Modelling Shooting Performance in Elite Men’s Wheelchair Basketball

Shooting is one of the key technical skills required by wheelchair basketball players (Zwakhoven, Evaggelinou, Daly, & Vanlandewijck, 2003, European Bulletin of Adapted Physical Activity, 2 (3)). Previous studies have focused on examining the technique of free-throw shooting (Goosey-Tolfrey, Butterworth, & Morriss, 2002, Adapted Physical Activity Quarterly, 19, 238–250); however, free-throws only equate to approximately 15–22% of the total number of shots taken during a game (Gomez, Perez, Molik, Szyman, & Sampaio, 2014, Journal of Sports Sciences, 32 (11), 1066–1075). To date, no attempts have been made to explore the determinants of two-point and three-point shooting. This study explores the key determinants of two-point and threepoint shooting and develops a valid prediction model. Following ethical approval, footage of all of the twopoint and three-point shots taken when the top five teams played each other (nine games; 1144 shots) at the 2015 Men’s European Wheelchair Basketball Championships was analysed in SportsCode (version 10, SportsTec Ltd). A valid and reliable shooting specific performance analysis template was developed by three elite wheelchair basketball coaches and the lead researcher (Inter-operator reliability ± 1: 0.00%; Intraoperator reliability ± 1: 0.00%). The template contained 61 action variables within 20 categories: Point, Quarter, Match Status, Classification, Shot Hand, Shot Handed, Shot Type, Shot Clock, Pre-Shot, Shot Movement, Shot Positioning, Zone, Pressure, Number of Defenders, Defender Marking-Shooting Hand, Defender Marking-Non-Shooting Hand, Defender Marking-Space and Defenders Positioning-Defender In Front, Defenders Positioning-Defender Behind and Defenders Positioning-Defender on Side. Chi-squared test highlighted seven categories were significant (P < 0.001) when shot outcome was compared (successful versus unsuccessful). The Shot Positioning category was the most significant category and produced a P-value of 2.2e-16. The forward selection method was used and a 30% sample of the data set selected to create a shooting specific generalised linear regression model that involved 10 categories. The model was tested against the entire data set and an area under the curve value of 0.850 was achieved. The model demonstrated that when an athlete was able to “catch and shoot” in a “square to basket” position whilst “stationary” and taking a “set shot” from the “two-point centre long” zone whilst a defender is placing “no pressure” on the shooter the individual achieved a 96.87% success rate. The findings provide objective evidence of the key determinants of shooting success. Coaches are able to utilise this information to adjust training sessions and game strategies to meet the needs of the performer.