The memory model that was developed was compared with a very simple memory model, which ignores the the inaccessible aspects of the state. This memory stores the positions of all objects that are visible. In order for the model to be useful at all, the client must at least be able to turn away from an object (e.g. the ball) while looking for another object (e.g. the goal). This was done by adding the capability for the model to remember the position of objects when the client turns, by simply updating their position by the exact amount of the turn. If the client dashes then the positions of all the objects are forgotten and only the position of the newly visible objects are known. This is the simplest model for memory that provides enough functionality to be usable by the client.