运动成绩预测的数学方法初探

A Preliminary Study of the Use of Mathematical Methods in Predicting Sport Performances

运动成绩的预测与体育发展战略及训练决策休戚相关。在经验判断和情报分析基础上发展起来的模型预测,其效果日益显著,故预测研究更为世界各体育强国所关注。本文仅就美国加州大学P.贝蒂芬尼教授对奥运会部分单项成绩预测的方法做一讨论,并探讨滑动平均、指数平滑、回归分析等七种确定型时间序列预测方法的实用性,对各种方法进行比较。对1952—1984年奥运会部分项目成绩进行拟合,从中选取最佳模型,再对1988年奥运会部分单项成绩进行试预测。

Predictions of athletic performances have much to do with strategies for sports development. Mathematical models in performance prediction based on past experience, judgement and analysis of information have produced increasingly remarkable results and consequently the study of prediction methods has receiwed ever greater attention in the world's leading sports nations. This paper discusses the methods used by Dr. P. Betipheny of the University of California for predicting the performances in certain Olympic events. The practical values of seven prediction methods are examined and compared with each other. Performances in some events at the Olympic Games from 1952 to 1984 are analysed and the best model is derived for predicting performances in these events at the 1988 Olympics on a tried basis.