Die Zukunft der Leistungsdiagnostik im Nachwuchsfußball: Quantitative Spieldatenanalyse zur Erstellung von positionsspezifischen Anforderungsprofilen und Talentprädiktoren

Publication: Book/ReportDissertations


The aim of this study was to show the possibilities of how objective game data can be integrated into the talent development process and thus contribute to improving the development of young talents in Germany. In the first part of the study it can be shown that objective parameters, which were selected by a group of experts to describe the playing philosophy of a team, vary in number and effect strength within the examined clubs across the age groups, depending on the club statistically relevant differ from each other. Across all clubs, in a sample of 6166 games, the parameter „game pace“ shows the greatest effect size between the youth and professional teams with ɲ² = 0.147, but also for other parameters, such as the „shots at goal per minute“ with ɲ² = 0.093 a moderate effect is shown. In the context of this study it can also be shown that the „consistent instruction“ (measured by indices that take into account the significant differences in the parameters collected) in the sense of the EiJP, shows a significant relationship (rP = .560) to the number of future professional players. This effect on the future BIG5 players is even greater (rP = .693). On the other hand, no statistically relevant influence (p = .697) can be shown on the later average market value of the players. The same can be seen for the number of professional debuts in one's own club, but with a different level of significance (.092≤p≥.972) for the various indices. This study also shows that the U23 teams are most similar to the professional team in terms of their playing philosophy. While the great effect in the area of the „game pace“ is in line with the prevailing literature, the findings in the consistent instruction of the U23 to the professional team and the form of the parameters from the offensive game in the junior area are critically discussed, as the results of this study are contrary to the non-evidence-based opinion of the media. In the second section of the study, numerous position-specific parameters can be identified at the individual player level, which differ statistically relevant between the individual age groups. For this purpose, a total of 39173 single player data sets were analyzed (CD: 11862, FB: 9002, CM: 11899, W: 2159 and S: 4251). single player data set analyzedThe position-specific demand profiles available for the individual age groups can be used to adapt the training content in the youth teams and, if necessary, to revise the competition mode in order to prepare the players even better for the later demands in the professional teams. Likewise, existing confidence intervals can also be used to find alternative positions for players. In the last part of the study, predictors can be identified by means of multinominal logistic regression using the later performance class of the examined players, which make it possible (with weak to large effect sizes (0.1 <d> 1.7)) to differentiate between the later performance classes „no professional league“, „professional leagues outside the BIG5 leagues” and „BIG5 leagues”. The results differ in their effect size and significance level between the individual age groups and positions. For fast, practicable scouting information, the specified confidence intervals can also be used here to better assess the player and his perspective. The present study shows the enormous potential of objective game data and outlines the wide possibilities from training control to scouting relevance to modern in-game performance diagnostics and thus offers a good basis for further research in objective game data in (youth) academies. Further research should aim to determine whether the identified predictors are also applicable to younger age groups and whether more complex KPIs, such as space control or the pressing index, are even more relevant within the meaning of the talent prediction.
Original languageGerman
Place of PublicationBerlin
PublisherWissenschaftlicher Verlag Berlin
Number of pages383
ISBN (Print)978-3-96138-330-6
Publication statusPublished - 2022