2) (slowly) setup the die until it just start make contact the brass... (bear in mind, at first contact, it may actually make the headspace measurement longer by a couple thousands)
Initially, I deliberately set my FL die too high so the shoulder doesn't get pushed back at all, only the body and neck gets squeezed. This moves the shoulder ahead. It has no where else to go.
This is how I get brass that won't fit in my rifle, to give me something to work with.
If I do that with maybe 10 fired cases, I'll usually get a couple that won't allow the bolt to close.
I measure the shoulder position on those cases for reference.
Then I slowly turn the die down and re-size the cases. I work in increments about the width of a Sharpie line marked on the 7/8" thread diameter with a corresponding mark on the lock nut.
I think it's Forster that makes a lock nut with etched increments for just that purpose. Hirsch Precision had them in stock a few months ago.
I measure the shoulder position of the case after each increment, and check it in the rifle. Keep track of the measurements, again for reference.
As the shoulder of the die comes down, the shoulder of the case will begin to follow it.
I may need to re-anneal the case after several increments, it will work harden to some degree.
Once I get to the point where the bolt will just close with very mild resistance, I consider that to be the absolute maximum length I can get away with. I record that measurement in a safe place (if there is such a thing) for future reference. It'll be the same for the life of the rifle, or at least that barrel and chamber.
That number has zero relevance to anyone else's rifle, because it was arrived at using my particular comparator bushing.
Anyone else's measurements will very probably vary from that dimension, every tool is slightly different, and so is the degree of "feel" while taking the measurement. Plus, every chamber isn't the same.
That number is purely a reference for my use, in my rifle, using my tools.
Your mileage will vary, without a doubt. There are too many variables involved for it not to.
The point is, when I measure a case that will just barely chamber and get X.###", I know that the case dimension I am trying to achieve is X.###" minus 0.002", maybe 0.003" to be on the safe side.
I set my die to consistently produce brass that size. I know that gives me between .002-.003" of clearance. Anything shorter is a sloppy fit, anything longer runs the risk of not chambering.
That's for my bolt gun. If it were a semi-auto or a lever, I'd want somewhat more clearance. Many reloading manuals have a recommendation about that.
That's my initial setup procedure, anyway. I'm sure others are different.