Part 1 - First is an Enhanced Arm Reticule:
Figure 1 is the current reticule.
Figure 2 shows the arm reticule grayed-out, which would apply when there are no weapons in the arms.
Figure 3 shows the left half of the reticule turned red, indicating that the left arm has been destroyed.
Figure 4 shows the right arm destroyed.
Figure 5 shows both arms destroyed.
I came up with this as I was working on other ideas. Combat in MWO can be very chaotic, particularly while under fire, pulling the player's attention in different directions at once. The one place that a pilot's eyes will fall most often is the targeting reticule. Thus, I thought it would be nice for the crosshair to provide more information at a glance. Of course, just about the only thing I could think of was the status of the 'Mech's arms, but that seems useful enough.
Part 2 - Next is a way to integrate a Cannon Arm (Such as the original Dragon's right arm) into the existing two-crosshair system:
Figure 1 shows how the reticule would behave if the 'Mech's right arm did not have a Lower Arm Actuator (Preventing horizontal traverse), and the left arm does have a Lower Arm Actuator. Note the transparent half circle 'ghost reticule' that stays with the torso reticule, showing that the right arm is aligned with the torso's point of aim. The arm reticule is grayed on the right half, showing that only the left arm is following the reticule.
Figure 2 demonstrates a possible way for the Cannon Arm to follow the arm reticule's vertical position while remaining aligned with the torso horizontally. Hopefully the 'ghost reticule' would not interfere with the range display in this case.
I don't think that this would be too difficult to implement, particularly with the recent addition of 'Lock Arms to Torso'. When the arm reticule moves left or right of the torso reticule, the Cannon Arm could act as if it was locked to the torso. If it would be preferable for the Cannon Arm to follow the arm reticule vertically as it does for 'Mechs that have both arms lacking a Lower Arm Actuator, then the 'ghost reticule' could move as shown in Figure 2.
While I was working on this idea, it occurred to me that this would also allow the destruction of an Arm Actuator to have an effect on aiming. Losing the Lower Arm Actuator would change the behavior of that arm to a Cannon Arm. Loss of the Upper Arm Actuator would lock the vertical movement of that arm, so the 'ghost reticule' could follow the horizontal movement of the arm reticule. Alternatively, loss of the Upper Arm Actuator could lock that arm to the torso as in Figure 1.
Part 3 - Last is what I'm calling the 'Convergence Reticule':
While I was considering changes to the reticule, I realized that a number of the problems with aiming could be reduced or eliminated by showing the status of weapon convergence. As often as First-Person Shooters are used as a negative comparison for MWO, I think we could take a cue from the way that such games inform the player of their accuracy.
When the reticule is moved, a pair of 'ghost reticules' would expand horizontally from the arm reticule based on the distance where the weapons will converge compared to the distance under the reticule. As convergence corrects to the new position of the reticule, the 'ghost reticules' would contract until they merge with the arm reticule. To show that convergence is not aligned with the target under the reticule, the arm reticule would turn gray.
Optionally, an additional layer of complexity could be incorporated. The position of the 'ghost reticules' could correspond to where the left and right side weapons would hit at the current distance of the arm reticule. A player with good timing could attack multiple targets or different locations on the same 'Mech intentionally.
With convergence set to instant with the recent patch, this would have to wait, but it could help.
(EDIT: Details on Convergence Function)
From "Developer Answers 5 - 'Mech Warfare", David Bradley - Game Designer - Overseeing 'Mech combat and BattleTech® Rules
Quote
The tutorial video referring to the arm crosshair and torso crosshair tracking as 'convergence' is incorrect.
Thoughts?
Edited by Renthrak, 12 April 2013 - 01:30 PM.