Swimming velocity (V) has been defined as a product of stroke rate (SR) and stroke length (SL). Stroke rate refers to the number of complete cycles of 1 arm in a given unit of time, and SL is the distance the swimmer moves forward per stroke (9,15). Because SR and SL are directly related to swimming V, many authors have suggested that the measurement of SR and SL is important for improving the knowledge of the combinations that produce V and informative with respect to performance and performance potential. Research has shown that the SR and SL combinations may change with V (8,17), stroke (14,18), race distance (1,19), and anthropometric characteristics (6,12). It has been reported that swimmers reach maximum V by increasing SR and decreasing SL (17), while an increase in SR above an optimal point may decrease V (8). SL has been generally identified as the main performance determinant (5), but some investigators have reported SR to be the most influential parameter for a few swimmers in competition events (11).
Blood lactate concentration (BLa), while not a direct measure of acidosis (21) or a true representation of the BLa in working muscle, reflects the balance between production and appearance in blood and removal processes. As swimming V increases, the rate limitation of aerobic metabolism leads to a greater proportion of contribution to adenosine triphosphate (ATP) synthesis by glycolysis, which causes increased lactic acid production, which increases hydrogen ion concentration, which causes acidosis. BLa values obtained at the end of a race or during training sessions could be an important indication of the contribution of anaerobic mechanisms to the total energy requirements (4). Thus, the measurement of BLa as a training assessment technique and for the purpose of designing training programs, especially during submaximal and maximal swimming efforts, has been recommended by several authors (13). The influence of BLa on performance, according to age, gender, stroke, and distance of the event, has been examined during swimming tests (23,25) or after competition (2,22). Peak postcompetition BLa levels seem to differ according to stroke and distance, and the ability of a swimmer to reach and sustain higher BLa levels has been shown to have a positive influence on performance (2,15).
Because most competitive swimming events require significant contributions from both aerobic and anaerobic sources, incremental tests have been suggested as useful tools in monitoring changes in several fitness aspects and performance characteristics of swimmers (20). Although several researchers have examined separately the influence of kinematic (i.e., SR and SL) and physiological (i.e., BLa) parameters in swimming performance, the relationship between stroke kinematics and BLa during swimming tests or competitions has been the topic of only a few studies (15-17,24). These studies have shown that swimmers generally increase V by increasing SR and decreasing SL, with a more dramatic decrease in SL above certain BLa levels (i.e., usually at the onset of blood lactate accumulation (OBLA) point, defined as 4 mmol·L−1) attributed to resulting local muscle fatigue or increased metabolic acidosis.
However, there is a major limitation in the existing studies in this area. Most of the swimmers tested were not specialized in the stroke (i.e., freestyle) and/or distance (i.e., 100, 300, or 400 m or combination thereof) used in these tests. Consequently, given that stroke kinematics and BLa are influenced by the stroke and event distance, the data reported in these studies are likely to provide misleading information regarding the stroke and distances used. Furthermore, these data have a limited ability to be generalized and limited applicability for swimmers of other strokes and distances. Finally, there is a lack of data for swimmers specialized in strokes other than the freestyle and in distances other than the 100, 300, and 400 m used in the aforementioned studies.
The purposes of the current study were to investigate and evaluate the relationships between stroke kinematics (i.e., SR and SL) and BLa for elite swimmers competing at the international level during incremental exercise in their specialist events and distances and, moreover, to examine the extent to which changes in these parameters are associated with swimming performance, as indicated by average V. Additionally, a case study for an elite swimmer was conducted to examine the relationships between these variables throughout a series of tests and whether performance in the test was associated with training status and competition performance.
Experimental Approach to the Problem
The participants for this study were members of the British National Swimming Team. As part of the national team testing requirements and for the purposes of assessing their training status and the kinematic and physiological characteristics of their performance, these swimmers performed a 7 × 200-m incremental test several times throughout the year. This is a popular swimming test that is being used worldwide by other national swimming teams and private swimming clubs (20). The 7 × 200-m test was designed to provide coaches and swimmers the aforementioned performance-related information, which coincided with the aims of this investigation. Moreover, as outlined above, there is a lack of information with respect to 200-m events. For these reasons, the 7 × 200-m incremental test was conducted to address the purposes of this study. In line with the purposes of this study, only swimmers specialized in 200-m events were selected to participate, with all swimmers performing their specialist stroke throughout the test. To explore the relationship between kinematic and physiological parameters and swimming performance, the following variables were calculated for each 200-m swim: SR, SL, V, and BLa.
Eleven men (age, 23.0 ± 2.5 years; height, 187.0 ± 5.7 cm; and weight, 82.3 ± 6.9 kg) and 10 women (age, 21.3 ± 3.5 years; height, 171.4 ± 5.3 cm; and weight, 61.4 ± 5.2 kg) participated in this study. The personal best performance of the year, expressed as a percentage of the world record average speed, was 96.7% ± 0.9% for men and 95.7% ± 1.0% for women. According to each swimmer's specialist stroke, the group could be divided as follows: 10 freestyle (6 men and 4 women), 5 backstroke (3 men and 3 women), 4 breaststroke (2 men and 2 women), and 2 butterfly swimmers (1 man and 1 woman). The participant for the case study was a male butterfly swimmer (age, 25 years; height, 181 cm; weight, 75 kg; and personal best performance, 98.1% of the world record).
All tests were conducted in a 50-m indoor swimming pool. For each 200-m swim, the swimmers started with a push start and were instructed to complete 7 evenly paced swims on either a 5-minute cycle for freestyle and backstroke or a 6-minute cycle for breaststroke and butterfly. Swimmers were instructed to swim the last 200 m with maximal effort. Ten seconds were added to each swimmer's personal 200-m performance of the year for the sixth swim, and then, working in reverse order, 5 seconds were added for each subsequent interval to establish the full test protocol. Standardized procedures and guidelines were followed regarding participation in the test, nutrition, training before each test, and warming up. For the case study, the performance of 1 swimmer was examined across 4 different tests, performed every 4 or 5 weeks, during a 20-week period. Additionally, training programs and competition performance were recorded. The research protocol and the methods used in this test were approved by the Leeds Metropolitan University Ethics Committee. Written informed consent was obtained for each swimmer before each test.
Stroke rate (cycles·min−1) was measured with the use of portable SR counters while the swimmers were swimming in the midsection of the pool. Three stroke cycles were used for each SR measurement for each 50 m, and the overall 200-m SR value was the average of the 4 measurements (i.e., 1 for each 50 m). The total and split times for each 50 m were recorded with the use of portable stopwatches. Average V (m·s−1) was calculated by dividing the total distance (200 m) by the final time (s). By using the calculated values for SR and V, SL (m·cycle−1) was estimated by dividing the average V by the average SR (SR/60). After each 200-m swim, 10-μL blood samples were taken from a hyperemic earlobe and analyzed immediately with a Lactate Pro lactate analyzer (Arkay Factory Inc, Shiga, Japan).
Estimation of Errors in Stroke Kinematics Calculations
The methods used for the calculation of stroke kinematics in the current study have been previously validated against a reference method (7). Although these methods are possibly subject to some errors in measurements, East (10) indicated that such errors occur in a systematic manner and do not greatly influence comparisons of stroke kinematics in different groups of swimmers. Therefore, in accordance with other studies in this area (9,19) and based on the analysis of East's data, it was decided not to attempt correction of the stroke kinematics data. Nevertheless, for the purpose of obtaining an indication of the approximate magnitude of errors that might have occurred in the current study, the performance of 1 swimmer was recorded on a VHS video camera with a sampling rate of 50 Hz. The camera was placed on a rigid tripod on the poolside, approximately 20 m away from the participant and 3 m above the pool level, and an 18-m scale line served as calibration. SR, SL, and V were calculated from the recorded video files following the guidelines for reference methods provided in the literature.
The differences between the practical values, using methods employed in this study, and the criterion values, using reference methods, were calculated and used to establish the validity and reliability of the methods used in the current study. Validity and reliability were assessed using a Pearson product moment validity correlation coefficient (r) computed between the practical and criterion measures; the standard error of measurement (SEM) (i.e., SD of the differences between the study and criterion measures); the coefficient of variation (i.e., ratio of the SEM to the mean of the values measured multiplied by 100 to express it as a percentage); and the systematic bias (i.e., mean of the differences between the study and criterion measures) ±1.96 × SD of the differences to give the limits of agreement according to the methods of Bland and Altman (3). The values for all calculations are shown in Table 1.
All the validity correlations showed very strong agreement between practical and criterion measures for all variables (P < 0.001). The SEM and coefficients of variation were small in magnitude, indicating, in both absolute and relative terms, that the variability between the practical and criterion measures was small and acceptable in terms of the precision associated with previously published research (7). The limits of agreement values revealed small systematic bias and small 95% confidence intervals defining the limits of agreement, again supporting the consistency between the practical and criterion measures.
A Pearson product moment correlation was used to determine the relationships between the pairs of variables (i.e., SR and SL; SR and V; SL and V; and V and BLa) for all strokes and for both genders. Significance was accepted at the p < 0.01 level. The magnitude of variation in SR and SL with V was also evaluated with a repeated-measures analysis of variance (ANOVA), using speed as a within-subjects factor and sex as a group factor. The results are presented in graphs and tables and expressed as mean ± SD.
Relationships Between Parameters for All Strokes
Swimmers generally increased V by increasing SR and decreasing SL, regardless of stroke or gender. BLa increased with V, as expected. However, deviations from this pattern were observed for SL in a few women freestyle swimmers. Statistically significant correlations were found for all strokes and both genders, between all pairs of variables, with the exception of SL and V for women's freestyle and breaststroke. Table 2 shows the correlation coefficients for the pairs of variables. Figures 1 and 2 show the changes of SR and SL in relation to V and BLa, respectively, for men and women freestyle swimmers.
As shown in Figure 1, the changes in SR and SL, with the exception of SL for women's freestyle, occurred in a linear manner in relation to V. SR increased significantly with V in both men and women (p < 0.001, ANOVA), with women having a significantly greater SR (p < 0.001, ANOVA) than men for any common speed of swimming. As expected, there was a significant decrease in SL for men and women (p < 0.001, ANOVA), although the relationship between SL and V was not so consistent. Similar trends were observed for backstroke, breaststroke, and butterfly. Men reached higher velocities by using either lower or similar SR, but higher SL compared to women. Moreover, similar velocities were produced by different combinations of SR and SL, with the combination of stroke kinematics for reaching high V varying among swimmers. The fastest times, reached in the final stage of the test, were on average 6.7 ± 2.5 seconds or 5.6% ± 2.2% slower compared to the swimmers' best competition performance of the year, with the magnitude of variation not being related to gender or stroke. Considering that swimmers were instructed to swim the last 200 m at maximal effort and that, in most cases, the personal best performance of the year was slower than the personal record of each swimmer, swimmers generally failed to reach the expected performance level in the late stages of the test.
As shown in Figure 2, SR increased and SL decreased with increasing BLa for freestyle swimmers, with the exception of SL for women. However, the changes in stroke kinematics in relation to BLa were higher at the early stages of the test, when BLa was low (1-3.5 mmol·L−1) and, in some cases, remained constant. Moreover, the changes in SR and SL did not seem to be affected by the elevation of BLa during the later stages of the test. The observations were similar for the other strokes. Drawing a vertical line at 4 mmol·L−1 on Figure 2 allows a comparison of the general OBLA values and shows that on average SR was 2 cycles·s−1 higher at OBLA in women than in men, whereas SL at OBLA was 0.35 m shorter for women than for men. Velocity at OBLA was on average 0.19 m·s−1 slower for the women than for the men freestyle swimmers, but was not defined by any specific change in the general relationships between SL and SR with V.
Case Study of Longitudinal Testing
Figures 3 and 4 show the changes in SR and SL, respectively, in relation to V, across the 4 tests for the single swimmer. The results were similar to those observed for the group, with significant correlations between all pairs of variables (i.e., SR and SL; SR and V; SL and V; and V and BLa) (p < 0.1). The swimmer increased V by increasing SR and decreasing SL, while BLa increased. Peak BLa values ranged from 10.8 to 15.7 mmol·L−1. In all tests, changes in stroke kinematics in relation to BLa were higher at the early stages of the tests, when the BLa values were low (1-3.5 mmol·L−1). Velocity and SL were significantly higher during the last test (p < 0.1) compared to the first three, while no difference was found for SR.
However, the swimmer was expected to have the best performance in the second test, which followed a period of tapering, compared to the rest of the tests, which followed periods of intense training. The best competition performance of the year occurred before the second test, as expected. The results also showed that the swimmer reached similar performance levels between tests 1 and 2 by increasing SR and decreasing SL. This finding indicated that similar velocities can be produced by different combinations of SR and SL. However, similar velocities did not always correspond to similar BLa levels. The latter emphasizes the usefulness of the test for the purposes of identifying the SR and SL combination that is more economical in terms of energy expenditure, as indicated by BLa, for a given V. The best times achieved during the tests were on average 10.7 ± 0.8 seconds slower (9.1% ± 0.7%) than the best competition performance of this swimmer for the year, indicating that the swimmer failed to reach the expected performance level during the last stage of the tests.
Swimmers increased V by increasing SR and decreasing SL, while BLa increased. This finding was in agreement with other studies (17,24), which also showed statistically significant relationships between these variables. A possible explanation for the different SL pattern observed in women freestyle swimmers could be that at the early stages of the test, these swimmers altered their individual stroking strategy to minimize the energy expenditure while swimming at a slower speed. Nevertheless, such differences may occur during a test situation and have been reported in other studies (24).
Changes in SR and SL were found to occur rapidly in relation to BLa at the early stages of the test, whereas changes appeared to be relatively smaller when swimmers reached and passed a potential OBLA or lactate threshold (i.e., 4 mmol·L−1). Changes in SR and SL in the first stages of the test should be attributed to individual stroking patterns of swimmers, as the lactate levels were so low that they were unlikely to influence their stroke kinematics. In contrast, other studies (15-17,24) have found relatively linear changes in SR and SL up to a lactate threshold or OBLA point, which, in some cases, became progressively greater for SL after that point. The level of the participants in the current study was higher compared to those in the aforementioned studies, and the swimmers were tested in their specialized strokes and event distances. These were possibly the main reasons that these swimmers sustained a greater decrease in their SL values when BLa increased. The latter underlines the importance of testing swimmers in their specialist strokes and event distances.
The results also showed that the combination of SR and SL in producing similar velocities varied within and between participants. This variation was expected and could be possibly explained by differences between swimmers in anthropometric characteristics, stroking technique, muscle flexibility, and coordination. Similarly, in agreement with other studies (14,18), the best combination of SR and SL for the production of the highest V varied between swimmers. Interestingly, as shown in the case study, the SR and SL combination for production of a given V may vary even within swimmers. This finding implies that the most appropriate combination for a swimmer may change when factors such as technique, body mass, strength, and physical condition change. It could also be possible that more than one combination of SR and SL is equally effective in achieving maximum V. Considering that different combinations of stroke kinematics produced similar velocities with different fatigue levels, as indicated by the BLa values, it could be concluded that repeated tests may enable the identification of the most economical SR and SL combination for a given V. Therefore, the SR and SL combinations for the V of interest should be reassessed on a regular basis for each athlete.
When assessing a swimmer's most effective SR and SL combination for maximizing performance, it is essential for the V reached during the swimming test to be similar to the competitive V. Consequently, the fact that international level swimmers failed to reach the expected V in the final stages of the test should be given careful consideration, because it indicates that the test is not suitable for providing such information that would allow the assessment of the best combination of SR and SL for maximal V production.
The higher SL and V and lower SR of men compared to women suggested that SL was the most important factor influencing performance. This finding was in agreement with most investigations carried out during competition (5,6), which reported that SL rather than SR was the significant contributor to higher V and identified SL as the dominant feature in achieving high performance levels.
The findings of the case study supported the results for the group. The changes in SR, SL, V, and BLa were similar across the 4 tests and in line with the trends observed for the group. The fact that the swimmer in the case study reached the highest V in the test that he used higher SL and lower or similar SR to the other tests underlined the importance of SL in achieving high V. Finally, the test did not directly reflect the overall training status and physical condition, as represented by competition performance, of the swimmer throughout the 20-week period. Similarly, Pyne et al. (20), after using the same test in a group of elite swimmers throughout a period of 8 months, reported that performance in the test did not seem to be directly associated with competition performance.
This study examined the relationships between kinematic (i.e., SR and SL) and physiological (i.e., BLa) characteristics and performance in elite swimmers tested in their specialist strokes and event distances. Velocity increased by increasing SR and decreasing SL. Similar velocities were produced with different combinations of SR and SL, while the best combination for reaching a given V varied both between and within swimmers. Stroke length appeared to be the most important parameter in achieving high V. Changes in SR and SL occurred rapidly in relation to BLa during the early stages of the test, when BLa levels were low (1-3.5 mmol·L−1). In most cases, swimmers failed to reach a V as high as expected toward the end of the test.
Given the importance of SL in performance improvement, it would be beneficial for swimmers to practice the lengthening of their stroke cycles during training and to try to resist the decrease in SL when V increases. The test was informative with regard to stroke kinematics combinations and relationships with BLa for slow to submaximal velocities, but not for maximal V. Therefore, repeated tests with the protocol used in this study could allow coaches to identify the most economical stroke kinematics combination for the velocities of interest and to focus on training drills that would enable swimmers to optimize the kinematic combinations for these velocities. However, for maximal V, a test that would enable swimmers to produce velocities close to those used during competition would be more informative for the purpose of assessing the most effective individual combination of SR and SL and would provide more useful information related to stroke kinematics and lactate tolerance of swimmers. Moreover, considering that test performance did not directly reflect competition performance, the test should not be used as a sole indicator of the training status or physical condition of a swimmer. Finally, the results emphasized that swimming tests should be individualized according to the specialist distance and stroke of each swimmer.
The authors gratefully acknowledge the swimmers, coaches, and scientific support team of the British Swimming Team who participated in this study. This project was supported by grants from the Greek State's Scholarships Foundation and by Leeds Metropolitan University.
1. Arellano, R, Brown, P, Cappaert, J, and Nelson, RC. Analysis of 50-, 100-, and 200-m freestyle swimmers at the 1992 Olympic Games. J Appl Biomech
10: 189-199, 1994.
2. Avlonitou, E. Maximal lactate
values following competitive performance varying according to age, sex and swimming
style. J Sports Med Phys Fitness
36: 24-30, 1996.
3. Bland, JM and Altman, DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet
8: 307-310, 1986.
4. Bonifazi, M, Martelli, G, Marugo, L, Sardella, F, and Carli, G. Blood lactate
accumulation in top level swimmers following competition. J Sports Med Phys Fitness
33: 13-18, 1993.
5. Chatard, JC, Caudal, N, Cossor, J, and Mason, B. Specific strategy for the medallists versus finalists and semi-finalists in the men's 200 m breaststroke at the Sydney Olympic Games. In: XIX International Symposium on Biomechanics in Sports: Proceedings of Swim Sessions
J.R. Blackwell and R.H. Sanders, eds. San Francisco, University of San Francisco; 2001. pp. 10-13.
6. Chengalur, SN and Brown, PL. An analysis of male and female swimmers in the 200-meter events. Can J Sport Sci
17: 104-109, 1992.
7. Chollet, D and Pelayo, P. Effects of different methodologies in calculating stroke length
. J Hum Mov Stud
36: 127-136, 1999.
8. Craig, AB and Pendergast, DR. Relationships of stroke rate
, distance per stroke, and velocity in competitive swimming
. Med Sci Sports Exerc
11: 278-283, 1979.
9. Craig, AB, Skehan, PL, Pawelczyk, JA, and Boomer, WL. Velocity, stroke rate
and distance per stroke during elite swimming
competition. Med Sci Sports Exerc
17: 625-634, 1985.
10. East, DJ. Swimming
: an analysis of stroke frequency, stroke length
and performance. N Z J Health Phys Educ Recr
3: 16-27, 1970.
11. Girold, S, Chatard, JC, Cossor, J, and Mason, B. Specific strategy for the medallists versus finalists and semi-finalists in the women's 200m freestyle at the Sydney Olympic Games. In: XIX International Symposium on Biomechanics in Sports: Proceedings of Swim Sessions
Blackwell, JR and Sanders, RH, eds. San Francisco, University of San Francisco; 2001. pp. 61-64.
12. Grimston, SK and Hay, JG. Relationships among anthropometric and stroking characteristics of college swimmers. Med Sci Sports Exerc
18: 60-68, 1986.
13. Kelly, M, Gibney, G, Mullins, J, Ward, T, Donne, B, and O'Brien, M. A study of blood lactate
profiles across different swimming
strokes. In: Biomechanics and Medicine in Swimming. Swimming Science VI
MacLaren, D, Lees, A, and Reilly, T, eds. London: E. & F.N. Spon; 1992. pp. 227-234.
14. Kennedy, P, Brown, P, Chengalur, SN, and Nelson, RC. Analysis of male and female Olympic swimmers in the 100-meters events. Int J Sport Biomech
6: 187-197, 1990.
15. Keskinen, KL and Komi, PV. Interaction between aerobic/anaerobic loading and Biomechanical performance in freestyle swimming
. In: Swimming Science V
Ungerechts, BE, Wilke, K, and Reischle, K, eds. Champaign, IL: Human Kinetics Publishers, 1988. pp. 285-293.
16. Keskinen, KL and Komi, PV. The stroking characteristics in four different exercises in freestyle swimming
. In: Biomechanics XI-B
De Groot, G, Hollander, AP, Huijing, PA, and Van Ingen Schenau, GJ, eds. Amsterdam: Free University Press; 1988. pp. 839-843.
17. Keskinen, KL and Komi, PV. Stroking characteristics of front crawl swimming
during exercise. J Appl Biomech
9: 219-226, 1993.
18. Pai, YC, James, HG, and Wilson, BD. Stroking techniques of elite swimmers. J Sports Sci
2: 225-239, 1984.
19. Pelayo, P, Chollet, D, Sidney, M, and Tourny, C. Stroking characteristics in freestyle during elite swimming
competition. In: XIVth Congress of the International Society of Biomechanics
International Society of Biomechanics
, ed. Paris, International Society of Biomechanics
; 1993. pp. 1018-1019.
20. Pyne, DB, Lee, H, and Swanwick, KM. Monitoring the lactate
threshold in world-ranked swimmers. Med Sci Sports Exerc
33: 291-297, 2001.
21. Robergs, RA, Ghiasvand, F, and Parker, D. Biochemistry of exercise-induced metabolic acidosis. Am J Physiol Regul Integr Comp Physiol
287: R502-R516, 2004
22. Telford, RD, Hahn, AG, Catchpole, EA, Parker, AR, and Sweetenham, WF. Post competition blood lactate
concentration in highly ranked Australian swimmers. In: Swimming Science V
Ungerechts, BE, Wilke, K, and Reischle, K, eds. Champaign, IL: Human Kinetics Publishers; 1988. pp. 277-283.
23. Wakayoshi, K, Tatesada, E, Ono, K, Terada, A, and Ogita, F. Blood lactate
response to various combinations of swimming
velocity and rest period of interval training. In: Biomechanics and Medicine in Swimming VIII
Keskinen, KL, Komi, PV, and Hollander, AP, eds. Jyvaskyla, Finland: University of Jyvaskyla; 1999. pp. 401-406.
24. Weiss, M, Reischle, K, Bouws, N, Simon, G, and Weicker, H. Relationship of blood lactate
accumulation to stroke rate
and distance per stroke in top female swimmers. In: Swimming Science V
Ungerechts, BE, Wilke, K, and Reischle, K, eds. Champaign, IL: Human Kinetics Publishers; 1988. pp. 295-303.
25. Weng, Q, Chen, Y, Niu, Y, and Wu, Z. The application of blood lactate
testing in swimming
training. Sports Sci (Beijing)
8: 47-51, 1988.
Keywords:© 2008 National Strength and Conditioning Association
swimming; biomechanics; stroke rate; stroke length; lactate; exercise test