The assumption behind the movement of the goaltender is that the worst thing that could happen to the goaltender is to lose sight of the ball. The sooner the goaltender sees a shot coming, the greater chance it has of preventing a goal. Therefore, the goaltender generally uses the widest view mode and uses backwards dashing when appropriate to keep the ball in view to position itself in situations that are not time-critical.
Every cycle that the ball is in the defensive zone, the goaltender looks to see if the ball is in the midst of a shot. It does this by extending the ray of the ball's position and velocity and intersecting that with the baseline of the field. If the intersection point is in the goaltender box and the ball has sufficient velocity to get there, the ball is considered to be a shot (though special care is used if an opponent can kick the ball this cycle). Using the passive interception if possible (see Section 4.4), the goaltender tries to get in the path of the ball and then run at the ball to grab it. This way, if the goaltender misses a catch or kick, the ball may still collide with the goaltender and thus be stopped.
When there is no shot coming the goaltender positions itself in anticipation of a future shot. Based on the angle of the ball relative to the goal, the goaltender picks a spot in the goal to guard; call this the ``guard point.'' The further the ball is to the side of the field, the further the goaltender guards to that side. Then, a rectangle is computed that shrinks as the ball gets closer (though it never shrinks smaller than the goaltender box). The line from the guard point to the ball's current position is intersected with the rectangle, and that is the desired position of the goaltender.