|（School of Physical Education，South China Normal University，Guangzhou 510006，China)
Abstract: The authors carried out mathematical modeling on the technical and tactical performance indexes and match results of Chinese Football Association Super League games in the game seasons of 2014, and analyzed and probed into match technical and tactical performance. The research samples consist of 240 matches and 480 groups of techni-cal and tactical statistic data, while the research variables include goal correlated variables, 11 offense correlated vari-ables, 5 defense correlated variables, and 1 game scenario variable (home or away). Firstly, the authors defined score balanced matches by means of cluster analysis method; secondly, the authors established a general linear model for every tactical performance index value and match result of every score balanced match, so as to define the linear rela-tionship between match technical and tactical performance index and match winning probability; lastly, the authors de-fined the significance of linear relationship between every technical and tactical performance index and match winning probability by using magnitude-based inference. The research results show the followings: 1) the matches with a score difference of 2 goals or less are score balanced matches; 2) in goal correlated variables, every increase of standard de-viation of 2 goal attempts, the team winning probability can increase by 16.3% (±90% confidence interval: ±14.6%), while every increase of standard deviation of 2 on-target goal attempts, the team winning probability increases by 33.8% (±16.2%); 3) in offense correlated variables, an increase of standard deviation of 2 passes, successful passes or through passes can bring an increase of 21.6% (±15.9%), 27.3% (±17.7%) and 16.9% (±22.9%) respectively to team winning possibility, while every increase of standard deviation of 2 fouled times, the team winning probability would decrease by 25.4% (±18.6%); 4) in defense correlated variables, every increase of standard deviation of 2 steals can in-crease the winning probability by 14% (±13.6%), while every increase of 1 red card would decrease the winning prob-ability by 30.9% (±26.7%); 5) in the score balanced matches in Chinese Football Association Super League games in the game seasons of 2014, the winning probability of home matches can be 9.5% (±15.4%) higher than that of away matches.. The research results indicate the followings: the general linear model can effectively define the causal rela-tionship between various match technical and tactical indexes and match winning or losing, thus can be used to deter-mine which match technical and tactical indexes are match winning indexes. Information provided by the established model can be applied to football practice such as game performance evaluation, opponent information detection, se-lecting and revising training or game preparation plans, etc.
Key words: competition and training；football technical and tactical performance；big data analysis；general linear model；magnitude-based inference；winning index