OH BOY! Another heated debate!
I'm not even going to put my thoughts into this one simply because it always turns out to be a waste of time, but I will toss this in for the sake of pondering this subject.
This isn't the only time this has happened during my vast amount of load testing over the years, but it's a good example that can be repeated by anyone else.
I was working up a load in a Browning A-Bolt II Varmint 223 with 40g Balistic Tips, CCI BR-4 primers, Federal cases, and varying weights of AA 2015 BR powder. My velocities were running in the expected ranges, and absolutely zero signs(yes I know how to check the right ways) of high, or even maximum pressure with these combinations. Being the eternal experimenter I am, I decided to increase charge weights 1/10 grain at a time, and see how it went. I'm not going to list here exactly my charge weights, but I've repeated this same thing in a couple other 223's, at slightly different charge weight ranges so I'm sure you can repeat it too if you choose.
I, as expected, gained a slight but definate increase in velocity(5 shot averages) with each increase in charge weight to a point, and then with only one more 1/10 grain increase, I lost about 15-20 fps. I then shot 5 more at that level, and got the same results. This is with Federal cases, and there's no doubt that if I was running high in pressure, it'd easily start to show when I miked the cases heads. I then increased another 1/10, and lost another 30 fps or so for a total decline in velocity of around 50 fps. Again, I tried that charge with identical results.
I scratched my head for a while, and literally gave myself a headache trying to figure this thing out at the time. I dismissed it that day as a fluke, or chrono glitches, but I ended up going through it again a few days later just to double check myself. Sure enough, exactly the same results. I decided that the only possible reason for this was unburned powder being pushed down the barrel behind the bullet, so I rigged up a back stop with plywood, and a cardboard layer taped solidly to the front. What I saw convinced me that my theory was right. With every shot at those charge levels there was a bullet hole with a little shotgun pattern around it. Once I dropped back to the charge weight where the last velocity "increase" was measured, this stopped. I imediately called the tech at Sierra to ask his opinion, and he confirmed my findings as something Sierra already knew.
You can draw your own conclusions from this...