I tracked a bulk lot of Winchester .45 ACP brass through 50 reload cycles. By then I had probably lost about half to the range grass and about a third of the remaining half had split, so I was down to the last third. These were mostly loaded with target loads. The headstamp letters had gradually peened over and were disappearing.
The most interesting thing was these cases were all about 0.025" shorter than when they started, losing about half a thousandth per load cycle. That happens when pressures aren't high enough to make brass stick to the chamber wall and the chamber is tapered. So instead of stretching, the whole case just backs up to the breech, where it expands shorter and fatter to fill the tapered chamber. The sizing operation never quite fully returns it to length.
I'll add that the reloading for that lot was all on my Dillon Square Deal. That matters because the Dillon dies are small base or at least smaller base than my other .45 Auto dies. That forces all brass to work. For example, the Remington brass that used to work harden to the point it was so springy my Lyman carbide sizing die could not make it small enough to hold onto a bullet, works fine in the Dillon. So the splits were partly the result of the brass being worked a bit more at each sizing cycle than was strictly required.
The annealing has a lot of variations. The stuff I've seen written by metallurgists seems to me to be in general agreement with the linked article. I'll add to it that a general symptom that tells if you are overheating during annealing is that the necks get weak enough to start splitting in as few as three reloads afterward. If you are annealing every few shots for consistency, that's one thing, but if the necks actually start splitting again in three to six reloads instead of in twelve to twenty or so, then you are getting the brass too hot.
Dropping annealed cases into water isn't to help the annealing process. It so you can more quickly pick the cases up without getting burned. It started with the old method of setting cases in water deep enough to keep the head from annealing, then getting the neck red hot (too hot, as it turns out) and then knocking the cases over into the water tray. This let you pick them out with your fingers and move on the next set without waiting. If you stick to reasonable temperatures in a short enough time, you can leave the cases sitting where it is until it is cool enough to handle. If you have enough work space, why not?