Re: Why does annealing work?
The primary issue is that brass is a work hardening material while iron alloys are generally precipitation hardening.
The internal crystalline structure of an iron alloy will undergo changes as it's constituent parts heat up and subsequently cool. The temperature that the material is heated to, the duration, the cooling rate, and the normalizing temp and time all make a difference what happens to that iron alloy. There are other things as well, but as you can see the point is that it has a lot going on.
Brass is work hardening. Work hardening happens with repeated plastic deformation, where "plastic" deformation is permanent changes in the shape/geometry through applied stresses.
As you fire/size/load/fire/size/load/fire/size/load the brass cases they work harden. This causes it to have stronger material properties before failure and subsequently the amount of force required to yield the material (in our case, bullet release) is changed. The way to make it repeatable for us (because it varies with each piece of brass) is to anneal it back to the dead soft state each time.
Tom Sarver anneals his 300 Hulk necks every time he fires them when he's shooting for 1k BR records. That's the way to make sure it's repeatable each and every time.
For shooting steel plates and the level of repeatability that most of us are striving for, annealing the necks each time isn't necessary.
I do it at various points for different rifle depending on what I've seen work. My 30-06 gets it every 3-4 firings, the 260 gets it about every 3rd, the 6mm AI that I had was every 7th (very tight neck that required turning so low tension values were per norm). The bolt 223 I usually don't bother, when the necks split, I toss them and move on.
Work hardening also embrittles the material, so when you fire it the high strain rates cause the neck and sometimes shoulders to crack during bullet release. With annealed necks that are properly cared for this becomes a very unusual occurrence.