that (i) BV_b and TP_b are two technical parameters that could discriminate JUN and PRO female players; (ii) PRO displayed higher sub-maximal and maximal aerobic fitness level than their less experienced counterparts. These results confirm only partly our initial hypothesis given that TP was not uniformly higher for PRO.

Technical responses

Interestingly, BA was higher for PRO at 70% and 75% of V˙O_2max (Fig. 2B) corresponding nearly to the typical averaged relative V˙O₂ (60%–75% of V˙O_2max) (^22,23) and blood lactate concentration (~2 mmol·L⁻¹) (¹) values reported for tennis competitions. Previous field test studies have assessed the effect of fatigue on technical skill performance and highlighted that BA is of paramount importance (^11,13,14,24). Recent study reported that international male players were able to maintain significantly higher levels of BA (11% greater on average) through high exercise intensities, and the stepwise discriminant analyses suggest that the ability to maintain high levels of BA at high intensities may also be a factor that differentiates players of international and national level. (²⁵). Reportedly, BA explained 37% of variability in competitive performance (r = 0.61; P = 0.001) as well as 53% and 55% in combination to V˙O_2max and VT2 values, respectively (¹⁰). Almost perfect correlations (r = 0.94, P < 0.001) have also been reported between stroke ratings (as a surrogate of BA) and players’ ranking (²⁶). The difference in BA_mean in our study (55%–60%) in reference to the study by Baiget et al. (¹⁰) (66% ± 9%) could possibly be explained by ball machine throwing velocity (86.5 ± 1.3 km·h⁻¹ vs 68.6 ± 1.9 km·h⁻¹). These authors also reported differences in the specific endurance tennis test (“SET test”) when they confronted elite (70% ± 6%) and subelite (63% ± 9%) male tennis players (²⁵), using the “target field” originally designed by Smekal et al. (¹²). When evaluating the reliability of stroke precision during the “on-court endurance testing for tennis,” they found lower values in BA (40.9% ± 11.7%) (¹²). No female data with similar field test designs are available. In our study, the increase in the number of missed strokes could be related to poor timing or an inability of players to position themselves properly for efficient stroke hitting (i.e., being “late”). In support, a decrement in maximal running speed likely occurs as players become fatigued, resulting in suboptimal stroke preparation (e.g., footwork and balance) and ultimately in a slowing of BV (²⁷). Additional biomechanical analysis of stroke production combining three-dimensional motion (segment positioning), electromyography (muscle activation) and pedobarography (plantar loading) analysis would help to explain the reasons for such suboptimal ball placement as fatigue develops. This seems particularly relevant in the context of tennis where important interindividual differences exist regarding effective groundstroke production (^28,29), which could be exacerbated under fatigue.

Although BV in PRO was on average higher than in JUN throughout TEST (Table 1), significantly higher values were only reported for BV_b (Table 1). This between-group difference in BV_b is an important observation that would contribute to partly explain the higher performance level in WTA-ranked players compared with ITF-ranked players. Recent advances in tracking technologies have made feasible to compare shots characteristics of JUN and PRO, in particular with more objective analysis during actual tennis competition. A recent study provided a comprehensive comparative analysis of JUN and PRO match play between 2000 and 2015 (⁷). Forehand and backhand speeds achieved during competitions were comparable in women PRO and JUN (forehands, 111 vs 110 km·h⁻¹; backhands, 106 vs 103 km·h⁻¹, respectively) (⁷). It seems difficult to conclude on skills comparison between different categories of players during actual competition as the opponent level and emotional influences likely play a crucial role. For these reasons, we believe that standardized conditions as encountered during TEST are more appropriate when comparing players of various standards.

Physiological responses

The load increments for PRO and JUN during TEST were relatively similar as evidenced by the progressive increases in V˙O₂, V˙_E, and HR (Fig. 1) responses. V˙O₂ and V˙_E curves for PRO were globally higher compared to their less experienced counterparts. It is known that there is coordination (“entrainment”) between limb (and especially upper limbs) movements and breathing (³⁰). Active breathing is probably increased in PRO due to moderate effect in BV_mean (110.4 ± 7.1 vs 104.4 ± 8.6 km·h⁻¹, ES = 0.76).

PRO possessed higher aerobic fitness (V˙O_2max, % HR_max at VT, and stage number reached at VT2) (Table 1). Our V˙O_2max values are higher than those reported by Ferrauti et al. (²⁷) with national level (49.0 ± 3.9 mL·min⁻¹·kg⁻¹ for adult (n = 13) and 47.3 ± 4.6 mL·min⁻¹·kg⁻¹ for under 16 female players (n = 14)) during a specific test without ball hitting. To our knowledge, no V˙O_2max data are available that have assessed female players practicing a specific test to exhaustion. Bergeron et al. (³¹) previously indicated that players with a well-developed V˙O_2max would better sustain cardiovascular load and improve their recovery between points. In support, a strong inverse relationship between V˙O_2max and ATP entry ranking over time in a professional tennis player has been reported (³²). Previous findings also postulated that for a top female player, it is important that V˙O_2max is above 50 mL·min⁻¹·kg⁻¹ but that higher V˙O_2max values (> 65 mL·min⁻¹·kg⁻¹) would not further improve on-court performance against a V˙O_2max of ~55–60 mL·min⁻¹·kg⁻¹ (^23,33). This is reflected here where PRO displayed averaged V˙O_2max values of 54.9 ± 3.3 mL·min⁻¹·kg⁻¹. However, the moderate correlation of competitive ranking with V˙O_2max (L·min⁻¹) (r = 0.53) underlines the pervasiveness of aerobic capacities without solely reflect a tennis level. This is because performance in tennis is largely dependent on the technical, explosive sequence, tactical, and motor control/coordination aspects.

PRO had smaller %HR_max values than JUN at VT (VT1: 87.3% ± 4.2% vs 91.6% ± 4.6%; VT2: 96.0% ± 1.3% vs 97.2% ± 1.0%). Surprisingly, only limited data are available regarding VT values in tennis players. Our aforementioned HR data are slightly higher than those generally reported in different categories of players (more often male than female, from regional to international level) undertaking a tennis specific field test ([79%–87% of HR_max] at VT1 (^10,20), [85%–95% of HR_max] at VT2 intensities (^{10,12,14,16,20})). According to König et al. (²³), high VT values could reflect the ability to tolerate high exercise intensity during tennis competitions. Low-to-moderate correlations were found elsewhere between VT and players’ competitive level (r = 0.35, P = 0.038 and r = 0.55, P = 0.001 for VT1 and VT2 respectively) (¹⁰). The discrepancies between our results and these studies are mainly the result of the training status of the tested players. Nevertheless, comparing results between studies is only anecdotal since subject characteristics, equipment, protocols, and test modes as well as the methods used to detect ventilatory breakpoints differ between studies. HR is affected by factors as emotional stress, dehydration, and illness can all cause changes in HR without associated changes in V˙O₂ (³⁴). Beyond physiological capabilities, an adapted HR is probably due to their “expertise.”

PRO showed significantly higher stage at VT2 and MAF. Differences can essentially be explained by physiological capacities because TP is also higher in PRO until 90% of V˙O_2max. Consequently it is unlikely that the higher stages reached can be explained by a lesser engagement in stroke production (cf. Fig. 2A). The elevated aerobic fitness of PRO would be explained by the need of higher aerobic condition to deal with the intensity of high-level competition. Similar findings have already been reported when International and National male players were compared on the SET test (²⁵). In this later study, an important finding was that international players showed better aerobic fitness (on average, V˙O_2max and VT2 were 8% and 10% greater, respectively) and better performance during the specific field test as compared with their national-level counterparts. Baiget et al. (¹⁰) indicated that a large part of the variability in the 38 competitive players that they have tested on a specific incremental test could be explained by time to exhaustion and physiological parameters (VT2 and maximal load). Although most of the important actions during the short-term periods of activity (i.e., strokes, accelerations or changes of direction) depend fundamentally on the anaerobic metabolism (intramus cular phosphates and glycolysis), the aerobic metabolism (oxidative phosphorylation) allows resynthesizing the high-energy phosphates during recovery periods (^26,35). Adequate aerobic fitness promotes better physiological regeneration between points, matches and tournaments to maintain a high competitive level throughout the season (³⁶). Our significant small to moderate correlation between the ranking level and V˙O_2max (r = −0.421), and stage at VT2 (r = −0.50) also confirm precedent findings mostly derived from male players (^10,16).

Of interest is also that [La-] increase above 90% of V˙O_2max was higher in PRO corresponding to lower TP than JUN at 100% V˙O_2max (Fig. 1C). BA and TP indices also became further altered compared to JUN (Fig. 3). This difference between our two groups of players above 90% of V˙O_2max is probably due to a permanent physical engagement in the strokes within professionals, suggesting increased involvement of anaerobic glycolytic processes to supply energy (³⁵). In support, an inverse correlation (r = − 0.51, P = 0.008) has been observed between changes in TP and blood lactate concentration from 60% to 100% of V˙O_2max. Even if our PRO athlete are probably not as aerobically-trained as long distance runners, VT values (cf. Table 1) observed in our players corresponded to values that are typically measured in high-level endurance athletes (³⁷). In a study where match-like conditions were reproduced in PRO tennis players, blood lactate values as high as 8.6 mmol·L⁻¹ have been reported (³⁸). This suggests a substantial participation of glycolytic processes to meet the energy supply and to perform high technical skills when game accelerates. Dramatic TP reductions are commonly observed when tennis players near or reach volitional exhaustion (blood lactate concentration, 9.6 mmol·L⁻¹) (¹¹). Whatever the exact physiological mechanisms, we hypothesized that peripheral disturbances partly caused deterioration in BV and BA at high intensities and explain the result in TP at 100% of V˙O_2max.

Practical applications

This study provides evidence that, when designing a training program for elite junior female tennis players, improvement of their aerobic capacity should represent a priority. Some of the largest differences between junior and professional are seen in the physical demands imposed on players during game play (⁷). To develop efficient groundstrokes (optimal speed and accuracy combined with high MAF), the usefulness of polarized training has recently been promoted (¹⁴). Although tennis performance analysis underlines the close relationship between physical and technical parameters, scientific or coaching approaches often neglect to concurrently develop these two aspects during the same training session. We therefore propose that high-intensity situations encountered during TEST (above 90% HR_max) could be used by coaches in the programming of an overload training stimulus during tennis practice of female players (¹⁸). For instance, when targeting improvement in technical skills, we recommend to train at low intensities, in zone 1 (V˙O₂ at or below VT1) (³⁹), corresponding to BA from 50% to ~70% and BV from 105 to ~115 km·h⁻¹, depending on the level of practice. Time in zone 1 brings physiological benefits and will indirectly lead to improved capacity to cope with the higher intensity of competitive tennis (zone 3: V˙O₂ at or above VT2) corresponding to the “money time,” in a small playing time (3% ± 5%) (⁴⁰).

Implementing TEST as a profiling tool seems meaningful when including both physiological and technical analysis, allowing possible comparison with data henceforth available. Based on observed difference in MAF, V˙O_2max, BA_b and TP_b we propose that TEST can be used as a discriminating tool to evaluate the potential of a young female player to eventually reach the professional level.

Strengths and limitations—perspectives

Although all efforts have been made to make TEST as specific as possible to the game of tennis, it is also acknowledged that simulated tennis practice does not completely reflect the actual competition situation (lack of visual cues, lower uncertainties, and anticipation). We cannot minor that tennis performance is multifactorial, and there are basic performance skills, such as the psychological, tactical, or strategic capabilities that could not be incorporated in our comparison of training backgrounds.

Although our study highlights the potential of TEST, that uses backhand and forehand strokes, as an evaluation and/or selection test for the higher level, the serve lacks an overall analysis. Hence, producing faster BV (+10 km·h⁻¹) for the first serve in particular or maintaining BA under fatigue have proven necessary skills for an efficient transition to professional level (⁷).

Although we have recruited top-level women players (including some of the best representative players for France), we cannot rule out that larger differences would have been noted if players with higher ITF junior and lower WTA rankings were recruited. Future studies comparing female players at the ITF Transition Tour level and others ranked within the Pro circuit (ITF and/or WTA) are needed. For successful completion, this approach might well require an international multicentric approach to reach the required sample size. A larger one would allow more fine grained assessment of the effects of TEST with varying rank among professional players.

CONCLUSIONS

By using TEST that offers the possibility to concurrently evaluate differences in technical and physiological parameters our novel findings indicate that professional compared with less experienced (juniors) players possess higher aerobic fitness level and more proficient TP for the backhand stroke. Professional players have lower % HR_max values combined to their technical skills. To reach the professional level in female players, it is therefore crucial that physical conditioners working with elite juniors put strong emphasis to further develop aerobic fitness of their players. Reinforcing the backhand technique and the muscular power involved in this stroke, notably in the context of fatigue development, also appears as a priority.