The point that the cable is clamped affects the distance and angle in relation to the pivot points on the derailleur. Although quite small visually, it has some effect on the actuation ratio of the derailleur. The resulting change could mean that the derailleur no longer has the 1.7:1 average ratio. It could now be + / - a few(?) % from the intended 1.7:1.

It would be interesting to see if the total movement of the derailleur, from low to high shift position, is substantially different depending on the attachment point.