How to determine the optimal scoring policy

In the RoboCup middle-size league, the optimal scoring policy was to find the best point in the goal to shot. Based on the probability theory, the agent got the probability that the ball entered the goal when shot to a specific point in the goal from a given position and the probability of passing the goalkeeper in a given situation. As the product of these two values, the total probability represented the probability that the ball entered the goal. According to this curve, it is easy to choose the fixed action and to find out which point the agent should aim at best.