Accurately identifying a teenager who will go on to succeed at the highest level is an on-going issue in all sports. Researchers interested in talent identification have attempted to develop a range of tests that cover the breadth of skills in a particular sport.
An issue with many attempts to develop assessments that accurately predict talent is that they are (a) biased towards one specific component (e.g., physical tests) or (b) they have focused on age-groups that are experiencing considerable maturation (hence, giving misleading results).
A multi-dimensional approach focusing on players in the latter years of the youth pathway is therefore likely to produce the most accurate results.
Carl Woods and colleagues assessed how accurately a multi-dimensional assessment approach could predict talent in Australian football. Of a sample of 379 under 18 players from the Western Australian Football League, 42 players were considered to be “talent identified” as they were selected in the state team (i.e., the highest level in under 18 Australian football), while another 42 players were randomly selected from the remaining cohort. These players were referred to as “non-talent identified”
The talent identified (n = 42) and non-talent identified (n = 42) players completed a battery of assessments. These included:
- Standing height
- Dynamic vertical jump non-dominant leg
- 20 m multi-stage fitness test
- Kicking test
- Handballing test
- Video decision-making task
The results of each assessment were included in a statistical model to predict whether the player was “talent identified” or the “non-talent identified”. The model was 95% accurate in its ability to predict talent.
This was more accurate than previous studies in Australian football that have included either all physical tests (84%) or technical (skills) tests (89%) in the assessment battery.
Given that the model was not 100% accurate, it means that some players classified were incorrectly. Two talent-identified players were not identified by the assessments, while six “non-talent identified players” were misclassified as talented (this might reflect the fact that only 42 players can be selected in a state squad).
This study demonstrates the effectiveness of a multi-dimensional approach for assessing players who are currently identified as talented. However, the biggest challenge with talent identification is being able to predict the players who will be skilled in the future. It would be great if the players in this study are tracked over the coming 5 years. This would provide strong empirical evidence regarding the effectiveness of the assessment battery.