Predicting alpine skis on-snow evaluations from measured physical attributes

Engineering of Sport 15 - Proceedings from the 15th International Conference on the Engineering of Sport (ISEA 2024)

Attribute selection methods have found extensive applications in various domains to gain a deeper understanding of user sensory experiences. For instance, these methods are employed to discern the physical attributes influencing coffee taste. Such analyses typically necessitate substantial testing to build large enough datasets appropriate for the use of statistical methods. This requirement elucidates the persistent challenge of correlating user feedback with the physical attributes of sporting equipment. In the context of alpine skis, numerous on-snow evaluations are imperative, incurring significant costs, time, and effort. Additionally, the physical attributes of skis, primarily provided by ski manufacturers, are often limited in scope and lack standardized formatting. The recent publication of two distinct datasets have created a new opportunity for evaluating and modeling the performance of alpine skis based on their physical attributes. These datasets include (1) the Blister Gear Guide, a well-known ski review website, which publishes ski reviews for over 200 skis each year, and (2) SoothSki, which provides the detailed measurements of over 4000 commercially available skis. Consequently, the primary objective of this research is to develop a methodology to construct simple predictive models of Blister’s rankings. These models will utilize SoothSki's measured attributes as input and Blister's rankings as output, to ultimately offer valuable insights on ski design while also helping skiers select skis.