chipmaker wrote on May 7^th, 2015 at 11:38am:

This is a possibility but not the only possibility.
Consider that at the moment of initial firing the case walls balloon out and grip the chamber preventing the case from sliding backwards to rest fully against the breach block. In fact the action of swelling the case walls will actually contract the length a small amount. Thereby giving the room to squirt the primer or separate the case

I have an early non-import Anschutz .22 Hornet..that thing used to eat brass..and showed the same primer set-back..yet headspace checked out perfect.  Finally determined the made-for rimfire action with it's bolt root locking into action was springy.  My Shillen barreled Sharps Borchardt on the other hand has been using the same RWS Hornet brass for 10+ years..primers stay flush.  Though I would have thought a Martini would have a stiffer lock-up?

In order to have a primer set back you have to have a gap between face of the bolt and the base of the brass. Angled bolt, rim cut to deep, thin rims, breech plug not flush, etc. something is allowing a gap. Is your head space gage using rim thickness variation or is it off of the shoulder ( not correct in my book). Can you measure the chamber cut for the rim

Thanks for all the suggestions regarding this problem.
The simple solution to testing whether a reloaded fired case would solve the problem would be to pull a couple of bullets from unfired cases and reuse the powder and bullet in the fired cases. Unfortunately, despite a determined effort, I can't get the bullet out of the case with my inertial bullet puller. So it will be a little while before I can locate a few bullets. I don't know much about reloading, so if anyone has any hints about removing bullets from unfired cases, I'd be interested. 
38 Cal
The donor action for the side lever experiment was an old Egyptian GP shotgun action complete with three prong firing pin and oversized barrel thread. The barreling and eventual making of the rifle was really an after thought, using parts left over from other projects. The initial attempt to make a side lever Martini, using the same type action, was unsuccessful and I converted that action back to an under lever action. The first attempt used a 32-20 barrel and showed no evidence of primer set back. The problem was in the side lever overcoming the heavy mainspring pressure. The solution to the cocking problem required the making of a new mainspring with slightly less pressure.
The current plan is to try using the fired cases, once I've located a few bullets. If this doesn't work, I'll re-barrel the rifle using a low pressure round and recheck firing pin protrusion/retraction, breach block face squareness, and side lever pivot pin movement at that time.
Otto

Is not the normal sequence thus:

firing pin pushed primer/ctg forward

primer ignites, pushing primer back

powder burns at some pressure, pushing ctg back or not depending on pressure developed.

I was told by some fellas back when I was having Hornet issues that re clambering to 22 Hornet Ackley Improved would eliminate the issue. And you could still fire factory ammo. 

I didn't try it because the solution  Marlinguy is suggesting worked fine for me.