It depends on where the direction changes happen - for the output to shift, you have to change direction in X at the same time as continuing a move in Y, or vice versa, and if you perform the opposite moves, it’ll “undo” the lost step.
I haven’t seen an offset due to the ‘rising edge’ setting on vectors before - usually it’s so small that it’s nearly imperceptible, but this test has enough direction changes in it that it’s obviously showing it as well.
The usual giveaway for the “rising edge” issue is when an engraving leans to one side, like this:
Thanks for jumping in here Brian - I’ll have to remember this one myself. 