Publication:

日本職棒先發投手每場比賽投球數對後續表現之影響

投球是一個容易受傷的動作，過多的投球負荷，短時間內可能造成疲勞，導致後續表現的下滑，長期的累積，則可能造成運動傷害，但如何評估投球負荷，仍然沒有明確的科學證據，而且僅有的少數研究，大多均以美國職棒為研究對象。日本職棒先發投手輪值比起美國較不固定，其使用模式不一樣，本研究以日本職棒投手為研究對象，探討在不同投球數之後，後續三場比賽的表現。研究範圍為2006-2009年間，單年度先發出賽場次12場以上，且後援場次6場以下之投手，刪除篩選範圍內選手休息天數14天以上的場次，以及該場次所影響的統計數據比率場次。統計數據比率則利用[(後3場成績總和 - 前3場成績總和) / 前3場成績總和] x 100%，以卡方檢定與交差表分析，針對不同工作量及不同投球類型的先發投手，其前3場比賽的PAP^3累積量(PAP^3，[投球數-100]^3，若投球數<100，則PAP^3=0)與ERA比率、WHIP比率、FIP比率、DERA比率、RA比率、H/PA比率、BB/PA比率、SO/PA比率與HR/PA比率的影響。

結果顯示：Workload 1 （每場先發投球中位數<94）的選手，不同PAP^3累積量各項比率均無顯著差異，可能因休息天數過多，或是兼任中繼與後援角色，無法看出低投球負荷量的投手，用球數與後續表現的關係；而Workload 2 （每場先發投球中位數94-100）投手投球臨界點約在100球左右，超過臨界點，ERA比率、WHIP比率、FIP比率、DERA比率、RA比率和H/PA比率即產生下滑的現象；Workload 3（每場先發投球中位數101-109）投手雖可承擔的投球數，大約在110球，但超過臨界點，ERA、FIP、DERA、RA、H/PA、SO/PA與HR/PA表現較差的比率會明顯增加；Workload 4 投手（每場先發投球中位數>109）在不同PAP^3累積量下，後續的各項統計數據表現大多數無顯著差異，即使前3場PAP^3累積量調整至投球中位數為125球時，表現較差的場次亦無顯著提升，因此日本職棒Workload 4的投手，似乎有能力負荷高投球量，而不至於影響後續比賽的表現。在投球類型部分，Finesse型（(BB+SO)/IP<0.93）和 Average Finesse/Power型（1.13< (BB+SO)/IP >0.93）投手，在用球數達100-109球後，被擊出安打的比率開始提升，顯示投手的能力下滑，而Power型（(BB+SO)/IP >1.13）投手在球數高過110球時，三振表現較差的比率大於表現較佳比率。由此看出，日本職棒投手可負荷的投球量很極端，未來可再增加研究範圍，更進一步探討投球量極端的投手群的投球上限。

Pitching is an action that has high risk for injury. In the short term, high pitching load may cause fatigue, leading to the decline in the subsequent performance. In the long term, it may cause sport injuries. So far there is no clear scientific evidence to to evaluate the pitching load. The few available focus on Major League The rotation for starting pitchers in Baseball. Nippon Professional Baseball is relatively unstable compared to Major League Baseball. The aim of this study was to investigate the effect of pitch count on the performance in the subsequent 3 games in the starting pitchers in Nippon Professional Baseball. The sample was collected based on more then 12 games as a starter and less than 6 games as a reliever in a season from 2006 to 2009. The games with more than 14 days between the previous start were excluded. The ratio of each pitching statistics was calculated as [(sum of previous 3 games – sum of subsequent 3 games)/ sum of previous 3 games]. Then Chi-square test and cross tabs were adopted to analyze the effect of workload in the previous 3 games on with the ratio of earned run average (ERA) rate, walk+hit/innings pitched (WHIP), fielding-independent pitching (FIP), defensive-independet ERA (DERA), runs allowed average (RA) rate, hit/plate appearance (H/PA), base on balls/plate appearance (BB/PA), strike out/plate appearance (SO/PA), and home run/plate appearance (HR/PA). The pitching load was calculated as Pitcher’s Abuse Point (PAP^3，[pitch count-100]^3，if pitch count <100，then PAP^3=0)

The results showed that pitching load had no effect on any statistics in Workload 1 (median pitch per game <94 ). These group of pitchers may rest for a long time or also be used as relievers. The effect of pitch count and the subsequent performance was inconclusive in these pitchers.. The pitchers in Workload 2 (median pitch per game 94-100 ) showed a threshold of pitch count at about 100. If the limit was exceeded, the ratios of ERA、WHIP、FIP、DERA、RA and H/PA were decreased in the subsequent games. The pitchers in Workload 3 (median pitch per game 101-109 ) showed a threshold of pitch count at about 110. If the limit was exceeded, the ratios of ERA, FIP, DERA, RA, H/PA, SO/PA, HR/PA were decreased in the subsequent games. There was no significant difference in most of the statistics in pitchers of Workload 4 ( median pitch per game>109 ). The results were the same even if the median pitch per game was changed to 125. Consequently, Workload 4 pitcher in Nippon Professional Baseball may be able to sustain high pitch count without the effect on the subsequent performance. As far as pitching types, when Finesse ((BB+SO)/IP<0.93) and Average Finesse/Power （1.13< (BB+SO)/IP >0.93 pitchers’ pitch count were up to 100 to 109, the ratio of H/PA began to rise. It suggested that the pitcher’s ability declined. When Power （(BB+SO)/IP >1.13）pitchers’ pitch count were higher than 110, the performance of SO/PA was declined. The results of this study suggested that the threshold of pitching load has large individual variations in Nippon Professional Baseball. The future studies can include more players and/or longer durations to further explore the extreme pitchers’ pitch count limit.