I don't think backlash is my issue but in answer to your first question I can say that there is no occurrence of the word 'backlash' in the StarGo User Guide and nothing in the application UI for configuring any such compensation. There is only one mention of backlash in the M-Zero User Guide but this is solely to promote Avalon's view that their Fast Reverse drives are backlash free.
On the Avalon Support web-page where PHD2 configuration is discussed they recommend not to configure backlash. I am compliant with this advice and have to say I am very content with my mount's guiding performance.
As you correctly say there is a 'Mount Settling Time' parameter in SGP Telescope TAB which I currently have set to a low value (20 sec) which is what Avalon recommended. I haven't noticed badly misshappened stars so think this isn't an issue. I understand the purpose of this parameter is to allow for vibrations that may be caused by elasticity in the mount drive coupled with the inertia of a heavy rig to dampen down. My OTA and camera are comparatively light so maybe even 20 seconds is overkill.
My current view is that my centering issue is caused by the displacement for the next corrective slew being determined without recognition of the latest mount positionning as communicated by the most recent SYNC. Any idea how long a SYNC action should take to become effective? Anyway I've been thinking I should perhaps try setting the SYNC behaviour parameter to 'None' or one of the other two available options, but am unclear what results to expect from such a change - haven't seen any mention in the SGP guides.
Unless there is something in the ASCOM interface standard regarding SYNCing then to my mind corrective action for this issue probably lies with Avalon, perhaps a state flag to indicate that a SYNC update is in progress, but I've yet to get a response from them on this issue.