Humans aren't machines. What is true from a mechanical standpoint may not be true from a biomechanical standpoint. We can easily constrain a robot to one stroke path, impossible to constrain a human to one stroke path without adding mechanical components. This is why some things are great in theory, not so great in practice.
For soft shots I tried to come to speed in the first half of my forward stroke then maintain a constant speed, zero acceleration. Multiple problems, not something my body was designed to do. My best speed and directional control was achieved when trying to slightly accelerate through the cue ball. Not something normally possible but in this case the attempt is what matters.
Afraid you lost me with the thought that coasting is the most efficient way to transfer energy. In layman's terms a moving object can be accelerating, maintaining a steady speed, or coasting. Anything coasting is slowing down from a greater speed so the energy used to achieve that greater speed is lost before hitting the cue ball. This would seem less efficient than maintaining a constant speed or accelerating until the moment of impact.
Maintaining a constant speed before impact is wasteful also. Reaching the desired speed at the
moment of impact is the most efficient use of energy. There are other considerations also. Compressed or compressing muscles provide more support and guidance than relaxed muscles. Another issue is that our arm muscles develop unevenly possibly negatively impacting the path of the cue when the arm muscles are in transition. Better for guidance to have all muscles in the arm contracting than some contracting and some relaxing.
My understanding of things anyway.
Hu