Do better things come in smaller packages? Reducing game duration slows game pace and alters statistics associated with winning in basketball

Basketball games are currently played across 48 and 40 minutes; however, little is known about the influence of game duration on team statistics. Therefore, we examined (i) differences in game-related statistics between 48- and 40-min games; (ii) statistical measures discriminating winning and losing teams relative to game duration; and (iii) the ability of statistical models to predict team success relative to game duration. Overall, 1,084 Australian professional men's basketball games were analysed. Ball possessions (·min^-1) had the largest difference (r_pb = -0.30; P < 0.001), between 48- (1.92 ± 0.12) and 40-min games (1.79 ± 0.26), while offensive efficiency and defensive statistics were significantly (P < 0.001) greater during 40-min games. Most statistical measures significantly discriminated winning from losing teams, with larger disparities evident during 48-min games. Offensive rating (32.3-43.0% vs. 22.4-29.8%, P < 0.001) and Four Factors (32.3-43.0% vs. 22.4-29.8%, P < 0.001) each explained a greater proportion of variance in game outcome during longer games. A reduced duration slowed game pace and increased offensive efficiency, possibly due to greater maintenance of defensive pressure slowing opposition transitions and promoting controlled offensive structures. The ability of game-related statistics to discriminate winning teams and predict game outcome diminished with reduced game duration.