A machine learning approach to “revisit” specialization and sampling in institutionalized practice

Abstract

Apart from a broad consensus statement stressing the essential role of practice for achieving success in international senior-level competitions, the nature and scope of developmental participation leading to that extraordinary success in sports have been controversially discussed in international literature for many years. The aim of this paper is to contribute to the existing body of literature in two respects: first, by reviewing the existing literature comparing the developmental activities of internationally and only nationally successful senior athletes. Second, a new methodical approach combining decision trees and gradient boosting is applied to data from a previous study, the results of which were internationally published. This does not only allow for the realization of a multivariate analysis (robustness check), but also gives reasonable hope of achieving a relatively better explanation than with the procedures applied in the past. The approach is realized by means of Extreme Gradient Boosting (XGBoost under the R environment). The results indicate that some formerly found differences in the volume of structured practice in main and other sports between internationally and only nationally successful athletes may represent rather artifacts of uncontrolled age effects than variables that differentiate the groups. In the context of the specialization-diversification debate, the present results indicate that from today’s perspective there is a debate about a “production function”, the structure of which is unknown. Obviously, practice-related recommendations on developmental practice volume are expressions of highly rationalized myths rather than evidence-based efficient norms.