It's been my observations that fired cases have the most uniform outside dimensions. They press hard against the chamber walls and bolt face at peak pressure then spring back a thousandth or so.
When they're resized, depending on the die used, how its set in the press, what case lube's used and how much case lube's put on it and where, how long the case spends all the way into the die (the press springs up, then slowly comes back down and can move the fired case shoulder back another thousandth or two without lowering the ram), the dimensions of a bunch of sized cases will have a greater dimensional spread than before they were sized. The biggest spread will be in case headspace; the dimension from the case head to the headspace reference diameter/point on their shoulder.
Wilson case gauges have SAAMI case headspace limits for the steps on their base for folks to see if their cases are within SAAMI specs. If your rifle's .223 Rem. chamber headspace is at the maximum SAAMI limit (1.4736"), fired cases from it will have headspace .001" to .002" less (approximately 1.4720").
Full length sizing those fired cases setting their shoulder back .001" to .002" is ideal for both accuracy and case life for bolt guns. Semiauto rifles may need a bit more; .002" to .003". If their shoulder are set back more than this, there'll be more case stretching when they're fired. The firing pin drives them hard into the chamber shoulder (sets the shoulder back a thousandth or two from the force they're pushed into the chamber shoulder just like in a full length sizing die), then the case body front half expands against the chamber while the back half gets pushed back until the case head stops against the bolt face as it expands against the chamber wall. This stretching and expanding also sucks brass back from the case neck and shoulder making the case length a thousandth or two shorter than before it was fired. But the case stretching and sizing tends to make the case thinner at its pressure ring about 2/10ths inch in front of the case head. And too much repeated case stretching at this point ends up cracking the brass and that's called head separation. So, minimal shoulder setback is the solution for safety as well as long case life.
Full length sizing such cases squeezes the brass diameters down lengthening the case back. Repeated firing and sizing them shortens their length then makes them longer; they grow in length a thousandth or so each cycle. They have to be trimmed back when they get too long so chambering will be easy as well as the case mouth has clearance to the chamber mouth and won't stop against it when fired.
Use your case headspace gauge (Hornady LNL, RCBS Precision Mic, or whatever) to see how much shoulder setback you get. Uniformly lubing cases helps keeps the spread to a minimum. Using a Redding competition shell holder of the right height lets you set the die in the press so the shell holder stops against it when the case is far enough into the die and its shoulder is set back the right amount. With standard shell holders, be sure the press ram's held at the top for the same amount of time for each case sized.
