OK, I see that for 850m in model, you have a difference of 1mm (with anchor -95mm; without -94mm), not 7 mm as you said.

But as I said, is important to have a good procedure to calculate the displacement received in the end of pipeline (in the hot bend, if there is one\ in front of your battery limit if you haven't that loop/offset/leg). Anyway the actual figure that software returns depends on the soil modeled and this is not an exact science, isn't it? Again, important is to be sure that with your modeled soil your calculation returns a correct value of the displacement.