Validity and Accuracy of Impulse-Response Models for Modeling and Predicting Training Effects on Performance of...

Area of Science:

• Sports Science
• Exercise Physiology
• Biomechanical Modeling

Background:

• Effective training planning is crucial for optimizing athletic performance.
• Impulse-response models are used to quantify the relationship between training load and physiological response.
• Previous models have varying abilities to account for training history and its impact on performance.

Purpose of the Study:

• To compare the suitability of different impulse-response models for practical application in training planning.
• To evaluate the predictive accuracy of selected models for individual athlete performance.
• To identify the best-performing models for optimizing training and taper strategies.

Main Methods:

• Six impulse-response models were tested, including Banister's model (Model Ba) and a variable dose-response model (Model Bu).
• Data from 11 swimmers over two seasons were collected, including daily training load and 50-m performance trials.
• Models were ranked using Akaike's information criterion and goodness-of-fit measures, with predictive accuracy assessed against validation datasets.

Main Results:

• Models Ba and Bu received the highest Akaike weights, indicating superior performance among the tested models.
• These models were used to estimate performance evolution and optimal taper characteristics.
• Mean absolute percentage errors for performance prediction ranged from 2.02% to 2.69% for Model Ba and 2.17% to 2.56% for Model Bu across validation datasets.

Conclusions:

• The top-ranked models (Ba and Bu) provided reasonable approximations of the training-performance relationship.
• However, their predictive capability for future performance was insufficient for effective individual training planning.
• Further refinement of models is needed to enhance their utility in personalized coaching and athlete development.