Predicting Bobsled Pushing Ability From Various Combine Testing Events.

Area of Science:

• Sports Science
• Biomechanics
• Human Performance

Background:

• Selecting elite bobsled athletes is challenging due to the difficulty in directly measuring pushing ability.
• Current combine testing protocols by USA Bobsled and Skeleton use numerous variables to predict push performance.

Purpose of the Study:

• To determine the most effective physical performance variables for predicting bobsled push ability.
• To identify redundant or irrelevant variables in existing combine testing for bobsled athletes.

Main Methods:

• Collected data on 11 physical performance variables from 75 subjects across two Olympic qualification years.
• Utilized discriminant analysis (DA) and principle component analysis (PCA) to analyze the data.
• Investigated two cases: Olympians vs. non-Olympians and National Team vs. non-National Team members.

Main Results:

• Initial DA with 11 variables achieved misclassification rates of 9.33% (Olympians) and 14.67% (National Team).
• PCA identified redundant variables, leading to optimized combinations.
• Reduced variable sets using DA resulted in significantly lower misclassification rates (as low as 6.67% and 13.33%).

Conclusions:

• Fewer, well-selected combine testing variables can accurately predict bobsled push performance.
• Optimized combine testing protocols can enhance athlete selection accuracy while minimizing irrelevant data collection.
• This approach can be applied to other sports to improve combine testing efficiency and effectiveness.