There are two questions in this thread.
One is about how to get the last tightening of the knobs such that it reflects the best alignment. That is, since the final torque on the set screws, etc. affects the position slightly, it is bound to be a bit off. This question has been discussed rather thoroughly in this thread. It is a point of "easing" into that final position and tightening. And, yes, you are right. There probably always will be some final effect!. But not enough to worry about. The differences in the torquing of the mount adjustments are probably dwarfed by the potential errors induced by flexing, flopping, refraction, tripod settling, weight shifts, and all that. And at any rate are less than will show up as rotation in the longest of exposures.
The second question, and it was actually more prominent in the original post, has not been addressed well.
It was the original posts "ponder" that:
>>>>>>Going back to your scope being at your index marks seems to me to have much more room to be "off" than the small torque adjustment at the end of the process.
Yeah, you can be a lot further off on the "index marks" than any torque on the mount adjusting bolts. Your perception of that is correct.
But the point is, that does not at all affect the alignment of your mount. A polar alignment affects the alignment of the MOUNT. It is not really about the pointing of the tube. Setting your tube to the index points just affects where the tube is pointed, not where the mount is aligned. To make sure you have a good alignment, you have to slew to a target, center it, and then synch the mount. And, considering the vagaries of the mounts, you probably need to do that to several stars so the mount software can calculate a "pointing model."
Even if you are fifteen degrees or more off in setting the tube back to the indexes, the mount still has the same polar alignment.