Hi Ichan,
When I commented this morning I [mis]understood, and thought you meant the components gradually shifted/drifted to the left as they got further away from your 0.0 reference.
Looking at your pic now on a large screen, I see that ALL of the components appear to be displaced by the SAME amount over to the left of where you expected them to be.
How did you load the PnP reference data into the N4?
Was is a manual input, or were the co-ordinates imported from a PnP file?
Seems simple to conclude that what you considered to be 0.0, is in fact not.
This could also be the result of using a 'reference' footprint to set up the panel on the machine where the 'anchor point' is not in the middle of the component, and yet the machine assumes it IS in the middle. Sounds like something fairly simple whatever it is.
Unsatisfactory result today, the placement shifted to the left (Y- direction on the machine), and the bad board detection feature seems not working.
Will try again tomorrow, will add a dummy part as "first chip" for a more accurate reference and will do calibrating the nozzles first.
-ichan