A mandrel can be used as a final NSing action. This pushes thickness variance outward, away from seating bearing, to allow low loaded runout.
The expansion also biases spring back continuance to an inward direction, so that tension does not relieve/reduce over time.
But tension itself, while sizing is shorter than seated bearing, is adjusted by sizing length.
There are three camps here where approaching bullet grip is completely different.
- FL sizing of necks
- Partial length sizing of necks
- Crimping
All three work provided a load is developed with it. If you want to include neck tension adjustments in development, then I suggest this as a last process. A final tweaking of load. Sometimes it makes logical sense to do it, sometimes it doesn't.
FL neck sizing brings interference as sized into tension, where beyond seated bullet bearing depth. This definitely causes the highest neck tension, but at a cost of highest tension variances. You bring donut area and shoulder angle influence, and binding at the bullet's base-bearing junction. It's not something I would ever do. If you suspect, through testing, that FL NSing will help, then you're using the wrong powder (too slow).
Partial neck sizing typically allows both interference and length of sizing adjustment (if bushing sizing). This offers high potentials with tension adjustment. You can reasonably downsize, and mandrel expand a reasonable amount.
Reasonable = less variance.
So you can start with sizing 1/3 neck length, and later try 1/4, or 1/2, or 3/4.
True crimping (mouth swaged into a bullet groove) is so dominant to bullet release that interference fitting with the rest of the neck matters less. While crimping is bullet grip, it is not tension. That can be a great advantage when you find a load that likes it, powder tune-wise.
A mandrel could be used with crimping done as a specific operation after mandrel use. Critical to crimp release would be neck hardness.