How different are soccer training sessions based on small-sided games? A cluster analysis to explore perceived exertion and training load

Background: Soccer coaching staffs may have difficulties in classifying properly the load accomplished during training sessions and understanding how players rate the effort and the training load during the micro-cycles. Objective: The aims of this study were to investigate if the features and duration of training tasks can automatically classify the different weekly training sessions into different clusters, and to describe and compare the features and duration of training tasks, rating of perceived exertion and training load from the different training sessions across both classifications (prescribed by staff per day vs. automatically per cluster). Methods: Eighteen elite youth male soccer players reported their rating of perceived exertion 10 min after each practice during twelve micro-cycles. In each micro-cycle, differentiated management of the task characteristics for each training day was implemented. A Random Forest Clustering was used to automatically assign each training session to one cluster and allowed to create similar groups and contrast them with the sessions prescribed by the coaching staff. Results: Proper manipulation of different variables of small-sided games (number of players, relative playing area per player, game orientation, training regimen) can automatically differentiate training sessions. Youth soccer players can also perceive each training session differently (p < .05). Conclusions: Using different formats of small-sided games in each training session may be interesting to modulate the player's perceived load during the micro-cycle. However, results should be interpreted with caution, due to representing a single team and coaching staff.