A Novel Method of Anaerobic Performance Assessment in Swimming

Abstract

Smolka, L and Ochmann, B. A novel method of anaerobic performance assessment in swimming. J Strength Cond Res 27(2): 533–539, 2013—The purpose of this study was to develop a novel, anaerobic efficiency test in swimming. The concept of this test was build upon the classic Wingate anaerobic test, with the difference that the dependent variables are extracted from an instantaneous speed curve. Twelve highly trained male swimmers, 19.3 (±3.3) years of age, 1.84 (±0.08) m in height, and 77 (±12.8) kg in weight, participated in the study. The swimmers swam a 100-m race in a 25-m pool, following the instructions to attain maximal swimming speed as quickly as possible and to keep maximal speed as long as possible. The race was recorded by means of video cameras (50 samples per second). Raw data were filtered by means of standard mathematical functions: smoothing and a polynomial function fit. The following speed parameters: maximal speed (V_max), time to reach V_max

, time at V_max

, minimal speed (V_min), and fatigue index were extracted from the filtered speed curve. Thorough assessment of anaerobic performance in swimming requires a swimming test leading to development of a speed curve obtained during a maximal effort race and not race time alone.

Introduction

Investigating the effects of training is one of the most important aspects of the training process. The gold standard in evaluation of anaerobic efficiency in sport is the laboratory, ergometric, Wingate anaerobic test (WAnT) (^5,12). In this test, a power curve obtained during maximal intensity effort illustrates the mechanical-biochemical relationship of the muscles tested (⁶). The Wingate test is widely used in sports regardless of the athlete's specialization: running (^2,3,8,11,21), jumping (^1,15), ice hockey (²²), wrestling (¹⁶), and swimming (^4,13,14,20).

In swimming, the relationship between the WAnT and swimming time has been studied by several authors. In a study involving competitive swimmers, Hawley and Williams determined a highly positive correlation between peak and mean power of WAnT and the 50-m swim time (r = 0.82; r = 0.83, respectively) (¹³). In another experiment, Hawley confirmed the existence of a highly correlative relationship between Arm and Leg Wingate tests and 50-m swim time (r = 0.63; r = 0.76, respectively) (¹⁴). Moreover, a high correlation between 50- and 200-m sprint trials and mean leg and arm power in the WAnT was observed in young swimmers (⁹). However, owing to the richness of laboratory data, the universality of WAnT has been questioned (^{2,4,7,10,12,15}). No correlation between peak, mean power, and fatigue index (FI) in the Wingate arm crank test and crawl sprints over 14, 25, 50, and 400 m was found in a cohort of competitive swimmers (¹²). The same conclusions have been put forward by Bampouras and Kelly in the experiment conducted with UK women's water polo players; there was no correlation between any of the cycloergometric WAnT parameters (peak, mean power, FI) and maximal and mean speed of 14 25-m swim sprint trials (⁴). These conflicting data may be explained by some factors that influence correlational relationships between WAnT and sport-specific tests. These are the type of ergometer used (adaptive or nonadaptive) biomechanical parameters of the arm ergometer (height of the ergometer from the floor, the adequate adjustment of the handles to the height of the subject's shoulders, angle of elbow flexion, and extension, distance of the subject from the ergometer (¹²), posture of the subject's body during performance of the trial (¹²), bicycle frame geometry (¹⁹), and the subject's demographic and anthropometrics characteristics (¹²).

The obvious limitation of the use of the WAnT in evaluation anaerobic efficiency in swimmers is its nonwater environment. Ergometers impose a strict, nonecological for swimmers, movement pattern on the tested subject. The arm movements are confounded to a closed chain, almost only sagittal plane motion, barely mimicking the reach and retract motion in swimming. During swimming, the arm movements are not confounded to the end point, and more importantly, they deal with the surrounding environment of the water. Moreover, the speed of swimming depends also on the swimmer's technique and level of skill, which cannot be expressed in an environment other than swimming (²⁰). Thus, we argue that the test must correspond to a concrete execution of movement and match the water conditions (¹⁸). However, the construction of a sport-specific anaerobic test for swimming cannot rely solely on race time measurements. Akin to the WAnT, it is crucial to determine a profile of instantaneous swimming speed during a race. For example, 2 swimmers may reach the same race time, yet the trajectory of instantaneous velocities might be distinctly different from each other. It is the trajectory of change in power that provides for the diagnostic capacity of the WAnT. We purport that a comprehensive understanding of a swimmer's anaerobic efficiency may be gain by investigating the course of changes in speed curve.

Therefore, the purpose of this study was to develop a novel, sport-specific test to determine the anaerobic efficiency in swimmers inferred from changes in speed during maximal effort swimming test. The concept of this test was build upon the classic WAnT.

Methods

Experimental Approach to the Problem

A 100-m, maximal effort swimming races were recorded using five 50 frames per second video cameras located equidistant at the side of the swimming pool to allow for computation of speed data. Raw displacement data were further processed by means of mathematical functions using smoothing and polynomial function fit. The obtained speed in time curve was used to assess of speed parameters for each swimming race. The dependent variable was instantaneous swimming speed.

Subjects

Twelve healthy, highly trained male swimmers: 19.3 (±3.3) years of age, 1.84 (±0.08) m in height, and 77 (±12.8) kg in weight, took a part in this study. The subjects engaged in regular swimming training for 8.9 (±2.9) years. The swimmers were members of the city of Wroclaw (Poland) swimming team. Three of them were members of the National Polish Junior Swimming Team.

All the athletes were involved in an intense training program. They performed regular training sessions of 9 km·d⁻¹, 5 d·wk⁻¹, whereas once a week, they engaged in a stretching training program. All coaches, subjects, and their parents if required gave written informed consent for participation in this research. The study was also approved by the local ethics committee.

Procedures

The tests were conducted after 2 days of rest. All the subjects were well nourished and hydrated. No recent injury or pain episodes had been reported by the participating athletes.

The study was conducted in a 25-m swimming pool. After a warm-up session, including short sprint races, the swimmers rested until they gained a total subjective readiness to perform the maximal effort test. The swimmers lay on the water with legs held by a coach's assistant, close to the wall. On the sound of a whistle, and without pushing off the wall with their legs, the swimmers begun the race. The swimmers were instructed to attain maximal swimming speed as quickly as possible and to keep maximal speed as long as possible. At the end of the swimming pool, they were instructed to perform open turns (touching the wall with one hand, no flip turn). The race time was recorded by means of manual chronometer, from the moment of first whistle, until the swimmer completed the 100-m test signified by touching the wall with one hand.

The tests were recorded at a frequency of 50 frames per second, by means of 5 analog video cameras (GR-DVL9800, JVC, Yokohama, Japan). The cameras were placed every 5 m, next to, and along the swimming track. Before the tests, calibration markers (plastic balls located on the ropes) were placed every 1 m along the middle of the track and the length of the swimming pool, then recorded. The swimmers performed the races with a spherical marker (7 cm in diameter) painted onto their swimming caps.

The recorded video-material was analyzed by video analysis software (SIMI Motion2D 2001, Simi Reality Motion Systems GmbH, Unterschleissheim, Germany). Analysis consisted of identifying the marker placed in the swimmer's cap, in every frame of the video. Knowing the rate of video recording, the instantaneous swimming speed was calculated. Then, raw, nonfiltered velocity data were exported to data analysis software (OriginPro7.5, OriginLab, Northampton, MA, USA) and filtered in a standard way by means of smoothing (Fast Fourier Transformation) and polynomial function Fit.

Then following speed parameters have been assessed: maximal swimming speed (V_max), time to reach V_max

, time at V_max

(criterion of 5%·V_max decrease), minimal swimming speed (V_min) (pointed out at the 90th meter of swimming) and FI (percent decrease of V_max at the 90th meter of swimming:

were determined. Speed parameters, with reference to time, were assessed according to the equation of 2 frames = 0.02 seconds. Speed parameters are also presented with reference to distance (position of the swimmer's head on the swimming track): Distance covered when V_max was reached (D to V_max) and distance covered while keeping (D at V_max).

Statistical Analyses

All statistical tests were processed by means of Statistica software (Statistica 9, StatSoft, USA). The Shapiro-Wilk test was used to test the normality distribution. Significant relationships between selected parameters were identified using Pearson moment product correlation coefficient (r). Significance level was set at p ≤ 0.05.

Results

The Shapiro-Wilk test revealed that the data met criteria for normality distribution.

Table 1 gives numeric data of speed parameters for all the participants in this study. The results are displayed from the best to the poorest race time. Sample raw and filtered data of instantaneous swimming speed are shown in Figures 1A and B, respectively.

Table 1: Anaerobic test data and race time for each of the 12 study participants.*

Figure 1: Sample raw (A) and filtered (B) data of instantaneous swimming speed (meters per second) during a maximal effort 100-m swimming test. The test was performed in a 25-m swimming pool. Graphs represent subject 12.

The mean race time for all 12 subjects was 60.63 seconds. On average, the subject's results were V_max: 1.9 m·s⁻¹,

: 6.9 seconds,

: 3.5 seconds, V_min: 1.5 m·s⁻¹ and the FI amount was 20.1%.

Table 2 contains a correlation matrix among all speed parameters.

Table 2: Correlation matrix among all swimming speed test parameters.*†

Discussion

This study presents a conceptual argument for a novel, sport-specific test to determine the anaerobic efficiency in swimmers inferred from changes in speed during a maximal effort swimming test. Experimental data provide proof of concept: that is, the values extracted from the speed curve allow for better representation of anaerobic efficiency then race time alone. For example, as presented in Table 1, swimmers 9–11 reached the same race time, but their speed parameters are distinctly different from each other: V_max: 1,7; 1,82; 1,73 m·s⁻¹, time to reach V_max: 6,91; 7,74; 7,94 seconds, time at V_max: 3,06; 4,08; 4,21, V_min: 1,38; 1,25; 1,38 m·s⁻¹, respectively. These examples confirm our assumptions that the swimming test based merely on race time measurements does not provide sufficient data for anaerobic assessment. To gain a comprehensive insight into a swimmer's performance, it is necessary to determine instantaneous speed trajectory. Moreover, speed trajectory and speed parameters values are instrumental in determining a swimmer's training program. That is, the individual training program for the 3 subjects discussed above would differ from each other in dose of training.

The second main observation made in this study is that the swimming test of anaerobic performance must be based on instructions employed during the Wingate test. Specifically, a swimmer during the test must attain maximal swimming speed as quickly as possible and keep maximal speed as long as possible. Performance recorded as a result of this testing instruction will produce a distinctly shaped speed curve. This speed curve serves as a source of typical anaerobic efficiency parameters, such as maximal speed, time to reach maximal speed, and time at maximum speed. As can be seen in Figures 1 and 2, the instantaneous speed trajectories are similar in shape to those attained during a Wingate test.

Figure 2: The swimming speed curve obtained during a maximal effort 75-m swimming distance test, performed in 25-m swimming pool. Example of a swimmer's pretraining and posttraining speed curves. The training consisted of 6 glycolytic training sessions during a period of 2 weeks. Data have been filtered by means of fast Fourier transformation (FFT).

Although the velocity parameters are highly correlated with the race time, none of the velocity or temporal parameters determined the race time alone. As presented in Table 2, high correlation between race time and 3 parameters V_max, V_min,

were observed. This illustrates the point why the highest and the lowest values of speed parameters are not always attained by the swimmer with the shortest and the longest race time, respectively.

For example, the lowest value of V_max was determined by a swimmer with the ninth race time (not by the swimmer with the longest race time, as might had been expected), the shortest

was achieved by swimmer number 3 (not by the last swimmer, as might had been expected) and the longest

was recorded by the swimmer in the eighth race time (not by the best swimmer, as might had been expected), and so on. These examples illustrate that the speed curve provides an insight into the dominant variable influencing a swimmers performance. A future study, with a larger number of subjects may allow for development of a regression analysis and therefore better predictions of an individual swimmer's parameters on race time. This illustrates the limitations of race time alone.

The parameters extracted from the speed curve provide the coach with information allowing him to categorize a swimmer's current abilities as predominantly that of a sprinter or endurance swimmer. For example, it can be expected that, independently of race time, swimmer with better “sprinting” abilities would attain better values of V_max,

, and

, yet on the contrary—a swimmer with better “glycolytic” abilities would gain better values of V_min, and FI. For comparison, note data for swimmers 9 and 10 in Table 1.

The presented method of assessment of anaerobic performance has some methodological limitations and caveats, which need to be more widely discussed. First, this study was conducted in a 25-m swimming pool. Consequently, 3 turns were performed during a 100-m race producing the 3 characteristic modulations in speed presented on Figures 1A, B and 2. Turning allows for a slight rest and enables acceleration after a push-off from the wall with the legs. The speed curve would surely appear slightly different in a race performed in a 50-m pool. We therefore suggest that for proper comparisons, the tests need to be conducted in pool of the same length. Second, the swimmers were instructed to swim using the crawl technique regardless of their swimming expertise. We acknowledge that the crawl technique enables reaching of maximal stroke frequency and it is usually the fastest technique. Swimming with another technique does not elicit a swimmer's full potential to gain the best performance. The third of the methodological concerns is the length of the race. A typical Wingate ergometric test lasts 30 seconds (^5,6). It is known from experience that a 30-second swimming race is too short and consequently evaluation of the “glycolytic” potential is difficult to attain. After several pilot studies, we determined that the best test distances are either 75 or 100 m. The swimmers seem unable to undertake longer distances (such as 150 m) at 100% of effort and therefore tended to divide intensity along the distance of the race and not reach their true maximal speed. Fourth, after many attempts, with the use of various devices, such as: radar, ultrasound, accelerometer, and global positioning system (which proves to be quite useful outdoors) the single reliable method we found so far is the video method. That said, video analysis is time consuming, for example, time is required to digitize the head marker location of a swimmer in every frame. Furthermore, selecting the same pixel within the marker every time is impossible, which contributes to additional instantaneous variability. Improvement to video analysis is still ongoing; there are some possibilities to improve tracking by means of a special program based on neural networks (¹⁷).

In conclusion, we have demonstrated that to better understand anaerobic performance of swimmers, a specific swimming test cannot rely merely on race time measurements but must be based upon a swimming speed curve obtained during a maximal test. To obtain appropriate anaerobic data (graphic and numeric), it is necessary to perform test race following Wingate Test performing instructions. It has been observed that there is no speed parameter that individually determines race time and that the speed parameters (equally speed curve) might reflect current sprinting or glycolytic abilities of the swimmer.

Practical Applications

There are 3 practical aspects, identified by the content of this article. First, the presented test shows how to determine the swimming speed curve and to extract the necessary parameters, which are indispensable in determining a swimmer's performance.

Second, we suggest that the proposed test might be used to evaluate a swimmer's response to a given type of training. To illustrate the second point, we included an example of a swimmer's pretraining and posttraining speed curves (Figure 2). This swimmer's training consisted of 6 glycolytic training sessions, which resulted in a race time that improved by 3 seconds, despite the lowered V_max from 1.69 to 1.56 m·s⁻¹. Third, the swimming speed curve might be used to determine the training dose. For example, the time of drop in the speed curve, after the period of holding maximal speed, may guide the coach in setting the goals for training a swimmer's sprinting abilities.

Acknowledgments

The authors are grateful to the members of “Juvenia” Wroclaw Swimming Team and for participation in this study. All technical and equipment support have been provided by University School of Physical Education in Wroclaw.