目的:棒球比賽的統計數據相當多元且豐富,哪些指標可以用來預測球隊比賽的勝負是球團管理階層和球迷所關心的,因此本研究的具體目的在以球隊的攻守數據和競賽賽程因素來對個別球隊的比賽勝負進行預測,同時也探討這些影響因素在預測四隊勝負時的異同。方法:取材自中華職棒30年的例行賽,以防禦型指標、攻擊型指標、競賽制度和球隊戰績等4大類10項指標對個別球隊的比賽勝負進行預測,使用的統計分析方法是多元邏輯斯迴歸。結果:10項指標中分別有4、2、4項指標可有效預測Lamigo桃猿、中信兄弟和富邦悍將的勝負,但在預測統一7-ELEVEn獅的勝負時卻是失敗的。結論:可有效預測勝負的攻擊性指標和防禦率指標都出現在相對戰兩支球隊的比較基礎上,而非對戰所有隊伍的數據上。這些指標中有些是跨不同球隊的發現,也有些是某特定球隊所特有的。根據本研究之發現,也給未來研究提出數點建議。
Purpose: The statistical data of baseball games is quite diverse and rich. Which indicators can be used to predict the outcome of a team game is of concern to the team management and fans. Therefore, the specific purpose of this research was to use the team's offensive and defensive data and competition systems factors to predict the outcome of individual teams' matches, and discuss the similarities and differences of these factors in predicting the outcome of the four teams. Method: Data retrieved from the 30-year regular season of Chinese professional baseball, it used 10 indicators in four categories, including defensive indicators, offensive indicators, competition systems, and team performance, to predict the outcome of individual teams. The statistical analysis methods used were multiple logistic regression analysis. Results: Among the 10 indicators, 4, 2 and 4 respectively can effectively predict the outcome of Lamigo Monkeys, CTBC Brothers and Fubon Guardians, but they failed in predicting the outcome of the Unilion. Conclusion: The offensive indicators and defense indicators that can effectively predict the outcome of the game appear on the basis of comparing the two teams, rather than on the data of all the teams. Some of these indicators were found across different teams, and some were unique to a particular team. Based on the findings above, several suggestions were also made for future research.