International Journal of Performance Analysis of Sport 2010, 10, 221-228.

General trends in European men’s handball: a longitudinal study Meletakos P1, and Bayios I1 1

Faculty of Physical Education and Sport Sciences, University of Athens.

Abstract The present study investigated in a longitudinal manner from seven consecutive competition years (2002-2003 through 2008-2009) the final results of men handball National Major League matches from seven European countries (Denmark, France, Germany, Greece, Poland, Spain and Sweden). The results showed that there were significant differences between the countries with regards to the total number of goals scored per match. Nonetheless all the countries followed a consistent pattern of significant increase in the total number of goals scored per match, which over the years amounted to an increase of roughly five goals per match, from 52.9 to 57.9 goals, meaning an increase of 10%. The seven countries were heterogeneous also with regards to the percentage of close games i.e. of matches with a goal difference of two or less, as well as with regards to the outcome of the matches (home win, away win, and draw). The present findings provide experts in handball with valuable information regarding general trends in modern handball with regards to pace of the game, home advantage importance and the quantification of competitiveness based on the results of close games as well as specific differences between countries in Europe. Key words: pace of the game, home advantage.

1. Introduction Handball is a sport which is relatively new in the world sports arena as it is played today. Initially it was played by 11 players on a team and on an outdoor court. It transformed into a 7 players per team sport in the 1960s in indoor courts. Its development is based on technical, tactical, and physical preparation. It is believed that another important aspect for the development of team sports is the competitive environment of play where the structural constituents of the coaching process are interdependent and codependent. The concept of environment is mentioned in the science of coaching as enhancement of athletic performance.We understand the competitive environment of play to be a dynamic situation in which the teams’ performances come into play and the performances themselves in their own right form and develop this concept.The importance of the term environment is seen in the field of psychology as laid down in the causal structure involving triadic reciprocal causation (Bandura, 1986). One of the constituents of environment in sports is the home advantage (the probability of home win versus visitor win). According to Courneya and Carron (1992), home advantage is defined as the consistent finding that home teams in

221

sports competitions win over 50% of the games played under a balanced home and away schedule. Nevill and Holder (1999) identified 4 parameters connected with home advantage, these being crowd, learning, travel, and rule factors. Other researchers dealing with Home Winning Percentages (HWP) in different sports and different leagues support the importance of Home Advantage (Adams and Kupper, 1994; Agnew and Carron, 1994; Moore and Brylinsky, 1995; Nevill et al., 1996). However there are studies dealing with Competitive Balance (Zimbalist, 2002; Rodney and Maxcy, 2003; Sanderson and Siegfried, 2003) which approached this aspect solely from an economics and attendance viewpoint, ignoring the phenomenon of home advantage (Forrest et al., 2005) and its role in the match outcome. Besides the percentage wins of the home teams versus the visiting teams, this study will examine the total goals recorded in a match, the range of the difference in goals between winning and defeated teams and the percentage of close games (outcome uncertainty). Research by Taborsky (2003) indicated that the total number of goals scored has an upward trend in Women World Championships from 1995 to 2003, as the average number of goals scored per match ranged from 46,3 to 55,5. Similarly, research by Johansson and Spate (2004) and Spate (2005) in analysing the 2004 Athens Olympic Games (men’s tournament) and the Men’s World Championship in Tunisia, found that the total number of attacks per game showed a pronounced increase. In particular, in Athens 2004 the average was 115,7 attacks per game. This was thought to be the peak and that no further increase was possible but the results of World Championship in Tunisia (2005) disproved this hypothesis where 120,2 attacks per game was recorded, 60 attacks per team on average for all 86 games. This meant that the attacks were completed after an average of 30 seconds. With respect to goal difference between winner and loser in the Women’s World Championships between 1995 and 2003, the average difference recorded was 3,6. However, this outcome refers to the results of games between the top 8 teams only (Taborsky, 2003). Finally, it is possible that close games (the equal possibility of either team winning) is an important ingredient which defines the competitiveness of competition in League or Olympic Tournaments, European and World Championships. The above variable has been used to investigate, together with balanced and unbalanced games, the discriminatory power of game statistics between winning and losing teams in basketball (Sambaio, 2003 & 2005; Gomez et al., 2006) and Drikos (2009) in volleyball did not manage to create a discriminant function for those sets with a final difference between the two teams of 2-3 points.

2. Methods 2.1. Samples and variables Our data were collected from the official final results from the archive of the National Federations of Handball. Playoffs, games of relegation, forfeit games, and National Cup games were exempted from the statistical sample. Therefore the only handball games which were analysed, were those played during the regular season in the Men’s National Major League. Our sample contained 10,358 final scores from seven European National Championships; France (1274), Germany (2142), Spain (1686), Poland (1243),

222

Denmark (1317), Sweden (1641) and Greece (1055). These data commence from the season beginning 2002/2003 up until the season ending 2008/2009.The final scores per year was, 1405 (2002), 1437 (2003), 1630 (2004), 1473 (2005), 1449 (2006), 1446 (2007), 1498 (2008). The dependent variables were the total number of goals scored per game, the goal difference between the winner and the loser, expressed through the dichotomous variable of close matches with a goal difference of two or less and open matches with a goal difference of more than two goals and, finally, the match outcome, i.e. whether the match was won by the home or the away team or whether it was a draw. The independent variables were firstly the country and secondly the competition year. 2.2. Statistical analysis The dependent variable of goals scored per game by both teams was entered in a twoway analysis of variance model with the seven countries and the seven competition years as the independent factors. This was followed by post-hoc tests with corrections for multiple comparisons in order to establish the homogeneous subsets of countries or com petition years. The k-means clustering method (Tabachnik and Fidell, 2007; Norusis, 2009) was employed in order to categorize the games according to the game final goal differences. The first cluster included all the games with a goal difference of two or less. These games were categorized as close games, while all the other games were categorized as open games. This binary variable was called game kind. Chi-square tests tested the dependency of this variable on the country and the competition year. Chi-square tests with the same independent variables were also employed with regards to the match outcome, i.e. whether it was a home or an away win or whether it was a draw. The level of statistical significance was set at 0.05.Statistical significance was set at 5% and all analyses were performed in SPSS 16.0.

3. Results The two-way analysis of variance with the total number of goals scored per game as the dependent variable revealed a very significant effect both of the country factor (F6,10309=90.0, p