This study aimed to detect the factors affecting successful shots in basketball games by comparing the results of simple cross-tabulation analysis and multivariate regression analysis after recording 13 items presumed to affect successful shots in basketball games.
Subjects were nation-wide top-ranking university players. The outcomes of all shots and 13 shooting-condition items (“Difference of gains and losses,” “Remaining seconds of shot clock,” “Locations of attempted shots,” “Directions of attempted shots,” “Ways of shooting,” “Plays leading to shot attempts,” “Whether screen play or not,” “Actions before shots,” “Movement of the ball before a shot,” “Distance between a shooter and the defense,” “Handwork of the defender,” “Block shots,” and “Whether fouls were committed or not”) relating to the four factors of “Game conditions,” “Shot attempts,” “Tactics leading to a shot,” and “Defensive conditions” were recorded at 10 games after the quarterfinal games of the 66th All-Japan Collegiate Basketball Championship.
First, relationships were investigated using a Chi-square test after forming a cross-table between the outcomes of shots and various condition items. Second, logistic regression analysis was conducted using the outcomes of shots as a dependent variable and the 13 shooting-condition items as independent variables in order to investigate comprehensively their relationships. Finally, the two results were compared.
As a result, the cross-table analysis found significant relationships in 11 shooting-condition items. In logistic regression analysis, the best-fitting model based on “Locations of attempted shots” in “Shot attempts,” “Whether screen play or not” in “Tactics leading to a shot” and “Distance between a shooter and the defense” in “Defensive conditions” etc. was derived. Although the result of the multivariate analysis was different from that of the simple cross-tabulation analysis, it is considered that the items chosen as the best-fitting variables in a logistic regression analysis are essential because the multivariate analysis keeps the relationship of other variables constant.
View full abstract