Using telecasting to identify key performance indicators in alpine skiing and evaluate the inter-analyst reliability of alpine ski coaches
2016 (English)Conference paper, Poster (Refereed)
ntroduction The most frequently used analysis method in alpine skiing is video footage. However, using video footage to identifying key performance indicators (KPIs) and other characteristics are difficult as each coach analyzes video footage in a subjective manner, which decreases the reliability of the analyses. Even though alpine skiing is a “closed sport”, without any external disturbing moments, each race is unique therefore never identical. It is thus difficult for coaches to know the validity and reliability of their analyses as the majority of the video footage of alpine skiers is captured during training, not racing. A large number of analyzed races with high reliability could therefore facilitate to identify KPIs in alpine skiing. The purpose of this study was to use the standardized typical error to explore the possibility to use telecasting to analyze performance and skiing characteristics in world cup slalom races. Method Standard telecasting footage (25 fps) from four different WC-races (eight runs) was used for analyzing turn times for eleven skiers (nine males and two females). The footage was analyzed at three separate situations, by two different alpine World Cup coaches and one video analyst. Turn time was defined as boot passage of the gate and all video analyses were made in DartFish ProSuite 6. Typical error (TE) was calculated by; TE=σ_diff/√(N_obs ) The standardized typical error (STE) was calculated by; TE= TE/√(((n_1-1)^2 σ_1^2+(n_2-1)^2 σ_2^2 ) /((n_1+n_2-2))-TE^2 ) The index described by Hopkins  was used for evaluating the influence of the STE. Results The mean turn time for the male skiers was 0.83 ± 0.18 s, with a coefficient of variation of 22%. The mean turn time for the two female skiers was 0.85 ± 0.19 s and the coefficient of variation was 22%. The TE between the different analyses was 0.03 s and the STE was 0.14. The results provide a detailed analysis of gate-to-gate times for each skier. Discussion The TE value of 0.03 s is most likely due to the 25 fps telecasting footage, where each frame is 0.04 s. The analysts must therefore choose one frame if the point of interest is between two frames. Hence, TE will presumably decrease with increase framerate. As the STE < 0.2, the disagreement between the different analysts can be considered as trivial [1, 2]. We here show how the STE can be used to identify inter-analyst reliability of alpine skiing video analyses. Furthermore, the presented method provides a robust, cheap and effective method to objectively analyze skiing performance and identifying plausible key performance indicators in alpine skiing. References 1. Hopkins, W., Reliability from consecutive pairs of trials (Excel spreadsheet). A new view of statistics, 2000. 2. Liu, H., et al., Inter-operator reliability of live football match statistics from OPTA Sportsdata. International Journal of Performance Analysis in Sport, 2013. 13(3): p. 803-821.
Place, publisher, year, edition, pages
Sport and Fitness Sciences
IdentifiersURN: urn:nbn:se:miun:diva-29690OAI: oai:DiVA.org:miun-29690DiVA: diva2:1058032
ICSS 2016 - International Congress on Science and Skiing, 10-15 December, 2016, St. Cristoph, Austria.