Paw'd got me immediately hooked, and I feel that the game is an incredible experience as a whole overall. However, there has been one subject of slight to extreme discontent that has recurred time and time again for me, and likely for many other players: the auto-target function of the "e" key. While extremely useful in cases during which a target must quickly be selected because using the cursor to click is too time consuming, at times, due to a larger variety of targets to select from, the key deviates from its usual consistency of selecting the correct target. This is due mainly to the presence and positioning of potential targets that in reality are not of first and utmost priority to attack.
Personal examples are cases during which Tormented Argans were attacking me in the room directly east of Argan Village. In this room, there is cotton that periodically respawns, a commonplace sight after a few times passing through the room. The cotton can be interacted with to produce one or two cottons, but otherwise is fairly superfluous and should not be considered by the "e" key as a target of priority. This is especially true during instances during which there is a mob in proximity with the cotton, such as an Argan, or Venomfang. Nevertheless, the auto-target will at times target the cotton even when I am explicitly the object of the Argan's attention. I have come to realize that this and other instances during which auto-target behaves in such a way are due to its fundamental algorithm for what should be prioritized and what should not be.
Auto-target nigh always prioritizes based on distance, rather than combat behavior. Should a hostile monster be in the vicinity of a non-hostile monster or an NPC, the closest entity, monster or NPC, hostile or neutral alike will be targeted first. Such an algorithm is counter productive and anti-intuitive to what the "e" key's function is really supposed to be. I propose an alternative algorithm for "e" in which priority is assigned based on distance and combat behavior, with the latter being considered first, as opposed to distance alone. Potential targets could thus be divided into three separate categories: Attacking, Hostile, and Not Hostile or Attacking. An Attacking monster is a monster that is focusing aggression against you or an ally. Such monsters should be prioritized first, nearest to farthest. Hostile monsters are greater threats than non-hostile ones, so they should be the second level of priority, nearest to farthest. The monsters and NPCs that are not hostile or attacking are therefore the least threatening and can be prioritized last, nearest to farthest.
This revised algorithm would much better suit the need for an instant target because it values potential targets in a more accurate, combat based way. Hostile monsters that are not attacking are typically less a threat than a monster already attacking, hostile or not. Hostile monsters that are not attacking are more threatening than non-hostile monsters that are not attacking. In each case, nearest to farthest could be applied because the player would be able to evaluate exactly which target they want to be targeted and their relative position to them. This would be a much more efficient way for auto-target to operate as a whole, minimizing cases during which simply stated, the wrong targets get selected in situations that call for immediate action and thus exact selection right away.
As a final thought, at times during which there are many targets and one presses "e" too quickly, perhaps SHIFT+e could rotate to the previous target in case the desired target is skipped over?
