| dc.description.abstract | Background: Soccer is dynamic and characterized by continual changes in physical and tactical movements. While having ball-possession and attacking, a common tactical movement of a player is to run behind the defensive line of the opposition. This movement aims to disrupt the defensive structure of the opposing team and gain positional advantage. A run behind the opponents’ defensive line, also known as run in behind [RIB] might end in a one versus one situation with the goalkeeper, which significantly increases the opportunity to score a goal. There is a gap in literature analysing the efficiency of such running behaviour, determined by the creation of goal scoring opportunities [GSO]. Moreover, GPS metrics like velocity and acceleration are rarely used to analyse specific tactical movements. This study aimed to investigate the relationship between velocity and acceleration characteristics of RIBs and GSOs in professional soccer matches. This was further examined for different types of attacks.
Methods: A senior professional male team of the Norwegian Eliteserien was followed across 30 matches of the 2023 season. Video was used to classify movements as RIBs by two qualified coaches. RIBs were further classified in “pass attempted” when a pass to the player was attempted, “pass successful” when the player gained control over the ball after the pass, and “goal scoring opportunity” when the player had a significant involvement in an attack where a GSO was created. Furthermore, each RIB was categorized into one of the four types of attack: build-up play, direct play, counterattack, counter-press. The difference in maximal velocity, maximal acceleration, and time to maximal acceleration between RIBs ending in GSOs and RIBs not ending in GSOs were assessed. This was done using the independent samples t-test or alternatively the Mann-Whitney U test. Effect sizes [ES] of differences were reported as Cohens’d or the point-biserial correlation r respectively. Furthermore, inter-group differences for RIBs grouped by types of attack were investigated. This was done using the Kruskal-Wallis test. The ES of inter-group differences was reported as the point-biserial correlation r.
Results: Out of the 1089 RIBs, 59.4% were classified as build-up play, 16.4% as direct play, 18.2% as counterattack and 6.0% as counter-press. The proportion of RIBs ending in GSOs ranged from 7.26% to 10.10% across types of attack. Maximal velocity was higher for RIBs ending in GSOs for all attacks together (p < 0.001; r = 0.17, small ES) as well as for RIBs during build-up play (p < 0.001; r = 0.2, small ES) and direct play (p = 0.022; r = 0.17, small ES). Time to maximal acceleration was higher for RIBs ending in GSOs for all attacks together (p = 0.007; r = 0.08, less than small ES) as well as for RIBs during build-up play (p = 0.019; r = 0.09, less than small ES). Inter-group differences for types of attack were found for maximal velocity for all RIBs (p < 0.001), where maximal velocity was higher in counterattack compared to build-up play (p < 0.001, r = 0.13, small ES), direct play (p < 0.001, r = 0.37, moderate ES), and counter-press (p = 0.001; r = 0,23, small ES). Maximal velocity was also higher in direct play compared to build-up play. Significant differences were also found for RIBs ending in GSOs (p = 0.027), with counterattack showing significantly higher maximal velocity than build-up play (p = 0.042; r = 0,1, small ES). Inter-group differences for types of attack were also found for time to maximal acceleration for all RIBs (p < 0.001), where time to maximal acceleration was significantly lower in build-up play compared to counterattack (p = 0.028; r = 0.1, small ES) and counter-press (p = 0.004; r = 0.13, small ES).
Conclusions: The results indicate that maximal velocity was a key component in creating GSOs. This was especially notable during build-up play and direct play. Differences in acceleration characteristics were present but due to sample size and many outliers the interpretation of the results is limited. Data processing techniques such as filtering are suggested for acceleration data. Further research should include larger sample sizes across several teams. This may allow to estimate regression models and give a more comprehensive understanding of the relationship between physical metrics and creation of GSOs through RIBs. The findings from this explorative approach are limited but might still be taken in consideration for training design, player selection and match strategy. | |
