Another engineer, here.
Two engineers meet coming from opposite directions on a campus path. One is on foot, while the other is on the most amazing bicycle the pedestrian engineer has ever seen. It has a carbon composition frame, hydraulically controlled gears, built-in heads-up navigation, physiology monitors, seats that adjust from recumbent to stand-up positions, aerodynamic wheel hub covers, and several features the pedestrian engineer cannot immediately discern the purpose of. So he waves the rider down and says: "That is the most amazing bicycle I've ever seen! Where did you get it?"
"Well," the bicycling engineer responds, "that's a funny story. The other day, I was walking down this same path just the way you are, when I saw the most beautiful woman I'd ever seen riding toward me on this bicycle. She was dressed in the most amazing set of bicycling clothes, with moisture-wicking sun-reflecting material, aerodynamic filler pads to control airflow over her body, pumped coolant for neck and underarms, and built-in LED safety light strings along the edges. As she drew near, she suddenly veered off the path and dumped the bike on the lawn, jumped aside and pulled off the suit and all her underclothes, and threw her arms wide open and said "take what you want".
The pedestrian engineer observed, "good choice. The suit would never have fit you anyway."
Petey308,
Several annealing questions come to mind. I some time ago read the John Klein paper linked to in an earlier post. One of the observations from reading it is the energy required to initiate and complete the stages of annealing decreases as the potential energy stored in dislocations in the brass crystal structure increases. Thus, the same time and temperature needed to anneal a 50%-hard piece of cartridge brass would also anneal a 90%-hard piece of brass but would do the latter, IIRC, about 6 times faster. In other words, achieved annealing doesn't depend only on time and temperature, but also on how hard the brass is when at the start of the annealing process. Presumably, that starting hardness is fairly consistent if you anneal every load cycle. But I would be interested to hear, have you compared the hardness of a neck annealed every load cycle with one annealed using the same time and temperature protocol but only after multiple reloadings when it is harder? Does the harder piece come out softer?
Another question has to do with the purpose for annealing, which I think you mentioned briefly. The manufacturer does it to prevent season cracking. The thrifty shooter may do it just to prevent neck splits, for which getting through recovery is all that is required to get adequate stress relief. Then we come to the AMP maker's assertion that getting down to around HV100 is needed to get precision improvements in group size. I know Bryan Litz borrowed and used an AMP annealer and could find no difference between the performance of a set of cases fired ten times and annealed every time as compared to a set fired ten times without annealing. Indeed, he reported a small, but not statistically significant superior performance of the unannealed cases. I am thinking the difference in results may lie in resizing methods. Brass, regardless of hardness, has a constant modulus of elasticity. But for that to keep performance the same, the amount of stretch has to be the same, and necks that are getting harder spring back more after either resizing or expanding tend to throw that off. I am wondering if you are aware of anyone having tried adjusting resizing with mandrels to keep neck ID constant for seating as hardness is increased and if that might not produce the same improvements credited to annealing? It would be a lot harder to do, but might be interesting to try out to satisfy curiosity.
Quiet Texan,
In post #100 you referred to the SD being a sixth of the ES. This would require a sample size of around 400. The statistic
ζ(n) value, is the number you divide the extreme spread by to get an estimate of SD of a normal distribution. It grows with sample size because the larger the sample, the more opportunities you are giving for less probable random values to show up, and they do show up and are farther from the mean than more probable samples. For a sample size of ten, the zeta of n value is 3.078 (follow the link). A plot of the value vs. sample size of up to 100 for an SD of 1 is below.
View attachment 324496
. If comments made here got complicated, that's just the direction it went I guess.
. Any others I should avoid 
?
