Modelling Shooting Performance Across Major International Tournaments in Elite Men’s Wheelchair Basketball

The ability to score baskets is the key technical skill that separates winning and losing teams in men’s wheelchair basketball (Gómez, Pérez, Molik, Szyman, & Sampaio, 2014). Previous studies have focused specifically on the technical aspects of free-throw shooting (Goosey-Tolfrey, Butterworth, & Morriss, 2002; Malone, Gervais, & Steadward, 2002) with limited research exploring the technical and tactical components of field-goal shooting (Francis, Owen, Molnar, & Peters, 2016). Francis et al.’s (2016) work explored the technical and tactical factors affecting two-point and three-point shooting in the 2015 men’s European Wheelchair Basketball Championships, identifying shot movement and shot positioning as being significantly related to shot outcome (successful versus unsuccessful; p < 0.001). This study explores the key determinants of two-point and three-point shooting across two major tournaments and develops a valid prediction model. Following ethical approval, footage of each two-point and three-point shot taken by the top five teams (30 games; 2432 shots) at the 2016 Paralympic Games was analysed in SportsCode (version 10, SportsTec Ltd) using the same performance analysis template and combined with the shooting data from the 2015 European Wheelchair Basketball Championships (nine games; 1144 shots; Francis et al, 2016) to provide data for 3576 shots. Chi-square tests highlighted 17 categorical variables that were significantly associated with shot outcome (p<0.05), however, the tournament in which the shot was taken at was found to be a non-significant categorical variable. The most significant categorical variable was the ‘Shot Location’ in which the shot was taken. A binary logistic regression model was developed to assess the impact of these categorical variables as predictors of the probability of shooting success, using an automated stepwise selection method with 70% of the sample (2,499 shots). The final model included thirteen statistically significant predictors and when tested against the remaining 30% of the data an area under the curve (AUC) value of 0.642 was achieved.
These findings support the previous model (Francis et al. 2016) and re-emphasize the importance of shot movement and shot positioning (p < 0.001) as key predictors of shot success. In addition, the new model highlights the odds of achieving a successful shot are increased when the shot is either taken early in the shot clock (Odds Ratios of seconds remaining in the shot clock: 0-6 seconds: -0.19; 7-12 seconds: 0.23; 13-17 seconds: 0.32; 18-24 seconds: 0.50). The information gained from this model could be used by coaches and support staff to provide technical and tactical feedback to enhance performance. During training sessions coaches can attempt to ensure players are taking shots in a square to basket position (Odds Ratio: 2.06) or can use the information to recreate shooting situations in order to work on areas of weakness to increase the players odds of success over time.