New EGR, new software

When the valve is closing, it stops the rotor from turning at the exact point where the valve is closed. (Because of normal thread clearances, if not stopped right there, the rotor could keep turning a bit further, and screw itself down on the closed valve.) The metal nib that it stops against has already melted a short distance into your ski jump, so your rotor is going a tiny bit past 'closed' now when it closes.

Because the valve has no position sensor, when the ECM wants to open the valve to a certain position, it simply has to assume that 'closed' was exactly closed (where the stop is), and then it counts out the right number of stepper-motor steps in the opening direction from there.

If the rotor screwed itself down some number of steps past 'closed', that strategy will end up leaving the valve that number of steps less open than the ECM wants it.

Another thing that may be possible, once the rut goes all the way through, is for the rotor to screw itself down on the valve so tightly when closing that it won't have the torque to open again later. The gas valve in my furnace, which uses a similar design, did that last winter, leading to no heat.