Introduction: The Chinese Professional Baseball League (CPBL) has been in operation for 28 years and, among all factors of player performance, including stealing bases, homeruns, strike outs, we want to find out what are the most important factors affecting fan attendance. Method: In this research, we use big data, auto model selection method to look for significant factors from all performance associated numbers, including winning percentage, hits, home runs, walks, strikeouts, and 23 other team statistics. We use a crawler program to collect data from the Chinese Professional Baseball League (CPBL) website and exam the teams' performance in more than six thousand games versus the fans' attendance of the games from 1990 to 2016, 27 years in total. We use auto model selection to find the best model and coefficients, and the most relevant factors to the fan attendance. Result: We found that errors, home runs, K/AB, steals, and pitchers' HR/AB are the most important factors that affect the attendance of CPBL. The R^2 of the selected model is 0.76, which is high correlated. All the p values in the model are less than .05. Conclusion: While errors is an indication to the fans that a player is not focused and thus drive the fans away, home runs, stealing bases, strike outs and the pitchers ability to prevent homeruns are all important factors to attract fans. The clubs can have these ideas in mind when they draft and trade players. And coaches can adjust the team's style of playing. Fans rather see the players they support play actively, stealing some bases, than wait for a walk passively.