Optimal trajectories is all about individual aircraft and we're dealing with a whole system - optimising that is a different problem to optimising an individual aircraft. It's nothing about trusting pilots. It's about limited ability to lose time with a tailwind vs a headwind. And which phase of flight pilots choose to lose the time - some reduce a lot in the cruise, others do it all on descent. The convergence is some distance from the feeder fix so you can have considerable difference in ground speed - e.g. 480kts ground speed from the north vs 380kts from the SW. The spacing will be good eventually, but until the ground speeds are comparable separation can be an issue. I've had extreme cases where ground speeds have been 200kts different due to wind direction and a huge headwind increase over a couple of thousand feet. Front aircraft 5 miles ahead at the turn opened up to over 10 before it has slowed. Maestro does not provide separation.