Secondary Logo

Journal Logo

Original Research

Testing and Training of the Eggbeater Kick Movement in Water Polo

Applicability of a New Method

Melchiorri, Giovanni; Viero, Valerio; Triossi, Tamara; Tancredi, Virginia; Galvani, Christel; Bonifazi, Marco

Author Information
Journal of Strength and Conditioning Research: October 2015 - Volume 29 - Issue 10 - p 2758-2764
doi: 10.1519/JSC.0000000000000946
  • Free



Water polo is a highly demanding sport with intense bursts of activity (17) and characterized by complex movements: swimming at high speeds for short durations even at maximum speed, fast, and rapid actions with pronounced counterattacks (9,17), and carried out with ball possession and in contact with the opponent; frequent passing actions of the ball to other players and strong and precise shoots at the goal; performing high-intensity nonswimming activities, including wrestling and blocking, to gain or maintain a position (8,18).

Many of the technical actions (shooting, passing) and the contacts with the opponent take place in quasi-vertical floating position and with hands out of the water (11). In water polo, lower limb muscle performance is therefore essential for the maintenance of the quasi-vertical floating position in individual technical actions, in wrestling and blocking actions, and also for tasks optimization of both attack and defense type (17). It has been demonstrated that players with higher levels of lower limb strength are able to generate greater elevation out of the water, hence throwing the ball faster (6). Moreover, it has been underlined that during both phases in the match, attacking and defending, characterized by the attempt to get closer to the goal (typical of the attacking player) and the opposite action of trying to distance the opponent from the goal (typical of the defending player) are performed using whole body muscles, but especially lower limb muscle (17). Two types of lower limb muscle actions can be identified: the eggbeater kick (cyclical movement) and the breaststroke kick (ballistic movement) (11,13). During technical actions, these movements are very often performed alternating or simultaneously. The ballistic movement is used most often in jumping and throwing, whereas the cyclical movement in fighting and passing. An “eggbeater kick” consists of alternating the circular and continuous movements of the legs, producing an upward force and maintaining players afloat in a vertical position. Upper limbs are kept free, giving the opportunity to do technical movements with or without the ball (passing, throwing, tackling an opponent, wrestling, catching or intercepting passes, and blocking shots on goal) remaining vertical or moving in any direction while in a vertical position (20).

In the current literature, few works have studied the performance of lower limb muscle in water polo, although many studies have underlined the importance of developing maximal dynamic force and power (4,11,12,20). Tan in 2009 published a study (19) focused mainly on players' fitness characteristics, including lower-body muscular power, verifying that some of these parameters can discriminate between players of different competition levels and playing positions. Some other authors studied vertical jumping ability related to performance (6,11), taking into consideration the vertical breaststroke ballistic component. Because Sanders' studies (13,14) aimed to investigate technique in the eggbeater kick through kinematics of the lower limbs, no other authors addressed the performance of lower limb muscle during the eggbeater kick in water polo until the research by Uljevic et al. in 2013 (20). In this study, Uljevic et al. examined a test of the dynamometric force produced while using the eggbeater kick, consisting of maximum intensity swimming using an eggbeater kick with a fast elastic line fixed to a special belt (tensiometric dynamometer). Another study (5) has applied the maximal dynamometric force in the eggbeater kick to junior water polo players. In our knowledge, no studies have considered a method aimed to evaluate and to train the ability to produce an upward force in this specific movement. Assessment and evaluation are crucial for an evidence-based training process, to plan tailored training sessions, to organize periodization improving training effectiveness, and to control for its effect (21). According to the specificity criteria, a functional test should be sport-specific, simulating real-game situations and functional movements performed during training sessions (7,16).

The practical question of the study was, therefore, to try to identify a nonexpensive and simple method to better evaluate and train the eggbeater kick. Thus, the aim of this work was to study, using an observational study design, the applicability, in terms of repeatability, sensitivity, and compliance, of a new method for the specific evaluation and training of the eggbeater kick movement in water polo players of different competitive levels.


Experimental Approach to the Problem

The observational study was developed in 3 different days separated by 7 days of recovery. On the first day, standard anthropometry (height, body weight) and a clinical history were collected. On the same day, subjects performed some tests in the water with a jacket allowing the application of an overload submerged in water (weighted jacket) to get familiarized with the method. On the second day, after a standardized warm-up, the test with increasing loads was performed. The standardized warm-up lasted about 30 minutes and consisted of 10 minutes of dry land mobility exercises and stretching, 400 m freestyle at moderate intensity, 4 × 25 m freestyle and breaststroke-only leg kicking, and 10 minutes of specific technical exercises for lower limbs including eggbeater kick movements. On the third day, the test with the constant load was completed to study the repeatability of the previous measurement.

The in-water measurements took place in December during the competitive season for all athletes and was performed in the same swimming pool with the same water conditions (temperature: 27° C, ph: 0.9–1.1) and same environmental conditions (temperature: 24° C, humidity: 70%).


Forty-two male water polo players voluntarily participated in the study (age = 23.4 ± 3.5 years; height = 185.8 ± 7.5 cm; body weight = 81.8 ± 9.8 kg; body mass index [BMI] = 23.7 ± 2.5 kg·m−2). Twenty athletes were water polo amateur players (nonelite subjects, NES), whereas 22 athletes played water polo at national and international level (elite subjects, ES). Table 1 summarizes the mean values (±SD) of age, body weight, height, and BMI of the players displayed by subgroups. Body mass index was calculated as weight divided by height squared.

Table 1
Table 1:
Subjects' characteristics.

The NES subgroup trained 3 times a week and played a match at the weekend for a total of 7-hour weekly training/playing in the water, whereas the ES group had 8 training sessions (6 in the water plus 2 at the gym) and also played a match every week for a total of 15 hours per week.

The participants read and signed statements of informed consent. Our sample did not include anyone who was younger than 18 years (minimum: 18 years, maximum: 31 years); thus, no parental or guardian consent was required. The research was approved by the Ethic Committee of the Swimming National Federation and performed in accordance with the Helsinki Treaty rules.

None of the athletes involved in the research had suffered from injuries to lower limbs in the past 4 months. Two athletes were excluded from the study because of the presence of other disorders not referred to musculoskeletal conditions that could affect the results of the tests.


In-Water Measurements

A jacket allowing the application of an overload submerged in water called water polo overload testing/training (WOT) was used (Figure 1). The jacket was assembled to be adaptable to fit the body size of the athlete. It was homemade and created by joining 5-cm-width seat belts and using some sewing thread specific for chlorinated water. It was possible to obtain a versatile, robust, and long-lasting jacket. Two parallel bands 100 cm long were attached at a distance of 20 cm from each other. Four transverse bands were sewn, 2 at the ends and 2 at intermediate distances (1 at chest height and 1 behind the shoulders) with enough space left for the introduction of the head. Two lateral bands, 2-cm width, and buckle fasteners on the waist were added to guarantee the highest adherence. In front and behind the body, at the end of the two 100-cm long bands, 4 other bands, 2.5-cm width and 80-cm long, were attached. They joined in a stainless snap hook where the overload was applied. Thus, during the test, the overload lied down between the legs of the subject without hindering the natural “eggbeater” movement (Figure 1). The jacket, being properly secure to your body, did not hinder breathing or mobility.

Figure 1
Figure 1:
Water polo overload testing, front and back view.

The test consisted of different trials until exhaustion with increasing overload from 5 to 25 kg, with a recovery of 20 minutes between each trial. The imposed external loads progression was the same for both groups, and time to exhaustion (EXT) was registered for each load.

Time to exhaustion was the total time the athlete was able to maintain the fundamental floating position (acromion out of the water). The fundamental floating position was performed with arms crossed at the wrists holding them against the chest. The overload estimated maximum value (OEMV) was calculated using a line of best fit for data and its equation (Figure 2). A video camera was placed facing the athlete and each trail was recorded while continuous timing was taken with a stopwatch.

Figure 2
Figure 2:
Differences in time to exhaustion data between elite subjects (ES) and nonelite subjects (NES). Overload is in relation to body weight. The equation to calculate overload estimated maximum value is reported.

Considering former experiences of similar overload training in water with water polo players (3), a maximum of 9 tests with increasing loads between 5 and 25 kg was planned. After preliminary tests, a protocol including more than 5 tests was considered more suitable to assess and compare athletes of different performance levels. The NES group performed 6 tests with increasing load (5, 7.5, 10, 12.5, 15, and 17.5 kg). In this group, the procedure was stopped at the sixth load for the inability of the athletes to maintain the fundamental floating position. The ES group performed 8 tests with increasing loads (5, 7.5, 10, 12.5, 15, 17.5, 20, and 22.5 kg).

The recorded data were reported in a dedicated database and plotted on a diagram in which the applied overload was on the x-axis (independent variable) and EXT was considered as the dependent variable and was on the y-axis.

Repeatability of the Measurement

The test with constant load was performed only by the athletes of the elite group because of their greater willingness to undergo a further test. Fifteen athletes of the ES group were evaluated. They performed a test with a fixed load (15 kg) at the fundamental floating position until exhaustion. The test was repeated 4 times with complete recovery (20 minutes) between the trials. After the test (12, 24, and 48 hours later), the athletes were interviewed by telephone to verify the onset of symptoms that could indicate the occurrence of musculoskeletal injuries and the necessity to undergo a medical examination.

Sensitivity and Specificity

The effectiveness of the different measured variables in both subgroups (EXT at 5, 7.5, 10, 12.5, 15, and 17.5 kg of increasing loads) and that of the OEMV and cutoff value for each variable were evaluated with receiver operating characteristic (ROC) curve analysis. Sensitivity and specificity were used to describe the utility of test in athlete's evaluation.

Statistical Analyses

The results were expressed as mean ± SD. Percentage values were reported without decimal places (sample size is less than 100).

Normal distribution was verified using normality plot, the Shapiro-Wilk test, and Kolmogorov-Smirnov test. Levene's test was used to inspect the homogeneity of variance. Logarithmic transformation was used to transform a skewed distribution to normality. The Pearson's correlation test was used to analyze the correlation between continuous outcome and explanatory variable. T-test was used to evaluate the difference between variables. Effect size was used to study the size of difference in mean value.

The curve fitting process was used to construct the curve that has the best fit to the series of data points and then calculate maximal test value. Receiver operating characteristic curve was used to find a cutoff value that delineates a “low” from “high” test result. In relation to cutoff value (optimal value), sensitivity and specificity of the test were calculated. Intraobserver variation was studied and reported as intraclass correlation (ICC) and 95% confidence interval (95% CI) (lower bound and upper bound) (10). Post hoc evaluation of research sample size and power of statistic was calculated as described by Cohen (2).

SPSS 20 was used to perform statistical analysis. The p ≤ 0.05 criterion was used for establishing statistical significance.


Some differences in height and body weight were found between ES and NES groups. Elite subjects were significantly (p = 0.002) taller and had a significantly higher (p = 0.03) body weight than NES. A 4% difference and an 8% difference were found for height and body weight, respectively (Table 1).

In-water measurement results were reported in absolute (Table 2) and relative (in relation to body weight) (Figure 2) values to verify whether the different body size could influence the performance in the tests.

Table 2
Table 2:
Differences in time to exhaustion data between ES and NES.*†

Table 2 describes the results obtained by performing the test with increasing loads. Significant (p < 0.01) differences among players of ES and NES were observed in any imposed overloads. The effect size describes the magnitude of the difference between 2 groups and in this case showed a “large effect” of different competitive level for all measurements, as Cohen's effect size benchmarks are small = 0.20, medium = 0.50, and large = 0.80 (2). The OEMV, calculated using the equation reported in Figure 2 accounting a time of 2 seconds, was 14.8 ± 2.1 for NES and 24.2 ± 3.5 for ES, respectively (effect size = 3.5; p = 0.0002).


The ICC for single measurement was 0.88, with a 95% CI from 0.66 to 0.98. A significant correlation (r = 0.94; p = 0.05) between items was found. All values demonstrated a good repeatability of the test.

An average difference of 1.9 seconds (1.3–2.1 seconds) between T1 and T4 was detected with a coefficient of variability of 5.5% (3.5%–6.5%).

Sensitivity and Specificity of the Test

The area under the ROC curve for the 4 exhaustion times and for OEMV value gives measurement in predicting the investigated variable. The greater area (0.90; lower bound 0.89; upper bound 0.96) was found for OEMV.

Areas under the ROC curve for the different overloads were 0.72 (0.53–0.92) for 5 kg overload, 0.80 (0.68–0.90) for 7.5 kg overload, 0.87 (0.77–0.91) for 10 kg overload, and 0.88 (0.84–0.92) for 12.5 kg overload, respectively. Overload estimated maximum value sensitivity and specificity are presented in Table 3.

Table 3
Table 3:
Overload estimated maximum value sensitivity and specificity.*


All testing sessions were completed by each athlete, and no athlete has interrupted testing procedures. Moreover, no musculoskeletal injuries occurred or were reported during testing sessions or during the following 3 days.

No significant differences were found between video camera and stopwatch EXT assessment (effect size: 0.1; p = 0.11; r = 0.90).

Power analysis for OEMV showed a good confidence and precision. Difference between groups is 9.4 ± 1.1 kg (95% CI: 8.2–10.5). Setting power to 0.90 and considering difference measured in all variables (EXT and OEMV), our sample size is for both groups 3 times more numerous than observations needed.


The applicability of a new method for the specific testing and training of muscle performance in water polo players of different competitive levels was studied in terms of repeatability, sensitivity, and compliance. Repeatability assessed how consistent the test was. Our results showed that no measurement error can affect the evaluation of differences in performance. Data on sensitivity displayed that the proposed test is sensitive and specific enough to measure water polo players' lower limb muscle performance. Compliance analysis indicated that the test was well tolerated by all athletes and could be performed without any risk of injury. This approach attempted to identify and demonstrate to coaches the ability of the test to describe water polo players' lower limb performance, and at the same time, to propose the exercise mode used for the test also as an exercise workout because of its high specificity. Actually, our data showed that the test can be considered a new method for the specific evaluation and training of the eggbeater kick movement in water polo players.

Water polo players of different competitive level (ES and NES) were recruited. The choice of 2 groups of different competitive level was performed to better study test validity. As expected, body weight, height, and lower limb muscle performance of the elite and nonelite players were significantly different from each other (Table 1). Effect size described the magnitude of difference between ES and NES groups in the different measurements performed with WOT (Table 2). The effect size can be considered “large” because it is always greater than 2 (large ≥0.80) (2). The greater amount of training (7 hours per week for NES compared with 15 hours per week for ES) can partially explain the differences between the 2 groups. Moreover because of the complexity of the eggbeater action, which is technically a very skillful action (13,14), the different performance of the test can be explained also by the greater technical skill of ES compared with NES (11).

As suggested by Meckel (7) and Sattler (16), a functional test should be sport-specific, simulating real-game situations. The eggbeater action used during the test is a specific movement performed by water polo players, and subjects participating in the study were accustomed to training this kick for several hours during each week using different types of overload: a weight belt commonly used for scuba diving, a medicine ball held in hands, elastic bands, etc. In general, 3 types of lower limb muscle actions can be identified for water polo players: ballistic movements, breaststroke and sidestroke (scissor) kicks, and the eggbeater kick. The eggbeater is the most important kick for water polo players. It provides the base of support for vertical passing and shooting and must be continually conditioned. The eggbeater kick movement is an alternating rotating kick to maintain body height and position. The hips start in a position close to 80 degrees of flexion and 90 degrees of abduction, with close to 30 degrees of lateral rotation. The knee is flexed close to 15° and laterally rotated at the start of the kick. During the kick, the hips and knees are extended, adducted, and medially rotated to produce power in the stroke. Propulsion occurs due to hydrodynamic lift forces and drag forces that are created by the rapid downward and inward movements of the foot and leg during the stroke (15). Technical skills and muscle performance are required to maximize the ability of efficiently sculling the water using the eggbeater kick.

Unfortunately, it is difficult to compare our results with other previously published results. Some authors have analyzed the action of the lower limbs in water polo underlying their importance to generate greater elevation out of the water, improving vertical jump capacity, but they have focused mainly on ballistic movements (4,6,11). Less extensive is the knowledge about eggbeater kick. Uljevic et al. (20) studied dynamometric force in a forward movement momentum (athletes were in the breaststroke position during the test). This specific movement is typical to some playing positions. In contrast, a sport-specific movement in water polo for all players, independent of their playing position, must generate an upward force in the water, keeping the player afloat in a vertical position, being able to stay up for very long, during a short action or for an entire game. For this purpose, this study has examined this specific capacity to achieve the maximal possible dragging force in water polo players performing the eggbeater kick, developing a new specific test and using a new specific jacket.

It was necessary to construct a new jacket meeting our needs because of previous negative experiences with other overload systems. The available jackets had part of the overload immersed and part of it was above the waterline; besides, the overload position could vary during the test itself. Athletes reported difficulties maintaining appropriate positioning and stability of the trunk and, with increasing load, a compressive effect of the jacket resulting in respiratory problems. The particular shape of these jackets hindered the natural movement of the hip and moreover did not allow a sufficient overload for the test of lower limb muscle performance.

The proposed test for lower limb muscle performance can be conducted with a stopwatch (a more ecological condition) and using video camera (a more technological and expensive methodology). The choice of the materials and methods of the test allows its utilization for both evaluation and training. Taking into account the results of the OEMV, it is possible to plan different training sessions, customizing the external workload, the training, and the recovery times and intensity levels, according to different aims: endurance, power, and strength. A specific training of the eggbeater kick could enhance the performance of lower limb muscle in the vertical position. The athlete would be able to maintain the position of the body with the head above the water and to raise the body higher out of the water for a longer time and with less fatigue; during a game, water polo players are required to raise the upper body for several seconds during both offensive and defensive actions.

With respect to the repeatability of the test, our results showed a good repeatability with a relatively good agreement displayed. The lower limb muscle performance can vary according to training, and our test can be used to assist with designing training sessions and controlling training effects.

Athletes performed the test to exhaustion level 4 times the same day with complete recovery (20 minutes) between the trials to check if learning effect or fatigue could influence data output. Differences in EXT were small between trials, although EXT values slightly decreased from T1 to T4. According to these data, both learning effect and the impact of fatigue could be considered negligible. However, further between-day (2 sessions on separate days) studies are needed to deeper investigate test-retest reliability.

The high intraoperator repeatability detected in this study indicated that the test for lower limb muscle performance could accurately be used to control longitudinal adaptation during a competitive season and during a multiyear training periodization.

Our results showed that the test for the evaluation of the performance of lower limb muscle was sensitive enough. The preferred value to take into consideration for lower limb muscle performance is the OEMV when compared with the measured variables (EXT at 5, 7.5, 10, and 12.5 kg) as showed by the greater area under the ROC curve (see Results). The OEMV presented good sensitivity and specificity, and the positive likelihood ratio showed a small and rarely important conclusive change from pretest to posttest probability. Small and rarely important effect indicated the utility of the test in identifying the level of players (elite and nonelite). The validity of a test refers to whether the test measures whatever it is supposed to. The ability of WOT to predict the agonist level of players was confirmed by sensibility and specific value. In this case, predictive validity (sometimes called empirical validity) referred to a test's ability to predict a relevant behavior that can be positively considered for WOT. According to the results of this study, an initial workload of 5–6% of the body weight is suggested, with a maximum increase in workload of 20–21% of the body weight.

All players, independently of their competitive level, correctly performed eggbeater movements, which they were well accustomed to, even with overloads. In fact, none of the players involved in the study and belonging to elite or nonelite group reported injuries 48–72 hours after the test session. Besides, none of the players interrupted the measurements, showing a good compliance with testing procedure.

In conclusion, the test reported in this study can be applied to investigate lower limb muscle performance in water polo players, even of different competitive level, and the procedure can be also used to train the eggbeater kick movement in a specific way.

However, in highly trained players, the testing procedure was very demanding and time consuming; therefore, future studies should consider different protocols to optimize test duration. Future studies are also required to further demonstrate the repeatability, sensitivity, and compliance of the test by examining players of different gender, pretraining and posttraining interventions, etc.

Practical Applications

This study analyzed a new in-water test for lower limb muscle by means of an important skill in water polo, used by the players to keep them afloat in an upright position and showed its applicability for water polo players of different competitive levels. The test was valid in terms of sensitivity and repeatability. The test was well tolerated and water polo–specific. Besides, the test is simple to administer and does not require expensive or intrusive instrumentations. The technical exercise performed during the test, the eggbeater kick movement, is very specific and therefore can also be used for training purposes (endurance, power, strength, etc.) to maximize water polo players' performance. Therefore, it can be easily used by trainers to plan tailored training sessions, to control adaptations, and to prevent injuries. This test takes into account the complex nature of water polo to ensure accurate and reliable information for the coach, as suggested by Bampouras and Marrin (1). Indeed, coaches may use this test to evaluate adaptations to training, to plan long-term training in young water polo players, or to assess recovery from injuries. It is a real training tool, and this is the reason why we consider WOT “a new method.”

Significant differences in some anthropometric and test data between elite players and nonelite players have been found and demonstrate the need to consider these physical and physiological characteristics when evaluating players and designing appropriate training programs.


We thank Anthea Marks for her linguistic review. The results of this study do not constitute endorsement of the product by the authors or the National Strength and Conditioning Association.


1. Bampouras TM, Marrin K. Comparison of two anaerobic water polo-specific tests with the Wingate test. J Strength Cond Res 23: 336–340, 2009.
2. Cohen J. A power primer. Psychol Bull 112: 155–159, 1992.
3. Dopsaj M. Testing technology for the elite water polo players. In: Proceeding of the 1er Congrès Mondial de Natation I.N.S.E.P. 17–21 Mai 2005, Paris FRA.
4. Gobbi M, D'Ercole C, D'Ercole A, Gobbi F. The components of the jumps in expert and intermediate water polo players. J Strength Cond Res 27: 2685–2689, 2013.
5. Idrizović K, Uljević O, Ban D, Spasić M, Rausavljević N. Sport-specific and anthropometric factors of quality in junior male water polo players. Coll Antropol 37: 1261–1266, 2013.
6. McCluskey L, Lynskey S, Leung CK, Woodhouse D, Briffa K, Hopper D. Throwing velocity and jump height in female water polo players: Performance predictors. J Sci Med Sport 13: 236–240, 2010.
7. Meckel Y, Machnai O, Eliakim A. Relationship among repeated sprint tests, aerobic fitness, and anaerobic fitness in elite adolescent soccer players. J Strength Cond Res 23: 163–169, 2009.
8. Melchiorri G, Castagna C, Sorge R, Bonifazi M. Game activity and blood lactate in men's elite water-polo players. J Strength Cond Res 24: 2647–2651, 2010.
9. Melchiorri G, Padua E, Sardella F, Manzi V, Tancredi V, Bonifazi M. Physiological profile of water polo players in different competitive levels. J Sports Med Phys Fitness 50: 19–24, 2010.
10. Peat JK, Barton B. Medical Statistics: A Guide to Data Analysis and Critical Appraisal. Malden, MA: Blackwell Publishing Ltd., 2005.
11. Platanou T. On-water and dryland vertical jump in water polo players. J Sports Med Phys Fitness 45: 26–31, 2005.
12. Platanou T, Varamenti E. Relationships between anthropometric and physiological characteristics with throwing velocity and on water jump of female water polo players. J Sports Med Phys Fitness 51: 185–193, 2011.
13. Sanders RH. Analysis of the eggbeater kick used to maintain height in water polo. J Appl Biomech 15: 284–291, 1999.
14. Sanders RH. A model of kinematic variables determining height achieved in water polo boosts. J Appl Biomech 15: 270–283, 1999.
15. Sanders RH. Strength, flexibility and timing in the eggbeater kick. 2005. Available at: Accessed May 15, 2014.
16. Sattler T, Sekulic D, Hadzic V, Uljevic O, Dervisevic E. Vertical jumping tests in volleyball: Reliability, validity, and playing-position specifics. J Strength Cond Res 26: 1532–1538, 2012.
17. Smith HK. Applied physiology of water polo. Sports Med 26: 317–334, 1998.
18. Tan F, Polglaze T, Dawson B. Activity profiles and physical demands of elite women's water polo match play. J Sports Sci 27: 1095–1104, 2009.
19. Tan FH, Polglaze T, Dawson B, Cox G. Anthropometric and fitness characteristics of elite Australian female water polo players. J Strength Cond Res 23: 1530–1536, 2009.
20. Uljevic O, Spasic M, Sekulic D. Sport-specific motor fitness tests in water polo: Reliability, validity and playing position differences. J Sports Sci Med 12: 646–654, 2013.
21. Zatsiorsky VM, Kraemer WJ. Science and Practice of Strength Training. Champaign, IL: Human Kinetics, 2006.

exhaustion test; overload training; overload estimated maximum value; lower limb; performance

Copyright © 2015 by the National Strength & Conditioning Association.