Your concerns are right, and what you're seeing is typical in that approach. The reason grouping falls apart with your seating testing is because your doing the testing from your powder node. That testing is then collapsing the powder node -causing 2 big changes at once. It's tail chasing. My approach is to mostly follow Berger's recommended full seating testing. A key point in that is to lower the charge considerably for the testing. I don't know, they may suggest that for safety, but it's actually critical to best results. It should reasonably take you away from any powder node, and with that you can focus more on results of seating, even while grouping is large at that point. THEN, with best coarse seating, you can move into powder testing for either tightest grouping or low Es, or best cold bore accuracy, etc. Your choice. THEN after powder with best coarse seating, move to fine tweaking of seating (inside it's window) for tightest group shaping. This doesn't take more shots, as full seating testing can be done while fire forming brass, and it doesn't take a lot of shots for coarse testing with Berger's recommended increments. Another worthy point in this approach is that you'll be going into powder with tested seating, instead of something pulled out of somebody's butt. After all, how can a bullet maker or anybody else predict where your COAL should be? They know nothing of YOUR chamber throat. If you go by their guess then how do you know your seating -during powder testing- was not the worst possible? This, potentially causing really crappy ladders, and masking best powder load. The same applies to primer swapping. It's likely a waste of effort to do that from a good powder node..