查詢結果分析
來源資料
頁籤選單縮合
題 名 | 世界棒球經典大賽日本隊投手對中華隊配球模式之分析及預測=The Analysis and Prediction of Pitcher's Pitching Strategy during the Game of Japan Team against Chinese Taipei in the First World Baseball Classic |
---|---|
作 者 | 王金成; 蘇茂菘; 陳良乾; 吳季龍; | 書刊名 | 長榮運動休閒學刊 |
卷 期 | 1 2007.06[民96.06] |
頁 次 | 頁24-33 |
分類號 | 528.955 |
關鍵詞 | 棒球投手; 配球模式; 類神經網路預測; Baseball; Pitching strategy; Neural network; |
語 文 | 中文(Chinese) |
中文摘要 | 本研究目的:欲探討世界棒球經典大賽日本隊投手對中華隊的配球模式分析與預測,方法:使用錄影帶系統觀察法,將紀錄所得之資料以Office2003 Excel作統計百分比分析後再由類神經網路進行投手配球之倒傳遞類神經網路模式預測。其研究結果如下:(一)不同球種、投球數比率分析發現,累計七局比賽114球數中,日本隊投手以直球(61.4%)為多數球種,並以滑球(23.6%)之球數做為主要配球。(二)在方向配球模式比率分析中發現,交叉配球模式、同方向配球模式個出現有14打席(45%),但同方向配球模式的上壘率(41.85)較高。(三)不同球數配球模式比率分析發現,第一球出現率最多的球種為直球(67.74%);兩好球之後的直球出現率高達(63.63%);兩好球一壞球情況下變化球出現率(87.5%)高於直球。(四)投手配球之倒傳遞類神經網路模式預測,球種預測率為(70%);位置預測率(71%);方向預測率為(84%);好壞球預測率為(82%)。結論:(一)訓練投手時先除了開發投球速度外,最主要的目地是訓練各球種(直球、曲球、滑球...等)控制能力,以讓打擊者較無法進行猜測全力揮擊,降低失分危機。(二)應培養投手在練習或比賽當中方向配球方式,以對角線交叉配球模式來進行配球,能使擊球員的上壘率較低。(三)投手對打擊者是鬥智、鬥力的開端,在訓練或比賽當中,應培養打擊者在不同球數下對投手配球慣性進行訓練,以提昇確實擊球的機率。(四)在投手配球之倒傳遞類神經網路模式預測,預測變項高達(71%)以上。本研究為研究者的初探,未來研究將進一步針對國內、外單一投手或捕手與打擊者對戰資料進行長期追蹤分析,建立大量資料使預測模式能使研究更為精準,並應用於實際的競賽。 |
英文摘要 | Purpose: The purpose of this study was to analyze and predict the pitcher's pitching strategy during the game of Japan team against Chinese Taipei in the first World Baseball Classic. Method: The approach of video tape observation was used to collect the necessary data. A computer program written by C language was utilized to run the results for Back-propagation neural network predict model. Results: 1. The Japanese pitcher in the total of 114 pitching balls pitched straight ball mostly (61.4%), and the slide ball (23.6%) was the second; 2. The cross direction pitching and the identical direction pitching showed the same percentage (45%), but the identical direction pitching had the higher possibility to touch the base; 3. The straight ball was the highest percentage used by Japanese pitchers in the first pitch, and the variant ball (87.5%) occurred more than the straight ball after the condition of two-strike and one-ball; 4. The Back-propagation neural network model for pitcher's pitching prediction showed rather good percentage (70%) to predict Japanese pitcher's pitching. Conclusion: The Japanese baseball pitchers had their unique pitching strategy, and the Back-propagation neural network prediction model obtained about 70% predict rate for predicting Japanese baseball pitcher's pitching. |
本系統中英文摘要資訊取自各篇刊載內容。