The Relationship between Specific Game-Based and General Performance in Young Adult Elite Male Team Handball Players

Featured Application

The presented specific game-based and general testing concept can be applied to measure performance in elite team handball and to improve the physical performance of young adult elite team handball players.

Abstract

Physical performance is an essential factor for becoming a top elite team handball player; however, the relationship between specific and general physical performance is not well known. Consequently, the aim of the study was (1) to analyze the relationship between specific game-based and general physical performance in young elite male team handball players, and (2) to reduce the number of tests for a more practical implementation of physical performance diagnostics in team handball. Twenty young adult elite male team handball field players (18.6 ± 2.1 years) performed the team handball game-based performance test (GBPT), including specific movements in offense and defense such as catching, passing and throwing a ball, sprinting, stops, changes in direction, jumping, checking and screening, as well as general tests, including a 20 m sprinting test, a repeated sprint ability test (RSA), a modified t-test, countermovement (CMJ), squat (SJ) and drop jump test (DJ), a standing long jump test, a single-leg lateral three jumps test, a standing throw test, and the determination of the one repetition maximum (1RM) in the bench press, bench pull, front squat, and deadlift. Significant correlations were mostly found between different sprinting and jumping tests as well as between different strength tests. The principal component factor analysis revealed four components (power and speed, strength, jump shot performance, and endurance) including 21 variables of high loads (>0.60 or <−0.60). Due to the correlations between the different tests, we suggest a more practicable testing procedure including the 20 m sprinting test, CMJ test, 1RM in the bench press and front squat, as well as the GBPT for elite teams, or the t-test and the YoYo intermittent recovery test for youth and non-elite teams. Despite some correlations between specific and general tests, we suggest using the GBPT to measure specific performance in team handball.

1. Introduction

Team handball is an Olympic team sport played on an indoor court (40 × 20 m²), in which two teams (one goalkeeper and six field players per team) attempt to score more goals than the opposing team. It is characterized by fast-paced, intermittent offensive and defensive actions that alternate between high and low intensity movements. The full-time match total distances (2 × 30 min) range between 3900 to 4700 m and the relative workload during a match in elite male team handball players is 70–80% of the maximal oxygen uptake (VO₂max). There are numerous intensity and technical changes (700–1500 per match) including stops (~30 stops per match), changes in direction (25–35 changes in direction per match), throws on goals (5–10 throws per match), passes (40–60 passes per match), jumps (10–20 jumps per match), and body tackles both light (45–55 light tackles per match) and hard (10–15 hard tackles per match). On average, elite male team handball players walk and stand still 70–75% and sprint only 1–2% of the total playing time [1,2,3].

Testing physical performance in team handball is fundamental to optimize the training process in elite male and female players. In previous studies, physical performance was mostly analyzed in a general manner, rather than being specific to team handball, utilizing physical performance tests such as different sprints (5 m, 10 m, 20 m, or 30 m), repeated sprint ability and agility tests, drop, squat, countermovement, hop and standing long jump tests, bench and leg press, upper body push and pull, push-ups, sit-ups, and grip strength tests, as well as the YoYo intermittent recovery or other aerobic power tests. As expected, elite team handball players performed better than players with lower performance levels in most of these general physical performance tests [4,5,6,7,8,9,10,11,12,13,14,15,16]. Especially in elite team handball academies, these general, unspecific tests were often used to determine the physical performance of the players for the selection process during adolescence or for young adults before transitioning to professional elite team handball. However, it is unknown if these unspecific physical performance tests have an influence on the team handball-specific (game-based) performance, although this knowledge is essential for an optimal selection of young players for the professional elite teams.

A suitable test to measure specific physical performance Is the team handball game-based performance test (GBPT). The GBPT is a reliable and valid test to determine specific game-based performance in team handball players, as found in previous studies [17,18,19,20]. The GBPT consists of eight heats, including sprints, stops, changes in direction, jumps, passes, throws, and body checking in offense and defense. During the GBPT, specific aerobic and anaerobic performance, throwing performance, speed, and agility were measured under conditions similar to those in competition. However, the GBPT requires a lot of equipment, manpower, and time. Consequently, for a practical implementation of physical performance diagnostics in the training process, the reduction of the overall testing time and the test equipment is also essential. Especially in elite team handball, there is less time for physical testing during a season due to trainings and competitions. If a few tests (around five to six general tests) are sufficient to predict the specific physical performance in team handball, this would simplify testing in team handball considerably.

Consequently, the aim of the study was to analyze (1) the relationship between specific physical (game-based) and general performance in young elite male team handball players, and (2) to reduce the number of tests for a more practical implementation of physical performance diagnostics in team handball. We hypothesized different relationships between the GBPT and the general unspecific performance tests, as well as that five to six tests would be suitable to measure the overall performance in team handball.

2. Materials and Methods

2.1. Participants

Twenty elite young adult male team handball field players (mean ± SD; age: 18.6 ± 2.1 years; height: 1.87 ± 0.06 m; weight: 86 ± 10 kg; 6 wings, 10 backcourt players, 4 pivots; 6 left-handed, 14 right-handed players; 8 trainings per week, each lasting 1.5 h) participated in our study. The subjects were recruited from the under-23 and under-19 teams of a professional team handball academy. Additionally, the goalkeepers of each team also performed the general tests, but no data from the goalkeepers were used for the subsequent evaluation or calculations.

All subjects were physically healthy and reported no COVID-19 infection or injuries during the time of the study. The Ethics Committee of the University of Salzburg approved the research protocol (GZ: 44-2015) in accordance with the Declaration of Helsinki, and all subjects signed the informed consent prior to participation. For players under 18 years of age, their legal guardians reviewed and signed the informed consent.

2.2. Study Design

The study was divided into a general and specific part. In part one (the general part), all subjects performed general physical performance tests, including a 20 m sprinting test, a repeated sprint ability test (RSA), a modified t-test, a countermovement (CMJ), a squat jump (SJ), a drop jump test (DJ), a standing long jump test, a single-leg lateral three jumps test, a standing throw test, and the determination of the one repetition maximum (1RM) in the bench press, bench pull, front squat, and deadlift. In part two (the specific part), all subjects performed a team handball game-based performance test (GBPT). To ensure maximal effort throughout the general tests and the GBPT, the subjects were constantly verbally supported by the test leaders and their teammates.

During the testing procedure, we used a standardized testing sequence for all subjects. On the first testing day (in the general part), we started by measuring anthropometric data (body weight and height) and collecting personal data (playing position, training experience, and age). Subsequently, all subjects performed a general 8-min jogging warm-up followed by a 15-min specific warm-up including mobilization, four maximal countermovement jumps, and 20–30 m run-ups. The subjects were divided into two randomized groups and performed the SJ, CMJ, DJ, and 20 m sprinting tests, as well as the single-leg three jumps test, the standing long jump test, and the standing throw test in the forenoon (the second group vice versa). After a 2–3 h lunch break, all subjects performed the modified T-test and the RSA test in the afternoon (the warm-up in the afternoon was the same as in the forenoon). On the following, second testing day (in the general part), all subjects performed a general 8-min jogging warm-up followed by a 15-min specific warm-up including mobilization, ten pushups, and squats. The subjects were again divided into two randomized groups and determined the 1RM in the bench press and bench pull in the forenoon, and the front squat and deadlift in the afternoon (after a 2–3 h lunch break, and the same warm-up as in the forenoon; the second group vice versa). All subjects performed all general physical performance tests within these two testing days. Fatigue should not influence the performance in any tests, as all subjects of the study were familiar with two intensive training sessions per day.

In the specific part, all subjects performed the GBPT after an 8-min general (jogging) warm-up and a 15-min team handball-specific warm-up, including passing, throwing, changing directions, and light tackling.

2.3. General Tests

To determine vertical jump height, the subjects performed three SJs from a squatting position (knee angle of 90°, without countermovement) and three CMJs from an upright position on a portable Kistler force plate (Kistler Force Plate, Kistler, Winterthur, Switzerland). Arm swing was not allowed during the SJ and CMJ. Jump height of the SJ and CMJ was calculated by the vertical velocity of the center of mass. Finally, all subjects performed three DJs, without arm swing, from a 40 cm box, with maximum effort (maximal jump height and minimal ground contact time). A jump was invalid if the subjects used their arms (SJ, CMJ, and DJ), performed a countermovement before jumping (SJ) or jumped vertically from the box (DJ). Data for the SJ, CMJ, and DJ tests were recorded and analyzed offline using Matlab (Matlab2020a, MathWorks, Natick, MA, USA). The maximal value of the three repetitions (1 min recovery between the repetitions) was subsequently used in the statistical analysis.

In the standing long jump test, the subjects stood with both feet tiptoeing behind the painted starting line and jumped as far forward as possible (Figure 1A). Arm swing was allowed during the jump. Jump distance was defined as the measured distance between the starting line and the heel of the rear foot of the last contact. A jump was considered invalid if the subjects touched the ground with their hands or lost balance. The maximal value of the three repetitions, with at least three minutes of recovery between repetitions, as the whole group performed each heat one after the other, was subsequently used for statistical analysis.

In the single-leg three-jump test (3er jump test), the subjects had to stand on one leg with the designated foot (tiptoe) behind the starting line (painted line). When ready, they had to perform three consecutive maximal hops with the designated leg, landing on both feet (Figure 1A). Arm swing was allowed during the 3er jump test. Jump distance was defined as the measured distance between the starting line and the heel of the rear foot of the last contact. A jump was invalid if the subjects touched the ground with their hands, the second foot, or lost balance. The maximal value of the three repetitions (with at least three minutes of recovery between repetitions) was subsequently used for statistical analysis.

In the standing throw test, the subjects had to throw the ball into the center of the team handball goal with maximal intentional velocity and no opposition (Figure 1A). Maximum ball velocity was measured with a radar gun (Perform Better, Warwick, UK) located behind the goal. The maximum value of the three repetitions (with at least three minutes of recovery between repetitions) was subsequently used for statistical analysis.

In the 20 m sprint test, each subject started with their front foot tiptoeing behind the starting line, which was a painted line half a meter behind the start timing gate (as shown in Figure 1A). They then sprinted as fast as possible to the finish line, 20 m away. The start time was self-determined. Two light beams (Micro Gate Witty, Micro Gate, Bolzano, Italy) placed at the start and finish line were used to measure sprinting time. Each subject repeated the sprint test twice, with at least three minutes of recovery between the two sprints. The fastest 20 m sprinting time was subsequently used for statistical analysis.

Regarding the guidelines of the National Team Handball Federation, we used the Federation’s modified t-test for our study. Each subject performed a warm-up heat (80% of maximum effort) and two maximum effort repetitions (with at least three minutes of recovery between repetitions), with the fastest time used for calculations. Subjects started in a standing position with their preferred foot (tiptoe) behind the starting line (a painted line, a half meter behind the start timing gate, as shown in Figure 1A), followed by a maximal acceleration forwards at maximal effort until reaching the first cone (placed 5 m from the starting line) with their right hand, followed by side-steps to the left, reaching the second cone (placed 2.5 m left from the first cone) with their left hand, followed by side-steps to the third cone on the right side (placed 5 m right from the second, respectively, 2.5 m right from the first cone), reaching with their right hand, back to the first cone, reaching with their left hand, and sprinting backwards to the starting line (painted line). During the side-steps, cross-over steps were not allowed, and the subjects must look forwards during the whole test. To measure the total t-test time, we used a light beam (Micro Gate Witty, Micro Gate, Bolzano, Italy) placed at the start (finish) line.

In the RSA test, the subjects started in a standing position with their preferred foot a half meter behind the start timing gate (Figure 1A) and sprinted forwards to a painted line (change-in-direction line) at 15 m, followed by a fast 180° change-in-direction (one foot must touch the line) and sprinted back forwards to the start (finish) line. To measure the sprinting time, we used the same light beam (Micro Gate Witty, Micro Gate, Bolzano, Italy) as in the modified t-test, placed at the start (finish) line. The RSA test consisted of six 2 × 15 m sprints, with 20-s intervals between each start. This break was controlled and timed by the test leader, while an acoustic countdown (“3, 2, 1, go”) was used to start on time. The mean (RSA_mean) and best (RSA_best) time of the six sprints as well as the RSA fatigue index (RSA_index) was calculated to determine the RSA.

$${\mathrm{RSA}}_{\mathrm{index}}=\left(\frac{{\mathrm{RSA}}_{\mathrm{mean}}}{{\mathrm{RSA}}_{\mathrm{best}}}\times 100\right)-100$$

The 1RM for the bench press, bench pull, front squat, and deadlift was used as a measure of dynamic strength, whereas the subjects progressively increased (around 10% of the 1RM, starting with 80% of the 1RM from the last test) their resistance across four heats until 1RM was reached in the final heat. Consequently, they performed five heats in total, consisting of a warm-up heat (50% of 1RM from the last test) and the four measuring heats. The bench press was performed lying horizontally (face up) on the bench with the feet planted on the floor. The subjects grasped the barbell with an overhand grip slightly wider than shoulder-width. After removing the barbell from the rack, the barbell was lowered until touching the chest. Without bouncing the barbell from the chest, the barbell was lifted in a self-determined tempo. The bench pull was performed lying horizontally (face down) on the bench with the arms extended below the bench. Like in the bench press, the subjects took a shoulder-width overhand grip on the barbell and pulled the barbell up until the barbell contacted the bottom of the bench. Once the barbell contacted the bench, the subjects lowered their weight in a controlled manner back to the starting position. In the front squat, the barbell was held in the front of the chest. The squats were performed to a knee angle of 90° and were controlled by the test leader using a set bar (in the lowest position, the barbell had to touch the set bar). Trials in which the subjects failed to reach the knee angle of 90° were considered unsuccessful. The front squat was finished when the barbell was lifted to the starting position. In the deadlift, the barbell was lifted off the ground, keeping a flat back, to the hips finishing in an upright position. During the deadlift, the feet stood shoulder-width apart, and the hands grasped the barbell just outside the legs. One trial was finished by lowering the barbell to the ground in a controlled manner.

2.4. Team Handball Game-Based Performance Test (GBPT)

In the GBPT, the subjects performed one warm-up heat (around 70% of maximum efficiency) to become familiar with the test, and eight test heats. The test heats in the GBPT included team handball-specific movements in offense and defense, transitions from offense to defense and defense to offense, as well as active recovery.

In defence, the subjects had to touch padded roll mats (Figure 1B) with both hands (like tackling a rival offensive player in a team handball match) while sprinting, starting at the 6 m line and finishing at the 9 m line. The right-handed players had to touch the left roll mat twice and the right roll mat once at the 9 m line, while the left-handed players did the opposite. In offense, the subjects had to touch squared (0.5 × 0.5 m) touching fields on the floor (Figure 1B) with at least one foot while sprinting, starting at 12 m and finishing at the 9 m line. During sprinting, they had to catch and pass the ball. The right-handed players had to touch the right squared touching field twice and the left squared touching field once, while the left-handed players did the opposite. All running distances in offense were standardized by the squared touching fields on the court and in defence by the positions of the padded roll mats (Figure 1B). In heats #2, #3, #4, #6, and #8, all subjects had to finish the offensive actions with a jump shot. In the jump shot, the right-handed players had to jump with maximal take-off from the left foot on the right squared touching field and throw as fast as possible to the lower left corner of the goal, while the left-handed players did the opposite. In heats #3 and #6, the subjects had to perform a fast break by sprinting as fast as possible from defence to offense (after touching the padded roll mat at the 6 m line), catching the ball at 12 m, and finishing with a jump shot at 9 m in offense. The right-handed players had to touch the right squared touching field at take-off, while the left-handed players did the opposite. In heats #4 and #6, the subjects had to perform a fast retreat by sprinting as fast as possible from offense to defence immediately after the jump shot.

To measure the time between the first and last contact on the padded roll mats (defence time) and squared touching fields (offence time) as well as the time between the 9 m line in defence and the middle line (10 m-fast break time), the middle line and the squared touching field in offence (20 m-fast break time) and the 9 m line in offence and defence (fast retreat time), we used three hand stopwatches (Hanhart Stratos 2, Hanhart GmbH, Gütenbach, Germany). Due to the different accelerations after the jump shot, we decided not to use the fast retreat time for subsequent evaluations. The breaks between offence and defence as well as defence and offence (20 s), between two defensive or offensive actions (15 s), and between two heats (40 s) were controlled by the Multi-Timer-Ultimate software (Multi-Timer-Ultimate 3.1, Wallroth, Berlin, Germany). All breaks started immediately after finishing the offensive or defensive actions, and the last three seconds of the breaks were counted through an acoustic countdown (“3, 2, 1, go”). To determine the ball velocity and jump height during the jump shots, the videos of all jump shots were recorded using a high-speed (200 frames per second) video camera (JVC-GC-PX100BE, JVC, Yokohama, Japan). Subsequently, the Tracker Video Analysing Software (Tracker 4.59, Douglas Brown, Aptos, CA, USA) was used to calculate the 2D-position of the centre of the ball and to determine the flight time. A team handball goal was used for calibration (4-point calibration, 2 × 3 m). Flight time was defined as the time between the last floor contact of the jump foot at take-off and the first floor contact of the landing foot. Ball velocity was determined by calculating the linear velocity of the ball 2D-positions after ball release (15 frames). Jump height was calculated using the following formula (g = gravity acceleration). For subsequent analysis and evaluation, the mean values (ball velocity and jump height) of the best three attempts were used.

$$h_{jump}=\frac{g}{8}\times t_{flight}^2$$

During the GBPT, we used a heart rate belt with a sensor module (Suunto T6d, Suunto, Vantaa, Finland) and a portable respiratory gas exchange measurement analysis system in breath-by-breath mode (K5, Cosmed, Rome, Italy) to measure heart rate (HR) and oxygen uptake. We calculated HR and oxygen uptake peak values (HR-peak and VO₂-peak) in all heats during the whole GBPT for subsequent evaluations. To prevent errors in determining VO₂-peak, only peak values were used where two breath-by-breath values before and after the peak value were not less than 90 percent.

2.5. Statistical Analyses

All statistical analyses were conducted using SPSS version 27 (IBM Corp., Armonk, NY, USA) with a significance level of p < 0.05 for all tests. Mean values ± standard deviations and 95% confidence intervals of all variables were calculated for descriptive statistics (

Table 1. Descriptive data, mean, standard deviations (±SD) and 95% confidence intervals (CI) for all variables in the game-based performance test (GBPT) and general tests.

GBPT Variables	Mean ± SD (95% CI)	General Tests Variables	Mean ± SD (95% CI)
VO2-peak [mL O2·min−1·kg−1]	69.1 ± 10.3 (64.8–73.2)	20 m sprinting time [s]	2.99 ± 0.11 (2.94–3.04)
HR-peak [beats·min−1]	189 ± 8 (186–193)	t-test time [s]	5.35 ± 0.24 (5.24–5.45)
10 m-fast break time [s]	1.80 ± 0.09 (1.77–1.84)	RSAbest [s]	5.93 ± 0.20 (5.84–6.01)
20 m-fast break time [s]	1.78 ± 0.13 (1.73–1.83)	RSAmean [s]	6.15 ± 0.20 (6.06–6.23)
Defense time [s]	5.53 ± 0.22 (5.44–5.63)	RSAindex [%]	3.7 ± 1.7 (3.0–4.5)
Offence time [s]	5.91 ± 0.32 (5.78–6.05)	CMJ height [m]	0.39 ± 0.05 (0.37–0.41)
Jump shot—ball velocity [m·s−1]	24.9 ± 1.3 (24.3–25.4)	SJ height [m]	0.38 ± 0.06 (0.36–0.41)
Jump shot—jump height [m]	0.36 ± 0.08 (0.32–0.40)	DJ height [m]	0.30 ± 0.05 (0.27–0.32)
		DJ contact time [ms]	180 ± 30 (170–195)
		Standing long jump distance [m]	2.56 ± 0.17 (2.48–2.63)
		Left 3er jump test distance [m]	7.69 ± 0.59 (7.41–7.94)
		Right 3er jump test distance [m]	7.59 ± 0.60 (7.35–7.85)
		Standing throw–ball velocity [m·s−1]	23.8 ± 1.6 (23.1–24.5)
		1RM bench press [kg]	88 ± 17 (80–95)
		1RM bench pull [kg]	81 ± 11 (77–86)
		1RM front squat [kg]	121 ± 18 (113–129)
		1RM deadlift [kg]	125 ± 22 (116–134)

Note: VO₂ = Oxygen uptake, HR = Heart rate, 1RM = One repetition maximum, RSA = repeated sprint ability, CMJ = Countermovement jump, SJ = Squat jump, DJ = Drop jump, 3er = Single-leg three.

). Normality of the data was verified using the Shapiro–Wilk test and was found to be normal for all variables used. The relationship between the general physical performance tests, the GBPT, and anthropometric variables (body height and weight) was determined by calculating Pearson Product–Moment correlation coefficients (r). r was categorized as high (0.8–1.0), moderate (0.5–0.8), and low (0.0–0.5) [21]. Linear regression was additionally calculated and diagrammed as linear regression analysis plots between selected variables. Finally, a principal component factor analysis (Varimax, eigenvalues > 1) was calculated to determine a few components that define multiple variables.

3. Results

Descriptive data, means, standard deviations (±SD), and 95% confidence intervals for all variables in the GBPT and general tests are presented in Table 1.

Pearson Product-Moment correlation analysis (

Table 2. Pearson product-moment correlation coefficients between all measured variables in the game-based performance tests and all general tests.

	Body Weight [kg]	VO2-Peak [ml O2·min−1·kg−1]	HR-Peak [beats·min−1]	10 m-Fast Break Time [s]	20 m-Fast Break Time [s]	Defense Time [s]	Offence Time [s]	Jump Shot—Jump Height [m]	Jump Shot—Ball Velocity [m·s−1]	RSAmean [s]	RSAbest [s]	RSAindex [%]	20 m Sprinting Time [s]	t-Test Time [s]	CMJ Height [m]	SJ Height [m]	DJ Height [m]	DJ Contact Time [ms]	Standing Long Jump Distance [m]	Right 3er Jump Test Distance [m]	Left 3er Jump Test Distance [m]	Standing Throw- Ball Velocity [m·s−1]	1RM Bench Press [kg]	1RM Bench Pull [kg]	1RM Front Squat [kg]	1RM Deadlift [kg]
Body height [m]	0.620 **	−0.426	−0.601 **	0.286	−0.016	0.063	0.070	0.543 *	0.412	0.214	−0.041	0.512 *	0.032	0.171	−0.053	−0.015	−0.119	−0.277	0.189	0.066	0.101	0.322	0.252	0.289	−0.168	0.410
Body weight [kg]		−0.473 *	−0.249	0.194	0.281	0.147	0.165	0.175	0.562 **	0.385	0.189	0.379	0.036	0.162	−0.166	−0.077	0.205	−0.530 *	0.157	−0.066	0.145	0.621 **	0.780 **	0.669 **	0.378	0.680 **
VO2-peak [mL O2·min−1·kg−1]			0.178	0.077	−0.390	−0.041	−0.273	−0.018	−0.105	−0.021	−0.044	0.032	0.213	0.059	−0.064	−0.148	0.096	−0.107	−0.248	−0.102	−0.035	−0.202	−0.180	−0.236	−0.104	−0.333
HR-peak [beats·min−1]				0.308	0.459 *	0.304	0.355	−0.501 *	−0.313	0.194	0.384	−0.410	0.335	0.346	−0.301	−0.405	0.093	−0.114	−0.303	−0.371	−0.214	−0.029	−0.219	−0.265	−0.086	−0.291
10 m-fast break time [s]					0.220	0.641 **	0.365	−0.082	−0.050	0.727 **	0.571 **	0.249	0.613 **	0.804 **	−0.537 *	−0.635 **	0.079	−0.706 **	−0.498 *	−0.618 **	−0.484 *	0.023	−0.165	−0.341	−0.503 *	−0.183
20 m-fast break time [s]						0.290	0.407	−0.245	−0.141	0.459 *	0.491 *	−0.104	0.471 *	0.476 *	−0.510 *	−0.609 **	−0.012	−0.036	−0.190	−0.489 *	−0.333	0.284	0.069	−0.136	−0.019	0.053
Defense time [s]							0.658 **	−0.401	−0.195	0.720 **	0.792 **	−0.216	0.558 *	0.805 **	−0.480 *	−0.577 **	0.319	−0.462 *	−0.672 **	−0.584 **	−0.682 **	−0.100	0.005	−0.294	−0.328	−0.296
Offence time [s]								−0.274	−0.305	0.422	0.462 *	−0.118	0.388	0.520 *	−0.399	−0.324	−0.031	−0.220	−0.353	−0.450 *	−0.475 *	0.199	0.134	−0.156	0.099	−0.017
Jump shot—jump height [m]									0.461 *	−0.103	−0.233	0.272	0.081	−0.276	0.138	0.169	−0.326	−0.077	0.257	0.186	0.315	0.416	0.068	0.127	−0.049	0.449 *
Jump shot—ball velocity [m·s−1]										0.028	−0.066	0.182	−0.092	−0.232	−0.106	0.024	0.197	−0.417	0.262	0.193	0.365	0.304	0.442	0.472 *	0.168	0.426
RSAmean [s]											0.881 **	0.145	0.752 **	0.845 **	−0.779 **	−0.781 **	0.036	−0.700 **	−0.708 **	−0.754 **	−0.700 **	0.130	0.154	−0.237	−0.264	−0.048
RSAbest [s]												−0.339	0.729 **	0.786 **	−0.708 **	−0.777 **	0.196	−0.519 *	−0.763 **	−0.773 **	−0.709 **	−0.001	0.035	−0.377	−0.288	−0.238
RSAindex [%]													−0.038	0.039	−0.061	0.079	−0.338	−0.301	0.187	0.124	0.091	0.273	0.246	0.334	0.082	0.414
20 m sprinting time [s]														0.687 **	−0.679 **	−0.815 **	−0.018	−0.416	−0.646 **	−0.664 **	−0.557 *	0.144	−0.168	−0.488 *	−0.356	−0.097
t-test time [s]															−0.707 **	−0.812 **	0.136	−0.520 *	−0.767 **	−0.832 **	−0.762 **	−0.051	−0.076	−0.379	−0.483 *	−0.283
CMJ height [m]																0.890 **	0.160	0.511 *	0.695 **	0.809 **	0.675 **	−0.034	−0.012	0.359	0.250	0.244
SJ height [m]																	−0.024	0.425	0.702 **	0.818 **	0.638 **	−0.006	0.128	0.479 *	0.405	0.337
DJ height [m]																		−0.265	−0.018	−0.035	0.177	−0.035	0.294	0.221	0.087	−0.043
DJ contact time [ms]																			0.339	0.490 *	0.224	−0.210	−0.414	−0.181	0.005	−0.201
Standing long jump distance [m]																				0.787 **	0.850 **	0.195	0.188	0.515 *	0.446 *	0.422
Right 3er jump test distance [m]																					0.778 **	0.034	0.060	0.460 *	0.303	0.293
Left 3er jump test distance [m]																						0.322	0.269	0.597 **	0.430	0.382
Standing throw—ball velocity [m·s−1]																							0.575 **	0.434	0.427	0.568 **
1RM bench press [kg]																								0.827 **	0.764 **	0.676 **
1RM bench pull [kg]																									0.668 **	0.751 **
1RM front squat [kg]																										0.599 **

Note: VO₂ = Oxygen uptake, HR = Heart rate, 1RM = One repetition maximum, RSA = repeated sprint ability, CMJ = Countermovement jump, SJ = Squat jump, DJ = Drop jump, 3er = Single-leg three. Significant correlation: p < 0.05 *, p < 0.01 **.

) revealed significant correlation (p < 0.01) between body height and body weight (r = 0.62), HR-peak (r = −0.60), between body weight and ball velocity in the jump shot (r = 0.56), ball velocity in the standing throw (r = 0.62), 1RM in the bench pull (r = 0.67), 1RM in the deadlift (r = 0.68), 10 m-fast break time in the GBPT and defense time in the GBPT (r = 0.64), RSA_mean (r = 0.73), RSA_best (r = 0.57), 20 m sprinting time (r = 0.61), t-test time (r = 0.80), SJ height (r = −0.64), DJ height (r = −0.71), right 3er jump test distance (r = −0.62), 20 m-fast break time in the GBPT and SJ height (r = −0.61), between defense and offense time in the GBPT (r = 0.61), RSA_mean (r = 0.72), RSA_best (r = 0.79), t-test time (r = 0.81), standing long jump test (r = −0.67), left 3er jump test distance (r = −0.68), between RSA_mean and RSA_best (r = 0.88), 20 m sprinting time (r = 0.75), t-test time (r = 0.85), CMJ height (r = −0.78), SJ height (r = −0.78), DJ height (r = −0.70), standing long jump distance (r = −0.71), left and right 3er jump distance (r = −0.75 and −0.70), between RSA_best and 20 m sprinting time (r = 0.73), t-test time (r = 0.79), CMJ height (r = −0.71), SJ height (r = −0.77), standing long jump distance (r = −0.76), left and right 3er jump distance (r = −0.77 and −0.71), between 20 m sprinting test time and t-test time (r = 0.69), CMJ height (r = −0.70), SJ height (r = −0.82), standing long jump distance (r = −0.65), right 3er jump distance (r = −0.66), between t-test time and CMJ height (r = −0.71), SJ height (r = −0.81), standing long jump distance (r = −0.70), left and right 3er jump distance (r = −0.83 and −0.76), between CMJ height and SJ height (r = −0.89), standing long jump distance (r = −0.70), left and right 3er jump distance (r = −0.81 and −0.68), between SJ height and standing long jump distance (R = −0.70), left and right 3er jump distance (r = −0.82 and −0.64), between standing long jump distance and left and right 3er jump distance (r = −0.79 and −0.85), between left and right 3er jump distance (R = −0.79), between the 1RM in the bench press and bench pull (r = 0.69), front squat (r = 0.84), between the 1RM in the bench pull and front squat (r = 0.67), deadlift (r = 0.75), as well as between the 1RM in the front squat and deadlift (r = 0.60).

The principal component factor analysis revealed six components (eigenvalues > 1) with a cumulative variance of 84.7%. All elements (loads of >0.60 and <−0.60, respectively) of the principal component factor analysis for each component, including the percentage of variance, are shown in Figure 2. In the principal component analysis, a Kaiser–Meyer–Olkin measure of sampling adequacy value of 0.76 was calculated.

For a detailed visualization of the results in Table 2, we selected some representative linear regression analysis scatterplots as shown in Figure 3. Figure 3A–D show the linear regression between a specific variable in the GBPT, such as the defense time (Figure 3A), 20 m-fast break time (Figure 3B), jump height (Figure 3C), and ball velocity in the jump shot (Figure 3D), and the appropriate variable in the general physical performance test, such as the t-test time (Figure 3A), 20 m sprinting time (Figure 3B), CMJ height (Figure 3C), and ball velocity in the standing throw (Figure 3D). Figure 3E shows the linear regression between the endurance variable in the specific GBPT (VO₂peak) and general RSA test (RSA_index), and Figure 3F shows the linear regression between body weight (an anthropometric variable) and the variable most determined in strength tests (1RM in the bench press).

4. Discussion

The aim of the study was to analyze the relationship between specific (game-based performance) and general performance in young elite team handball players. As expected, we found significant correlations between some specific and general performance variables. The factor analysis revealed two components (power and speed, strength) including more than seven variables, and two components (jump shot performance and endurance) including two variables with loads greater than 0.60.

In the present study, we measured the specific and general performance in young adult male team handball players recruited from a professional German team handball academy. Comparing the general physical performance, it was found that the jumping, sprinting, and throwing performance, as well as the 1RM in the bench press, was higher compared to Norwegian, Greek, Belgian, and Icelandic players of similar age [5,7,8,9], except the CMJ height, which was higher in the Icelandic U21 National Team players [9]. In the GBPT, the participants in the present study reached a better VO₂-peak, defense, and offense time, but a lower jump height and ball velocity in the jump shot compared to adult elite Austrian team handball players [17]. We suggest that the performance level of the young adult elite team handball players in our study is on the European top level and that the results are representative of the world’s best team handball players in this age (17–22 years).

A great problem in physical performance diagnostics is the number of different tests, the expenditure of time for these tests, and the crossover effects due to the requirements of the participants in the different tests as well as fatigue. In our study, we used an adapted test battery of the German Team Handball Association, including fourteen general physical performance tests. To minimize the crossover effects and fatigue, we used two testing days for the general tests and an additional testing day for the GBPT. It is not surprising that the coaches are not willing to spend three days for testing during their training periods. Consequently, a reduced test battery that minimizes the testing time to one or, at most, two days would be a huge improvement and would increase the readiness of the coaches to test physical performance frequently (3–4 times) per year.

As shown in Figure 2, seven variables of the GBPT with a load of >0.60 (<−0.60) were found for the components power and speed (10 m-fast break time and defense time), jump shot (ball velocity and jump height in the jump shot), strength (body weight), endurance (body height and VO₂peak), and heart rate (HR-peak); in summary, five out of six components. The GBPT is a team handball-specific test determining different physical performance parameters (power, speed, agility, jumping, and throwing performance, as well as endurance) and is suitable for discriminating between different performance levels, sex, and age [17,18,19,22]. Consequently, we suggest using the GBPT to measure physical performance in elite team handball instead of different general physical tests due to its overall performance diagnostics.

However, two additional strength tests for the upper (1RM bench press) and lower (1RM front squat) body should be used supplementary to determine strength. Due to the different body and muscle growth of the participants, we found a wide range in the four strength tests that could have influenced the significant correlations between them. Consequently, we would prefer two strength tests (upper and lower body) instead of only one, although the high correlations between the strength tests would suggest using only one representative strength test.

The test duration of ~20 min for the GBPT is optimal to combine the GBPT with the strength tests on one testing day. In the morning, starting with the strength tests, and after a 3–4 h break in the afternoon performing the GBPT (or vice versa), enables testing an elite handball team within one day. Additionally, a 20 m sprinting test (before the GBPT) and a CMJ test (before the 1RM front squat) could be included in the warm-up programs. We suggest that this combination of five physical performance tests (20 m sprinting test, GBPT, 1RM bench press, CMJ test, 1RM front squat) is optimal to determine physical performance in elite team handball. Because of its low expenditure of time (2–3 h on one testing day), these tests could be used several times per season (to determine the training effect) and one can find many comparative data in male and female team handball players of different performance levels and ages in the literature [4,5,6,7,8,9,11,12,13,14,15,16,17,18,19,22,23,24,25,26,27].

However, a limitation of the GBPT is the effort required for preparation (measuring and positioning of the touch fields and padded roll mats), the number of measuring devices (Multi-Timer-Ultimate, high-speed camera, and portable respiratory gas exchange measurement analysis system) and human resources (during measuring and data analysis). The consequence is high financial costs if the GBPT is not performed in a scientific study (like in the present study). Consequently, the GBPT is not practicable in youth or non-elite team handball. As shown in Figure 2 and Figure 3, as well as in Table 2, we found several significant correlations between different physical performance tests and the GBPT. We suggest using a classical sprinting test (20 m sprinting test), an agility test (T-test), a jumping test (CMJ test), and strength tests (1RM bench press and front squat) in youth or non-elite team handball.

However, the low correlation between the VO₂peak in the GBPT and the RSA_index was unexpected, as we suspected that the anaerobic performance in the RSA (RSA_index) would correlate with the aerobic performance (VO₂peak) in the GBPT [18]. Consequently, the RSA test is not suitable for determining endurance in team handball, and we would suggest utilizing the YoYo intermittent recovery test (YoYo IR test) instead. In a previous study, a significant correlation between the VO₂peak in the GBPT and the running distance in the YoYo IR2 (level 2) test was found in top-elite female team handball players [28]. To summarize, we suggest a combination of six physical performance tests (20 m sprinting test, t-test, CMJ test, 1RM bench press and front squat, as well as the YoYo IR test) in youth and non-elite team handball.

A limitation of the study was that the participants performed fourteen tests on two testing days, as the cross-over effect and fatigue could have influenced the results. However, the young adult elite team handball players in the present study are familiar with a high training load (eight trainings and 1–2 games per week), suggesting a negligible cross-over effect and influence of the results due to fatigue.

5. Conclusions

The aim of the study was to analyze the relationship between specific (game-based performance) and general performance in young elite team handball players. Several significant correlations between the different general and specific test variables were found, and the principal component analysis revealed four components (power and speed, strength, jump shot performance, and endurance) including 21 variables of high loads (>0.60 or <−0.60). To optimize the testing process in elite team handball, we suggest a testing procedure within one testing day for all players, including five physical performance tests (20 m sprinting test, GBPT, CMJ test, 1RM in the bench press and front squat) that enables testing several times in one season to determine the training effect. However, due to the high financial costs of the GBPT, we would suggest a different testing procedure in youth and non-elite team handball, including six physical performance tests (20 m sprinting test, t-test, CMJ test, 1RM bench press and front squat, as well as the YoYo IR test).