Rufous,
You didn't say which neck sizer, the Competition or the Type 'S' die. I've used the Competition dies in .223 and .220 Swift, and had no problems, since I normally neck turned the cases being used. I'd end up w/ loaded rounds w/ about 1-2 thou runout, ususally less. On the other hand, when I picked up my 700VS in .308 Win, I got a Type 'S' die, since in all honesty, I for one don't ever use the micrometer set on the Competition die once it's been set up initially. I was getting up to 10 thou run out on loaded rounds, and eventually backtracked the problem to the neck sizer giving me at least 4-5 thou at the neck, which for a variety of reasons ended up magnified on the loaded round runout. I had the standard sizing button, but even w/o it, I still got what I considered too much runout. Redding gave me some song and dance about sizing the brass down too much because of a sloppy factory chamber causing the runout. Yada yada yada.
My solution? It sounds hokey (and feels cheap as hell), but the Lee Collet die saved the day for me. I inadvertently ordered the whole 3 die set, but just the collet die costs something like $16 or so. I figured it was cheap enough to take a chance on.
My inital loads, w/ the Redding dies, were shooting about 3/4-1"@100yds, 1-1 1/4"@200yds, and 1 1/4-1 1/2"@300yds w/ the 175gr SMKs over 45.0 gr varget. Basically the bullet needed some time to stabilize itself and overcome the initial yaw from going into the rifling crooked due to the runout, I believe, as the loaded rounds had like I said, up to 10 thou runout. My followup rounds loaded w/ the Lee Collet die, have about 2-3, sometimes 4 thou TIR, and routinely shoot under 1/2"@100yds, w/ one 3-shot group punching a nice .40-.45 caliber hole
A few things to note: I haven't ever used the O-ring btwn the die and the press trick, more out of laziness in finding a suitable O-ring than anything else. The Lee Collet die does come w/ one installed, IIRC. I do however 'float' the shell holder, by removing the metal clip, and (again, since I am too lazy to hunt down an O-ring) wrapping a rubber band around it a few times to retain it, so it isn't being pushed one way or the other by the metal clip ring. Also of note: I initially didn't notice much reduction in runout in the cases necks themselves using the Lee Collet dies. I did however, notice a substantial improvement upon measuring TIR of a loaded round. My theory about this is as follows: The bushing dies work fine... if you neck turn. When I neck turn, I expand the necks first w/ a mandrel, to 'iron' any deformities to the outside surface of the neck and then use the cutter to reduce/eliminate those irregularities. Subsequently when the neck bushing dies squeeze the neck down, effectively ironing the neck in reverse, there isn't much in the way of irregularities or variations to transfer to the inside neck, and the bullet/neck junction, so TIR is low. If you don't neck turn, the bushing die irons any defects to the inside of the neck, where they are magnified in TIR of the loaded round. The Collet die forms the neck around a solid mandrel, effectively ironing the defects to the outside of the neck. Hence, you don't see much improvement in measuring the outside of the neck for runout, but when the bullet is seated, and loaded round TIR is measured, a noticeable benefit is realized.
These are simply my observations. I am by no means an expert, and your mileage may vary. If you try the above, and your stuff breaks, you get to keep both parts
HTH,
Monte