Moving towards an evidence informed approach to talent identification and development in Scottish soccer long term athlete development pathways.

Professional soccer academies aim to optimise the identification of talent and the development of players to create a resource as first team players or sellable assets. Since the turn of the millennium, the literature that examines the impact of talent ID and development practices has predominantly been cross-sectional in nature, lacking in longitudinal analysis. There is a lack of objective analysis that supports practices and decision making with progress and performance decisions being subjective in nature, normally lacking quantitative evidence. By understanding what practices have been successful and can increase efficiency and productivity for academies, applied research can support practitioners in their future planning and development. Papers I, II and III utilised physiological data collected over a ten year period in practice from an elite Scottish soccer academy. Paper IV examined the impact of seasonal data related to transition to first team collectively investigating both longitudinal trends and practices facilitating the journey to professionalism. Retrospective analysis was conducted utilising linear mixed effect models across physical performance parameters (papers I, III and IV), Pearson's Chi squared and odd's ratios on relative age impacts (papers II and III) and least absolute shrinkage and selection operator (LASSO) logistic regression with multiple imputation and bootstrapping (n = 10,000) to consider the predictive value of physical assessment on the likelihood of being awarded a professional contract (paper II). Over a ten year period a significant improvement in 5m, 10m, and 20m sprint performance time by 0.0055 s per year (t=−11.8, p<0.001), 0.008 s per year (t=−9.2, p<0.001), and 0.011 s per year (t=−7.8, p<0.001) was observed. Speed, power and aerobic progression were related to an increase in anthropometric measures (Paper I). Professional contracts awarded were impacted by the academy cohort showing a relative age bias resulting in relative age being the greatest predictor of whether players were awarded professional contracts (Q1= 0.50; Q2= 0.26; Q3= 0.20 and Q= 0.05 (x2 3= 40.9, p<0.001)) (Paper II). Predictive modelling performed similar to random guessing (0.45 (0.41–0.49) error proportion obtained with the test set) with large overlapping and variation between those awarded professional contracts and those released. Disappointingly, ANOVA and post hoc tests showed no significant effect of time in changes in RAE (p=0.3217). Relatively younger players who are continually selected do so as they have the physical capacity to match their older peers, potentially due to advanced maturation (Paper III). When considering transitions to facilitate players to first team, players on development loan may be experience lower physical exposures with non significant lower weekly values than squad players for all variables ranging from 5.2% (weekly sRPE) to 16.8% (weekly sprint distance covered) (paper IV). Academies should consider conscious or subconscious bias towards recruitment on physical or arthrometric terms influenced by relative age or maturation, rather than overall technical soccer ability. Whilst there is a lack of predictive capability from testing results, these results should be utilised towards individual development considerations for each player. These may include low cost interventions to address negative impacts of relative age or maturation. The ability to carry out advanced multidisciplinary modelling may soon be available to practitioners who have maintained valid and reliable longitudinal data related to both legislation such as EPPP and technical enhancements.

Author dc:creator

• Craig, Thomas Peter

Advisor dc:contributor.advisor

• D. Reid and K. Cooper

• Soccer
• Football
• Professional development
• Player development
• Talent development
• Talent identification
• Talent scouting
• Professional sports contracts
• Physiological factors

Language dc:language

en

Identifier

oai:rgu-repository.worktribe.com:3452409
https://doi.org/10.48526/rgu-wt-3452409

Author Identifier

0000-0003-4050-8522

OAI identifier oai:identifier

oai:rgu-repository.worktribe.com:3452409