Analysis of playing position and match status-related differences in external load demands on amateur handball: a case study

– Currently, the load quantification during training and competition in all sports is important to injury prevention, design specific training sessions, and player performance enhancement. The present study aimed to describe the external load profile of amateur-level handball players and to compare the playing position and match outcome-related differences in an official 2 nd Regional Division male official match. 19 handball players were monitored by WIMU PRO TM electronic performance and tracking systems (EPTS) with ultrawide-band (UWB) indoor location technology. Statistical analysis was composed of a t-student independ- ent sample test and one-way ANOVA with Bonferroni posthoc. In amateur handball, most demands are composed of aerobic intensity displacements (<12 km/h, 76% playing time) and low-intensity impacts (<5G). Differences were found between playing positions where wingers obtained the greatest values in high-intensity actions, centers in the volume of demands, and defensive specialist players in impacts. Besides, the winning teams performed greater high-intensity demands both in displacements and speed changes. In conclusion, the present results realized the first approximation to handball amateur demands, considering the playing role and match status-related demands to design specific training plans.


INTRODUCTION
Handball is a team sport that requires intermittent high-intensity actions, as accelerations, decelerations, changes of direction and speed, high-speed displacements, jumps, and landings 1 . Therefore, similarly to other team sports as basketball, football, or rugby, it is very important for team staff the design of training and the workload monitoring to analyze the demands during competition and prepare specifically for the worst-case scenario in handball matches 2 . In this sense, the study of physical demands of handballers has been one of the most investigated topics in the last decade, shown in standardized variables as total distance covered, speed profiles, high-intensity actions, or impacts during competition 3,4 .
Thanks to technological advances, electronic performance, and tracking systems (EPTS) have been developed that allow monitoring the skills and displacements of the players during training and competition 5 . Based on the specific demands of competition, different investigations have carried out an analysis to determine the physical, anthropometric, and physiological characteristics of handball players through EPTS 6,7 . These studies recommended that coaches should consider anthropometric and physical characteristics of handballers such as height, weight, fat percentage, muscle mass percentage, hand size, and arm perimeter when assessing the potential performance of young male and female players. In the same researches, other important key factors have been mentioned as throw speed, vertical jump height, aerobic capacity, maximum speed, and repeated sprint ability 8 .
Also, most studies realized in team sports have identified an effect of contextual variables as playing position 9,10 and match outcome 11,12 in soccer and basketball. In handball, it has been also identified that the physical demands performed are influenced by playing positions at elite-level 1,13,14 and recreational-level 7 . Instead, in our knowledge, a lack of research about the effect of match outcome on the workload demands of the handball players both at elite-level and amateur-level has been identified. Therefore, the main purpose of the present study is to describe the workload demands in amateur-level handball players during an official competition game and identify differences between playing positions (center, backcourt, wing, pivot, and defensive specialist) and the match outcome (win or lose).

METHODS Participants
Twenty-four handball players that competed in the 2 nd regional male division participated voluntarily in the present study. The inclusion criteria were: (a) not suffer any injury that prevents the sports practice before and during the game, (b) participate more than 50% of the playing time in each period, and (c) the goalkeepers were excluded from the study due to workload demands and technical-tactical actions differ to court players 8 . Therefore, a total of 19 handball players were analyzed in the present study.
A previous consent was requested from the Handball Regional Fed-eration, the referees, and both handball teams because the use of EPTS attached to the athlete during the regular handball season is prohibited, and all of them approved their utilization. Besides, all players signed informed consent before the data register. The study that was carried out following the bioethical considerations of the Declaration of Helsinki (2013), and was approved by the bioethics committee of the University of Murcia (Registration Code 2061/2018).  1a). To attach the device on the player's body (upper back), an anatomically adjusted harness was used that locates the device at the level of the interscapular line (vertebrae T2-T4) (see figure 1a). The device has previously been used both in outdoor sports 15 and indoor sports 10 . Also, its reliability and validity have been previously evaluated for indoor conditions positioning in handball 16 , and the accelerometer reliability for external workload measurement in the sport context 17 .

Variables
External workload: The external workload variables analyzed in the present study are divided into three groups: (a) distance and speed of displacements, (b) neuromuscular load, and (c) speed changes. Due to the particularity of the substitutions and playing position changes during the game, only the relative variables per minute (e.g. relative distance, accelerations per minute) and the maximum values of each variable (maximum acceleration or maximum speed) were used due to they are more representative of the total volume and intensity in team sports with intermittent efforts (e.g. basketball, handball, soccer…) and should be used to provide more accurate information on the demands, regardless of playing time 10  For the data monitoring, UWB system was installed on the court as follow: (a) six antennas were fixed around the track, two of them were located in the extension of the line midfield, one to each band, at a distance of 2 m. of the line that delimits the field by its sides and the remaining four were placed in the respective 4 corners of the field, at a distance of 2.5 m. in the prolongation of the diagonal line of each vertex (see figure 2); (b) the antennas were fixed through a tripod at 3 m. tall; (c) once installed, they were turned on one by one ensuring that the master antenna was the last; (d) after switching on, a process of autocalibration of the antennas occurred during 5 'in which they calculate the position in x and y coordinates of all of them; and (e) finally the devices were turned on leaving them static, on a flat surface and without magnetic elements around for 1', so that they communicated with the antennas and are configured within the Cartesian plane created by them 18 .
Devices were placed on the athletes using an anatomically adjusted harness, as usually used in team sports. The first and second half temporary marks and the selection of the playing time of each handballer were made in real-time using a laptop with S Vivo Software. At the end of the match, the devices were removed from the players and placed at the charging station to download the raw information recorded during the match. Raw data were exported in Excel format and imported into the statistical software for statistical analysis.

Data analysis
The mean and standard deviation (M±SD) of the analyzed variables were used for the descriptive analysis. The distribution of the variables and the homogeneity of the variances were analyzed using Kolmogorov-Smirnov and Levene test, reporting a parametric distribution in all of them. For the inferential analysis based on the match result and the specific positions, the T-statistics for independent samples and one-way ANOVA were used, comparing them in pairs using Bonferroni. The effect size of the differences was calculated using Cohen's d (d) and squared partial omega (ω p ²), respectively. The statistic d was interpreted as: d>0.2 small, d>0.6 moderate, d>1.2 large and d>2.0 very large and the statistic ω p ² as >0.01 small; >0.06 moderate and >0.14 large 19 . Statistical analysis was performed using the Statistical Package for the Social Sciences software (version 24, IBM Corporation, Armonk NY, USA).   Table 2 shows the descriptive and inferential analysis of the external load variables analyzed in the present investigation according to the specific position. Significant differences were found between specific positions with a moderate effect size in DR Z4 (F=6. 26 Table 3 shows the descriptive and inferential analysis of the external load variables analyzed in this investigation based on the final result of the match.

RESULTS
The winning teams obtained higher values with a very large effect size in the DRZ4 and DecZ4 variables and large effect size in DR, DRZ3, ImpZ2, steps, PL, AccZ3, AccZ4, DecMAX and DecZ3. The losers obtained higher values with a large effect size in the ImpZ1 variable.

DISCUSSION
The monitoring of the load in competition and the individualization according to the specific characteristics of the players is fundamental for the design of training tasks and specific fitness programs in team sports 2 . Therefore, the present study aimed to describe the external workload of senior amateur handball players in official competition and analyze the differences in the specific positions and the match outcome. The amateur handball player covers 72.12 m/min, of which almost all is done walking and jogging (76%), receiving a total of 54.23 impacts per minute, the majority being low intensity (84%) caused by the displacement, in addition to 18 accelerations per minute, almost all of them of low intensity (<3 m/ s 2 ) with 96% of all. Concerning the specific position, the wing is the one with the greatest requirements, especially in high-intensity actions, the backcourt one is the one with the greatest number of speed changes and finally, the centers and the defenses are the ones with the greatest demands at low intensity. Finally, the winner of the match presented higher values in the relative distance, high-intensity distance, and high-intensity speed changes with large effect sizes. Concerning the distance and speed parameters, after analyzing the results of all the players together, we obtain an average distance traveled of 2551.42±1102.12 meters lower value than in the studies of Belka et al. 20 and Michalsik et al. 13 with results of between 3399-to-4002 meters covered. Regarding the distance covered per minute, 72.12 meters were obtained, which resembles the results of other studies 20,21 . These differences between the total distance covered and the distance covered per minute are due to the time that each player was on the court since in amateur handball there is usually an equitable distribution of the minutes of play among all aligned players. In this way, a smaller total distance is covered, but the parameters of distance covered per minute are kept high. Regarding the intensity of the displacements, different studies the studies carried out in senior male and female elite-level handball showed that most of the distance traveled was done on walking or at low-medium intensity, below 12 km/h (76% of the distance covered) 13,14,20 . Therefore, in this sport modality, the development of aerobic metabolism is important for the improvement of sports performance. Regarding the neuromuscular load, most of the impacts made did not exceed 5G (92%), making a total of 54.23 impacts per minute over the entire range of intensities. There are no studies that show data on impacts greater than 8G forces performed by handball teams during the competition. However, if there is information about impacts made at more than 5G of force during the match 22 with results very similar to those of the present study on impacts made at more than 5 G (13.6 impacts >5G). Jumps are the specific action in which lower values are found in comparison with the literature in elite handball. Chelly et al. 23 and Póvoas et al. 14 in junior and senior handball found values of 6.9 and 4.9 jumps per player during the match respectively, higher than the 4.26 jumps recorded in this study. Finally, regarding the PL variable, there are demands per minute of 0.75±0.18 a.u./min. These demands are lower than those found in elite handball (9.52±1.1 vs. 7.5±1.8).
Finally, regarding the quantification of speed changes, it has not been studied in depth until now 24 . In this study finds an average of 0.70 accelerations per minute and 1.05 decelerations per minute above 2.5 m/s 2 , similar to those of the present study. Instead, Manchado et al. 25 find higher results in acceleration and deceleration 1 and 2 with respect to those obtained in the present work. The causes of this great difference are due to the level of the participants between investigations. Therefore, the individualization of the load is essential for the improvement of performance between categories and sports levels, due to its specific requirements.
In relation to the results obtained, the end is the specific position that carries out the greatest external load, since it obtains results greater than the rest in the parameters of distance traveled per minute, displacements at high-intensity, and accelerations and decelerations of high intensity. These results are similar to those found by Cardinale et al. 21 and Michalsik and Aagard 6 . These results could be explained because the extremes are the players that make up the first wave of the counterattack every time there is a change of possession, either after a goal kick of the opposing team that was not between the 3 sticks, for a goalkeeper stop, or a steal 26 . Also, in this match, team A maintained a 5:1 defensive system throughout the match, whose outpost was always occupied by the end of the side, depending on the moment of the match. This advanced defender deals with covering pass lines between the players of the first line of attack, and to be effective, he must make a large number of changes in direction and pace, being this outpost of high physical, technical and tactical qualification 27 .
In this study, the center, backcourt, and wing positions are those that receive the highest number of impacts greater than 8G forces per minute on average (15.67 center, 15.50 backcourt, 14.57 wing), while the pivot position and defensive specialist were those with the lowest results (3.00 and 3.50 respectively), which is the opposite of the results obtained by Póvoas et al. 14 where the highest values of impacts> 5G are received by the pivot. The difference between studies may be because most of the impacts received by the pivot are 5-8G of force, but to know for sure it would be convenient to carry out future studies with that objective.
Finally, concerning speed changes, the wings also obtained the highest number of accelerations and decelerations per minute at high intensity, while in zone 1 the results are more balanced among all players. In the study by Manchado et al. 25 the same results were obtained in the accelerations >3m/s 2 and in the decelerations <-3m/s 2 (zone 4), while in the acceleration and deceleration zone 1 of this study the wings remain the ones with the highest number of accelerations and decelerations per minute performed.
This study shows how the winning team has greater relative distance, high-intensity distance, and high-intensity accelerations and decelerations compared to the winning team. This fact has been analyzed in handball in relation to technical-tactical actions 28,29 but not related to external load demands. This aspect has been studied in other sports disciplines such as basketball 10 or soccer 30 finding greater efforts of high-intensity actions in winner teams than the losing teams. Therefore, teams that have the highest acceleration and high-intensity displacement capacity are more likely to win the games compared to teams with a higher volume of low-intensity demands in their game, especially in offensive actions.
Although the results of the present investigation make the first approximation in the description of the external load carried out by two senior handball teams of amateur level, finding differences between the requirements for specific positions and identifying physical performance indicators between the winning and losing teams, some limitations should be considered. The results obtained cannot be extrapolated to the rest of the sports categories and levels, even to the specific characteristics of the study population due to the small sample (an official competition game in regular phase). Although, this research is a pioneer because it is the first that describes the demands in indoor handball using a new radio frequency-based tracking technology (UWB). Future research should deepen the study of amateur handball, both at the senior level and the youth level, due to the lack of study at this sports level in addition to where there is the greatest number of licenses and research in this area would have a greater social impact.

CONCLUSIONS
Based on the results obtained from this study, we can identify the following conclusions and practical applications: 1. Amateur senior handball players covered an average distance of 2.5 km during the match, making 76% of that distance at a speed of less than 12 km/h, which shows the great prominence of aerobic metabolism during the handball competition. On average, they travel 309.56 meters with an acceleration greater than 1.12m/ s 2 , receive 11.47 impacts greater than 8G of force, and perform 4.26 jumps per game. 2. The wing is the player who receives the highest external load in almost all parameters in relation to the rest of the specific positions, and it is because that the wing is the player who performed most of the team's short, explosive and anaerobic actions during the competition. The center makes the highest number of low-intensity demands and defensive players are the ones who receive the highest number of impacts due to their role during the game. 3. The intermittent actions of high intensity (relative distance, distance at high intensity, accelerations, and decelerations of high intensity) are those that are performed in the most important situations of the game and that differentiate success in the final result. Therefore, handball players must have a high-level of anaerobic power and the ability to perform high-intensity action repeatedly during the competition.