In this part, we managed to disregard high-level tactics as much as possible. This was possible by using expert wisdom and supervised learning. In many cases, however, the weapon-selection behavior integrates with the tactics. Often, criteria used in choosing weapons depend on the tactical behavior and the current state. We'll cover high-level tactics in Part VII, and we'll discuss how to integrate the existing decision tree design then.
Parts III and IV focused on offensive behaviors, including shooting and weapon selection. Such capabilities often go hand in hand with defensive behaviors, such as dodging fire or picking up armor. The next part covers these capabilities and neutral behaviors, such as climbing ladders or using platforms. This will make our animats better-rounded individuals—and therefore more realistic to other players.