Discriminative power of valid game-related statistics in the National Basketball Association
Cardiff Metropolitan University
MetadataShow full item record
The aim of this study was to investigate the discriminatory power of valid gamerelated statistics, on the outcome of a basketball game. The sample included 246 games from the Men’s National Basketball Association; 2012 - 2013 Regular Season. The games were classified as balanced (equal or below eight points) or unbalanced (above eight points) based on the final score differential. Before analysis, reliability (95% limits of agreement) and validity (kappa) assessment procedures were completed on the raw data. Mann-Whitney U (post-hoc pairwise) tests with adjusted Bonferroni correction were used to compare 17 game-related statistics between the winning and losing teams, in each category. The statistical significance was set at P < 0.003. The analysis highlighted that field goals made, field goal percentage, three points made, three point percentage, defensive rebounds, total rebounds, assists and turn overs significantly differentiate between winning and losing teams, over all games (P < 0.001). In addition, unbalanced games had a total of nine performance indicators that were significantly different, with steals also becoming discriminative than in all games. Conversely, for balanced games, field goal percentage and field goals made were the only significant variables that distinguished between winning and losing teams (P < 0.001). These findings have numerous practical applications for an NBA team’s coach and recruitment scout. Specifically, the coach can utilise the findings to design conditioning programs, regulate training sessions and inform player or tactical selection.
DEGREE OF BACHELOR OF SCIENCE (HONOURS) SPORT AND PHYSICAL EDUCATION