| Summary: | 碩士 === 國立高雄第一科技大學 === 風險管理與保險系碩士班 === 105 === This research probes into the relationship between team winning percentage and offensive index of batting order of the Chinese Professional Baseball League. The research methods including correlation analysis, analysis of variance, factor analysis, cluster analysis and linear discriminant analysis are used to study the data derived from four teams in Chinese Professional Baseball League - the Chinatrust Brothers, Uni-President 7-Eleven Lions, Lamigo Monkeys and EDA Rhinos. The data originating from 240 regular season games in 2016, such as at bats, hits, runs batted in, home runs, strike outs, stolen bases as well as batting average, slugging percentage, on base percentage, stolen bases percentage, are taken into consideration so as to discuss the factors impacting on victory and to investigate the difference in performance between batters belonging to different teams, the results could be applied to the adjustments made by teams.
The results of factor analysis indicate that the abilities of hits and powerful hitting, getting on base and on base contribution are the main indice of offense. The post-hoc tests of analysis of variance are reviewed subsequently, which reveal that teams differ significantly among batting order, players in the Chinatrust Brothers present the best performance in terms of hits and getting on base, whereas players in the Lamigo Monkeys present the best performance in terms of on base contribution. In addition, there are no significant differences between four teams in respect of the abilities of powerful hitting and basic hitting. The cluster analysis is used to analyze the performance of hitters, hitters are divided into three classifications - powerful, quick and balanced. Powerful hitters are often in the center of batting order, quick hitters are usually the first, second and third batter. The linear discriminant analysis is used to assay the correlation between performance of batters and victory, the correct classificaion table shows that the total correct classification rate is 76.6%.
|
|---|
