Do it right in the middle only talking about .00005 inch(+- little) , ball screw will have 10 times more error than that. That's if by the looks of it that's a .003 error rolled screw middle will come out to .000025 error one way or the other.
remember to add in a error of one graduation per revolution on the indicator.
How far off are you? If we knew your counts to try for an inch, and the actual distance travelled, then we could help with where the error lays.
You say 500 lines per revolution, but 512 would be the next normal value. That would give you 28912.94 counts per inch. So the error introduced by your wrong counts per inch would make 1" actually 0.97656" ... is this what you're seeing?
2. I have not tuned the geckos yet. I was waiting until I had everything put together, so that it was not far off the final weight/power requirements. because of the lack of tuning, the geckos are not "snapping" to the final position like the should. (I think). Over the weekend, I'll tune them and see how it works.
3. I think I'm expecting too much. At least for now.