Soccer tournament simulation and analysis for South Africa World Cup with Poisson model of goal probability

Referencing Dyte and Clarke's model, we propose a new formula to calculate the expected goal numbers. The goal probabilities are described by Poisson distribution. Based on the ranking data of men's teams in FIFA website, the parameters of Poisson model for each pair of teams can be determined. According to the full game program of 2010 World Cup, the full soccer games can be simulated. Based on millions of simulations running, we obtain the statistical results of champion, half final, quarter final and others for 32 attended teams. The fortunate degrees of all game groups for all teams can be evaluated from the rank position changes. These results may provide an analysis tool to organizations and fans of soccer games.