I understand the process you illustrate, but having no actual experience, I can't think of any reason there would be variance due to principle or an "ideal situation" and any problems that occur seem, to me, that they would be coming from natural deviance in machine operation and the general chaos of happenstance.
How good are your operators? That makes all the difference in the world that no engineer or designer can plan for.
Ummmmm.... operators.... ( don't want to get fired here )... Good!!
I just know what people that clam to know more tell me. I can see why certain circumstances would affect the outcome of the pattern; mainly speed comes to mind. If they are moving at the same speed, then I don't see why there should be distortion ( in a perfect environment, when no metals expand / contract when heated / cooled ), but those who are in the know, say there is. I have no data to defunct their statement. This is a big learning idea for me, so anything someone can find, stating one way or the other, I'm willing to read.
My thought was that maybe your tolerance is too tight for the equipment.
What is the tolerance?
Not sure, but I don't think they are too tight right now. What I've heard is that, say for example the circle, will be out of spec by as much as half the circle at one point, and then at others, totally out of the rectangle pattern, until it comes all the way around, and will be back in spec for a little while.