Yes, with decomposition I meant a force causing the translation (your normal) and rotation (the perpendicular projection of your normal respect to the lever). By "radial" I mean the force vector applied to the centre of the object would give the direction of motion, which in your case it's the normal since you are dealing with a contact force. I tend to visualise things in a more abstract way (there are force vectors), and you in a causal way (the contact is a normal force). So all seems fine.   Yes, because translational and rotational impedance would be different. Imagine a ball moving slowly, but rotating really fast. If you push it off centre, doing so "in tandem" with its translation would benefit that more, while doing so "in tandem" with its rotation would benefit that more instead, because it's about applying the impulse as efficiently as possible (like pushing a swing with the correct timing), and there are two "independent" velocities in the system. Now, all that is just to justify that you can have independent scaling for translation and rotation. If choke points will be a mess or not, hard to tell without trying, that's why I said "hopefully", having in mind that in choke points relative speeds could become quite random, but hopefully low enough, although if objects accelerate instantly I have less hopes for things to work out (I hope ships need to accelerate though, otherwise my coming ramming ideas won't be as nice).

we are here forever....

It's much easier to use on Telegram, and it's very user-friendly.

We don't decompose the collision vector. We calculate the resulting impulse from the contact normal, and the spin actually comes from calculating the torque given the impact position to the center. Since the shape is (almost) a rectangle, then it's just about calculating the lever effect on a axis - the contact point projected on it.  Velocities are also calculated based on this closest point. I think to apply your idea it would be to add a scaling to the lever effect? But I'm not sure it's always desirable, for example on the choke point wouldn't that create more units facing the wrong direction of movement?