I have just learned that the AI remembers contacts even if they go out of sensor range and continues to react to them. For example, a ship might still go to protect an "control"-marked objective even if the enemy was just visible for a second or so in the area.
The problem is the player will likely not continuously stare at the map and see every brief contact. That
1) is a tactical disadvantage, because he has no way to learn of a contact he did not see himself.
2) can be confusing if the AI reacts to something the player can't see, making him question the AI behavior.
A way to fix that would be a brief contact shadow, an afterimage of the contact. The way I imagine it, the shadow would be fixed at the edge of the sensor radius, moving along with it. It would disappear after the object reappears or after some time has passed where it did not. Maybe use the same memory time as the AI, if that works.
The shadow would have to be clearly marked a such to avoid confusion.
