During the firing of a round, when the firing pin hits the primer, that force and that of the explosion of the primer pushes the case forward in the chamber until the case shoulder is stopped by the shoulder part of the chamber. At this point the primer is sticking out of the back of the case a little and the head is not touching the bolt face (but the primer is). As the pressure rises assuming a dry case and chamber, the sideways force in the case effectively locks the case in this forward position by means of the friction created between the case body and the chamber. The primer is still sticking out and the head is still not touching the bolt face. When the pressure gets high enough the case will stretch at the back and the head will come back to the bolt face, which pushes the primer back flush, in the case of a warm to hot load , and in the case of a medium to light load the head will sort of bounce off of the bolt face a little and the primer will end up very slightly above the head of the case. This and the possibility of primer cratering are why you must either remove the primer or reseat it well below the head before measuring a fired case. The measurement will be off by the amount that the primer may be above the head and by whatever slight amount of crater there may be. When we are dealing with thousandths of an inch these things matter.
With light loads thin brass and smaller shoulder angles a light load can actually shorten a case's "headspace" Those who shoot cast bullet loads in rimless cases need to be aware of this. As the shoulder is progressively moved back by sucessive firings, the neck will become longer faster than the overall case, which will cause it to eventually be crimped into the bullet during firing, by the angle at the front of the neck portion of the chamber. If the shooter is only looking at case length, he is likely not to catch this. The other sign that it is happening is that after each successive firing of a case at that pressure the primer will be sticking out of the case a little more than the time before. This is not conjecture. I have done the experiment....years ago.
Bottom line, in order to understand your measurements you need to understand the forces that have caused them. If we have clearance based on the longest measurement of the datum to head produced in that rifle barrel combination, then we do not need to increase it by setting our die to bump. Bump will vary with brass hardness as will case " headspace". Careful annealing will make all of this a lot more uniform. Even almost new brass can have a lot of difference, case to case in how it stretches and how it responds to sizing.
