Statistical Analysis And Prediction Of The Outflow About The Players Of ARPG

The rapid development of online games makes us see the bright future of animation industry and the game industry, and the outbreak of onlincgame has already become the biggest engine in the development of animation industry and the game industry. The behavioral datas about the players that are used in this article are the players’s level, the average tine, the dummy currency, cambat power and so on,and developing models which are base on these datas to predict the number of the loss of player, and to predict the probability of the final loss of player.This article use statistical principles and methods to do statistical analysis and prediction based on the behavioral datas of the Crazy Tribal players between 2013 to 2014. The statistical methods which are mainly used in this article are the main factor analysis, Jonckheere-Terpstra test, Krusleal-Wallis test, Linear-regression, nonlinear-regression and so on. According to the statistical results, I will provide advices for the game developer when they set design module, sent gifts to players, ask for charges and so on.