I am very sorry - but I must disgree with Chip once again about the adjustment of your number of steps per unit. For some reason he seems to think that you can fiddle about with that until you get things right. This is not correct.
Your Maths for working out the number of steps per unit are correct, and 400 is the right answer. Why on earth anyone should make a rotary table that required 421.05261 turns per rev is beyond me - or are standards so sloppy that is how it turns out. It must be the heat.
The correct way to test this is as follows. Jog the table clockwise. Mark a position on the table - where the rotating part meets the body. A pencil line will do (across the join). Instruct the computer (G0 or G1) to turn the table 360 degrees - again clockwise . The pencil mark should then line up again.
If the pencil marks do not line up then I will agree that the screw thread is not accurate - but I personally would send it back if it is only accurate to 5%
To check backlash - use the scale on the table. Move the table clockwise to a convenient position e.g. 0 degrees or similar. Instruct the computer to move the table again clockwise by 10 degrees. If your steps per unit are accurate and your screw thread is accurate and the scale is accurate your scale should now read 10 degrees. Instruct the computer to move it back to 0 and see how the scale reads. It should be zero, but if you have backlash, it will not be. This is the amount of backlash and should be entered in the backlash table.
Make sure A is angular is ticked on the general config page otherwise if you change from inches to mm and vica versa, the A axis will change too - you don't want it to.
As far as speed goes - I am not sure what you are saying - you seem to have the speed set at 4000, that is a revolution of the table in 5 seconds (at rapid speed) - I am not sure what you are set up in - mm or inches - looking at the other speeds, I would think mm or you are going into orbit.
When you put in a command g1, use the s command to dictate the speed. I do not know what rate is set for the A axis to revert to. Is the jog any different from a g1