Badminton is an intermittent sport characterized by multiple intense actions (12), and the specialized movement pattern consists of fast accelerations, decelerations, and many explosive movements with changes of direction over short distances (3,4,7,15,17). The duration of international single matches may vary from ∼15 to ∼90 minutes, with the length of each point varying from a few seconds to several minutes (2,28). Many rallies are decided in less than 10 seconds (7), and the intense actions during match play require sporadic movements of high intensity related to repetitive actions of short duration but highly explosive (2,15,17). Anaerobic power and explosive strength, therefore, seem to be of major importance for the physiological performance level, and accordingly, it has been reported that maximal strength and anaerobic alactacid power are high in elite badminton players (2,4). Performance in both individual and team sports relies on tactical and technical parameters and the players' physiological capacities. Therefore, it is essential for physiologists and coaches to develop tests and analyses to specify the players physiological capacities.
Sprint tests, repeated sprint tests, rate of force development measures, or squat jump tests may be used for a general determination of the players explosive exercise capacity (1,18,24,25,27), but the transfer ability from such tests to badminton or other racket sports is questionable because of the highly specialized movement pattern that characterize these sports (15,17,19,29,30). Wilkinson et al. (30) report that a squash-specific test for evaluation of multiple sprint ability in squash players correlates to the players' performance level, whereas repeated sprint tests could not discriminate between elite squash players and matched subjects without experience with racket sports. Specialized endurance and custom-made sprint tests of relevance for various sports have been evaluated and validated (4,9,10,21,29,31,32), but valid tests for the assessment of speed and agility in badminton players are lacking.
Thus, the available tests within badminton in scientific literature are mainly focused on endurance (4,8,16,32), and these tests, as well as the few published tests focusing on speed or repeated sprint ability, have used predetermined exercise patterns instead of randomized movement pattern, which is characteristic for badminton (16,17,23). Ooi et al. (26) performed sideways tests and 4 corner agility tests (14) in Malaysian elite and nonelite players, but both of these on-court badminton tests failed to differentiate between elite and nonelite players.
Therefore, the purpose of this study was to develop a badminton-specific speed test (BST), which includes a random movement pattern comparable with real game situations and evaluate the reproducibility and specificity of the developed test. It was the intention that the BST should be able to differentiate between elite, less skilled, and non-badminton players and with test reproducibility comparable with simple sprint tests. The developed test would be of specific relevance for badminton. However, we propose that the test design may be adopted by other sport disciplines characterized by an intermittent exercise pattern with many fast accelerations, decelerations, and explosive movements with changes of direction over short distances.
Experimental Approach to the Problem
The badminton-specific speed test was developed on the basis of video observations and review of literature with analysis of international competitive elite matches and tests (2,4,7,8,14,15,17,20,22,23,32). Pilot experiments with different positions of the sensors, time between rallies, and total number of actions were conducted to develop a test that challenged the player's agility on court, with movement and exercise patterns that resembled real match play.
Subsequently, the test was evaluated for specificity and reproducibility. To provide sufficient data, we recruited 61 healthy male subjects to fit into 1 of 3 categories: elite (n = 20), skilled (n = 21), and non-badminton players (n = 20). They completed 3 trails of a 30-m sprint tests and 3 trails of the BST. Furthermore, 6 of the elite players completed 3 trials of the BST on separate days to address day-to-day reproducibility.
Sixty-one healthy male subjects with an age of 23.7 ± 4.9 years (range 18-33 years), height 184 ± 6.2 cm, and weight 77.8 ± 7.6 kg participated in the study: elite (n = 20) (all in the top 100 on the Danish national badminton ranking list; DBF) with a mean age of 24.6 ± 4.5 years, height 186.7 ± 7.0 cm, weight 77.7 ± 6.0 kg; skilled (n = 21) (all players with a ranking between 100 and 1,000 on the Danish national badminton ranking list) with an age of 24.9 ± 4.8 years, height 182.2 ± 5.1 cm, and weight 77.8 ± 7.2 kg; and non-badminton players (n = 20) with an age of 21.7 ± 5.1 years, height 183.2 ± 5.9 cm, and weight 77.8 ± 9.5 kg. The elite players were recruited from Team Denmark national team training and local clubs in Copenhagen. By the time of testing, both elite and skilled players were active in the Danish team tournament, and they were ranked in the national rankings of male badminton players (www.badmintonpeople.dk/DBF/Ranglister/). The non-badminton players were physically active men matched for age and stature. All non-badminton players were engaged in regular training but with no previous record of participation in organized badminton training. The subjects were instructed not to participate in strenuous exercise for 24 hours before testing and attend the test well hydrated. Test-retest evaluation was performed on similar times of the day (scheduled within ∼1 hour for each individual), whereas different times of the day were used for testing of each individual. The study was approved by the University of Copenhagen. All participants volunteered to participate and gave written informed consent for the use of their test data as approved by the Ethic Committee of The Capital Region of Denmark (H-4-2012-FSP), and testing was performed according to the Declaration of Helsinki.
Thirty-Meter Sprint Test
The subject performed active warm-up for 15 minutes (running and stretching activities) before completing 3–4 habituation trials at submaximal intensities. Subsequently, the subject completed 30-m sprints from a standing start position, with the start and finish time recorded by photocells, and the time for 30-m mark stored electronically (Newtest Power Timer System, Oulu, Finland, in accordance with Ref (6)). Each subjects completed 3 attempts separated by 1 minute of slow jogging followed by 1 minute of standing rest. The fastest sprint time was used as the test result of each of the subject's 30-m sprint performance.
Badminton Speed Test
After completing 30-m sprint test, subjects rested for 4 minutes before starting BST trials. Subject first performed a familiarization trial followed by 3 maximal trials of the BST with approximately 5 minutes of recovery between each attempt. Subjects had to perform the BST as fast as possible and were motivated by the technical staff. For each subject, the best of the 3 total times was chosen as the person's maximum speed test performance.
Figure 1 provides an overview of the experimental test setup with the computer screen at the net in front of the subject, and the 4 sensors positioned in each of the 4 corners of the badminton court (see Video, Supplemental Digital Content 1, https://links.lww.com/JSCR/A8, which demonstrates the experimental test setup).
The BST starts in the center of the badminton court and involves a total of 20 maximal actions toward custom-made sensors (Larsen Elektronik, Praesto, Denmark) positioned on each of the 4 corners of the single badminton court. One-second interval between each action allows the player a brief of time to return to the center of the court before the next action. The sequence of the movements to each corner (where the subject should hit/activate the given sensor with his racket) was randomly dictated by a computer program (Direct RT 2008; Empirisoft Corporation, New York, NY, USA) and shown on a screen in front of the subject. Two sensors were placed on the front court 60 cm from the net and 50 cm above the floor on the single line, with 1 sensor in the forehand corner and another in the backhand corner. The other sensors were positioned on the back of the court in the forehand side 250 cm above the floor and in the backhand side 200 cm above the floor, vertically above the cross double serve line and the single side line. The subject was requested to hit all sensors from behind with the racket, similar to a real badminton stroke when returning a shuttle.
Before each action (the screen shot dictating the sensor to be activated), a visual countdown consisting of the numbers 3, 2, and 1 was displayed on the computer screen in front of the subject during a 1-second period. The subsequent activation of the correct sensor was then followed by the next 1-second interval (with the next countdown sequence), allowing time to return to the center of the court and allowing the player to time his preparation for the next movement and get a rhythm analogously to real play. The randomized sequence and the countdown before each action were constructed to mimic real match situations where the player must be able to react and move fast in any direction depending on the opponent's next stroke. The BST was completed when 5 actions to each corner had been completed in the correct order, and BST performance was expressed as the accumulated time for all 20 actions (i.e., the time from the visual input was displayed on the screen to the correct sensor was activated).
Differences between groups in terms of performance, anthropometric data, and physiological characteristics were tested with 1-way analysis of variance (ANOVA), whereas day-to-day variation was evaluated with 1-way ANOVA for repeated measurements. The coefficient of variation (CV) was calculated as a measure of test reproducibility as the SD between tests divided by the average test result. All values are presented as mean ± SD, unless otherwise specified. The significance level was set at p ≤ 0.05, with an estimated power of 80% to detect differences in performance between the 3 groups (with the sample size of ∼20 in each group). In the case of a significant difference in mean values, the Tukey's post hoc test is used to determine significant difference between the test groups. Furthermore, intraclass correlations were determined with Pearson's product-moment correlation, and statistical significance was accepted at p ≤ 0.05.
The elite group was significantly faster in the badminton-specific speed test with a best time of 32.3 ± 1.1 seconds compared with the skilled players (34.1 ± 2.0 seconds) and non-badminton players (35.7 ± 1.7 seconds) (Figure 2A), whereas there were no differences between groups in the 30-m sprint test (Figure 2B).
Considering all subjects in the 3 groups, there was a weak correlation between performance in the BST and 30-m sprint test (r = 0.50, p < 0.01). Within the elite group, there was no correlation between BST and 30-m sprint performance (Figure 3).
For the elite group, the average reaction and movement times were 1.6 seconds per action, and the average time per action was similar for the first 10 and last 10 actions in the BST (Figure 4). In contrast, both the skilled and non-badminton players were significantly slower during the last 10 actions of the BST compared with the first 10 actions.
For tests performed on the same day, there was no significant difference between trials 1, 2, and 3 for any of the 3 groups (elite: trial 1: 32.8 ± 1.3 seconds, trial 2: 32.8 ± 1.4 seconds, trial 3: 32.9 ± 1.3 seconds; skilled: trial 1: 35.4 ± 1.9 seconds, trial 2: 34.8 ± 2.3 seconds; trial 3: 34.6 ± 1.9 seconds; non-badminton players: trial 1: 36.9 ± 2.1 seconds; trial 2: 36.6 ± 2.0 seconds; trial 3: 36.5 ± 1.8 seconds). For tests performed on the same day, this corresponds to a CV of 1.7% for the elite, 2.6% for skilled, and 2.5% for non-badminton players. The interindividual variation between subjects on the same day was 1.1 seconds (range: 0.22–2.27 seconds) for elite, 1.8 seconds (range: 0.23–4.8 seconds) for skilled players, and 1.8 seconds (range: 0.24–3.94 seconds) for non-badminton players. Furthermore, for the 6 elite players completing the BST on 2 separate days, neither average test performance nor time for the fastest trial was different. The average difference in the best performance was −0.1 seconds (range: −0.56 to 0.24 seconds), and the CV was 0.7% (Figures 5A, B), and the average difference between average time for the 3 trails was −0.1 seconds (range: −0.59 to 0.25 seconds).
In this study, we have developed a novel badminton test with a randomized movement pattern for evaluation of maximal speed during game-like conditions. Our results let us suggest that this test is of relevance for the assessment of badminton players' physical exercise capacity. Furthermore, from a practical point of view, the BST may be used by physiologists and coaches for cross-sectional comparison of players or for evaluation of longitudinal changes, for example, individual training adaptations. The BST could discriminate between elite, skilled, and non-badminton players with similar performance in 30-m sprint tests, indicating the sport specificity of the test. Furthermore, the BST had reasonable test-retest reproducibility on successive trials performed on the same day and with a low day-to-day variation for elite players.
The CV of the BST is comparable with the CV reported in evaluation of endurance, sprint, or jump performance (5,6,13). The CV of the BST for the 3 trials performed on the same day was below 2% in the elite group and ∼2.5% for the less skilled players. Furthermore, the day-to-day variation of the present test is similar to speed tests, which includes straight forward running (6) or on-court squash and badminton tests with a predetermined fixed movement pattern (28,29). Therefore, the randomized movement pattern, used in the present test to resemble match play closer than stereotypical tests, did not compromise the reproducibility of the test.
We used 1 familiarization trial as habituation to the BST, which seemed to be sufficient to minimize variation, and we observed no systematic bias during the subsequent 3 trials preformed on the same day. It could be expected that further habituation would lower the variation, although the CV obtained for the present test is considered acceptable and comparable with those reported by similar types of tests (6,11,13,24).
The low day-to-day variation for the elite players and the low variability in the test on the same day may in part relate to the players habituation by match play and daily training in the particular movement pattern of importance for the test (13). Therefore, the low variation of the BST can be used by coaches and physiologist to track small improvements (or decrements) in individual players, thanks to training procedures.
The BST consists of 20 strokes to minimize the variability of the test, without compromising the elite player's ability to maintain maximal movement speed throughout the test. With only 10 actions, the CV would have been 6 ± 2% in the elite group, that is, more than 3 times greater than with 20 actions. However, as illustrated in Figure 4, the average movement time for each action would not have been different. For each movement to a given corner of the court, there is a considerable variation (ranging from 15 to 30% to the front and back court corners for the elite players), and evaluation of the fastest movement time for a single action would not be reproducible. However, when the actions are composed into a test with 20 actions, the overall variation is markedly minimized.
The BST was able to discriminate between elite, skilled, and non-badminton players, whereas performance in the 30-m sprint test was not different across the 3 groups. Our results, therefore, let us suggest that the BST is more relevant for testing on-court speed agility for badminton players than a simple sprint test. In accordance, Wilkinson et al. (30) report no difference in running speed tests but superior performance in elite squash players in a squash-specific speed tests compared with non-squash players. Furthermore, Walklate et al. (28) report significant improvements in sprint performance with changing directions after specific training but observed that this was not transferable to straight forward sprinting.
It has been shown that daily training lead to higher cross-sectional area of both type I and II fibers in the dominant leg compared with the nondominant leg in elite badminton players (22) and badminton players. Furthermore, they have greater difference in circumference and maximal power between the dominant and nondominant leg compared with athletes in sports where the movement pattern is less asymmetrical (20,22). Therefore, using a test that assesses the above-mentioned physiological adaptations may be superior to a nonspecific speed test. However, it seems that not only the inclusion of badminton-specific movements but also randomizing the order of the movements is a major strength of the current test. The study of Ooi et al. (26) failed to detect differences in test performance between elite and subelite badminton players in a BST with a predetermined and fixed movement pattern. Although the test was performed on a badminton court using a predetermined exercise pattern, it is not equivalent to real match play. This may explain why such tests are less relevant for testing of badminton players
The BST was developed to include relevant movements and resembles match play; however, some compromises were necessary to develop a reliable test. The total duration of the test was ∼50 seconds, with ∼30 seconds of the total time as the combined reaction and movement time toward the sensors and 19 seconds (1 second between each action) to return to the center of the court. Total test time was, therefore, longer than the average rally length in elite badminton matches (2,7), and the test was substantially longer than traditional sprint tests (1,11,24). As previously discussed, this was necessary to reduce test variation, and because it seemed that maximal movement speed could be maintained throughout the test in the elite trained players, we concluded that 20 maximal actions with an average duration of approximately 1.5 seconds interspersed by 1-second break to return to the center of the court was acceptable for testing maximal movement speed. Also, the total test duration was comparable with duel lengths, occasionally observed in international top matches (2,7).
In conclusion, the developed badminton-specific speed test has shown a low day-to-day variability in elite players and a low CV similar to other general tests. Therefore, it seems to be relevant for coaches and physiologist to use the BST to evaluate maximal movement speed in different groups.
We recognize that this study is a cross-sectional evaluation where different groups are compared. Next step may be to conduct a longitudinal study and investigate if seasonal variation, training, or conditioning may be reflected in test performance and develop a method to evaluate if test performance correlates with on-court moving speed.
From a practical perspective, we suggest that physiologists and coaches may use the test for establishing standards for elite players on-court movement speed. Knowledge about elite standards would provide a benchmark for training programs. This standard may also be used for comparison of players. The test can be used to evaluate longitudinal changes in performance, for example, seasonal variations, development of young players, or adaptations to explicit training interventions. We suggest that this test design could be used by other sport disciplines with intermittent exercise that match badminton physiological requirements.
The authors are grateful to the Danish Badminton Association and Stenhus Badminton College for supporting this study. Furthermore, they express their appreciation to the Danish badminton national team players and all the other subjects for their participation in the study.
1. Buchheit M, Spencer M, Ahmaidi S. Reliability, usefulness, and validity of a repeated sprint and jump ability test. Int J Sports Physiol Perform 5: 3–17, 2010.
2. Cabello Manrique D, González-Badillo JJ. Analysis of the characteristics of competitive badminton. Br J Sports Med 37: 62–66, 2003.
3. Chin M, Steininger K, So RC, Clarkt CR. Physiological profiles and sport specific fitness of Asian elite squash players. Br J Sports Med 29: 158–164, 1995.
4. Chin M, Wong AS, So RC, Siu OT, Steininger K, Lo DT. Sport specific fitness testing of elite badminton players. Br J Sports Med 29: 153–157, 1995.
5. Currell K, Jeukendrup AE. Validity, reliability, and sensitivity of measures of sporting performance. Sports Med 38: 297–316, 2008.
6. Enoksen E, Tønnessen E, Shalfawi S. Validity and reliability of the New Test Power timer 300-series testing system. J Sports Sci 27: 77–84, 2009.
7. Faude O, Meyer T, Rosenberger F, Fries M, Huber G, Kindermann W. Physiological characteristics of badminton match play. Eur J Appl Physiol 100: 479–485, 2007.
8. Fuchs M, Faude O, Wegmann M, Meyer T. Critical evaluation of a badminton-specific endurance test. Int J Sports Physiol Perform 9: 249–255, 2014.
9. Girard O, Chevalier R, Leveque F, Micallef JP, Millet GP. Specific incremental field test for aerobic fitness in tennis. Br J Sports Med 40: 791–796, 2006.
10. Girard O, Sciberras P, Habrard M, Hot P, Chevalier R, Millet GP. Specific incremental test in elite squash players. Br J Sports Med 39: 921–926, 2005.
11. Glaister M, Howatson G, Lockey R. Familiarization and reliability of multiple sprint running performance indices. J Strength Cond Res 21: 857–859, 2007.
12. Gowitzke BA, Waddell DB. Biomechanical studies of badminton underarm. Power strokes, court movement, and flexibility. 1989. Paper presented at: 7th International Symposium on Biomechanics in Sports 1989.
13. Hopkins WG, Schabort EJ, Hawley JA. Reliability of power in physical performance tests. Sports Med 31: 211–234, 2001.
14. Hughes M, Fullerton F. Development of an on-court aerobic test for elite badminton players. In: Science and Racket Sports. Reilly T, Hughes M, Less A, eds. London: E & FN Spon, London. 1995. pp. 51–54.
15. Hughes MG. Field-based assessment of speed and power in junior Badminton players. In: Science and Racket Sports IV. Lees A, Cabello D, Torres G, eds. Oxon, United Kingdom: Routledge, 2009. pp. 70–76.
16. Hughes MG, Andrew M, Ramsay R. A sport-specific endurance performance test for elite badminton players. “Communications to the 12th Common wealth International Sport Conference”. J Sports Sci 36: 277–278, 2003.
17. Hughes MG, Bopf G. Relationships between performance in jump tests and speed tests in elite badminton players. Annual Conference of the British Association of Sport and Exercise Sciences. J Sports Sci 23: 195–196, 2005.
18. Kale M, Asçi A, Bayrak C, Açikada C. Relationships among jumping performances and sprint parameters during maximum speed phase in sprinters. J Strength Cond Res 13: 2272–2279, 2009.
19. Lees A. Science and the major racket sports: A review. J Sports Sci 21: 707–732, 2003.
20. Majumdar P, Khanna GL, Malik V, Sachdeva S, Arif M, Mandal M. Physiological analysis to quantify training load in badminton. Br J Sports Med 31: 342–345, 1997.
21. Micklewright D, Papapdopoulou E. A new squash specific incremental field test. Int J Sports Med 29: 758–763, 2008.
22. Mikkelsen F. Physical demands and muscle adaptation in elite badminton players, 1979: 55–65.
23. Ming CL, Keong CC, Ghosh AK. Time motion and notational analysis of 21 point and 15 point badminton match play. Int J Sports Sci and Eng 02: 216–222, 2008.
24. Moir G, Button C, Glaister M, Stone M. Influence of familiarization on the reliability of vertical jump and acceleration sprinting performance in physically active men. J Strength Cond Res 18: 276–280, 2004.
25. Moir G, Sanders R, Button C, Glaister M. The influence of familiarization on the reliability of force variables measured during unloaded and loaded vertical jumps. J Strength Cond Res 19: 140–145, 2005.
26. Ooi CH, Tan A, Ahmad A, et al.. Physiological characteristics of elite and sub-elite badminton players. J Sports Sci 27: 1591–1599, 2009.
27. Requena B, González-Badillo JJ, Saez Saez de Villareal E, et al.. Functional performance, maximal strength, and power characteristics in isometric and dynamic actions of lower extremities in soccer players. J Strength Cond Res 23: 1391–1401, 2009.
28. Walklate BM, O'Brien BJ, Paton CD, Young W. Supplementing regular training with short-duration sprint-agility Training leads to a substantial increase in repeated sprint-agility performance with national level badminton players. J Strength Cond Res 23: 1477–1481, 2009.
29. Wilkinson M, Leedale-Brown D, Winter EM. Reproducibility of physiological and performance measures from a squash-specific fitness test. Int J Sports Physiol Perform 4: 41–53, 2009.
30. Wilkinson M, Leedale-Brown D, Winter EM. Validity of a squash-specific test of change-of-direction speed. Int J Sports Physiol Perform 4: 176–185, 2009.
31. Wilkinson M, Leedale-Brown D, Winter EM. Validity of a squash-specific fitness test. Int J Sports Physiol Perform 4: 29–40, 2009.
32. Wonisch M, Hofmann P, Schwaberger G, von Duvillard SP, Klein W. Validation of a field test for the noninvasive determination of badminton specific aerobic performance. Br J Sports Med 37: 115–118, 2003.