Ofoghi, B., Stefano, D., Zeleznikow, J., & McMahon, C. (2012). Modeling relationships between swimming attributes for performance prediction. Presentation 1935 at the 59th Annual Meeting of the American College of Sports Medicine, San Francisco, California; May 29-June 2, 2012.

Swimming is a high profile sport in which an athlete with the fastest overall time wins and even a very small difference in overall times is decisive for medal winning. The inter-relationships between different swimming attributes and the overall time can be used to predict and optimize overall times hence to increase the medal winning chance.

This study modeled pair-wise inter-relationships between swimming attributes and the overall time, for predicting overall swim times and training for optimal strategies. A linear regression analysis was performed to model the pair-wise interrelationships between swimming attributes and overall times in specific subsets of swims from a database of international events. Data were filtered based on gender, age group, stroke, pool size, and distance.

In many cases no significant relationship was found between attributes. This could have been the result of many missing values for those attributes. In some cases, no values were reported for the performance attributes. In other cases (e.g., breath count in M, freestyle, 50m, lcm, open swims), a significant relationship was found. The model could be used to predict overall times.

Implication. The linear regression analysis of swimming attributes requires a more sophisticated analysis that can handle a large number of missing values. Where enough data exists, the model can be used effectively to predict overall swim times.

Return to Table of Contents for Hydrodynamics of Swimming.