The tale of two cases
It was the best of brass. It was the worst of brass.
My Remington 700PSS in 308 will shoot scary-accurate groups when fed my personally developed handloads. These are 168gr Sierra BTHP bullets, seated to an OAL of exactly 2.815" over exactly 42.5gr of IMR 4064 powder in BHA match cases primed with CCI 200 primers.
However, when I build rounds using this same recipe but substituting the BHA cases for something else, the rifle will not group. I recently tried an experiment when I built up a batch using the recipe above, and another batch using the same recipe but substituting Federal Gold medal match brass instead of the BHA cases.
The chronograph tells the difference. The BHA loads clocked in at 2571.1+-21.4 fps, while the FC loads clocked in at 2643.3+-6.4 fps. So even though the velocity spread of the FC loads was smaller than the BHA spread, the BHA loads group much better. This says more about the rifle's barrel than the load; apparently the 700PSS really likes the 168gr Sierra bullets to be going at about 2550 fps.
So, what explains the almost 5% velocity difference between the BHA loads and the FC loads? Measuring the case volume (grains of water) showed that the BHA brass has a volume of 56.6 grs of water, while the FC brass has a, smaller, volume of 54.3 grs of water, a 4.2% difference.
So, there you have it: a 4.2% decrease in case volume results in enough increase in chamber pressure to cause a 5% difference in velocity, and a 5% change in bullet velocity can really throw off a finicky rifle.
MORAL OF THE STORY: Don't substitute components when handloading.
CORROLLARY: If you must substitute, start over with the workup.