Here's the way I approach the issue. I load 3-5 rounds at the start load which is 10% less than maximum. I then load the next test batchs at -8%, -6%, -4%, -2%, maximum. If the maximum load is 50 grains of powder X then the inrements would be at 45gr, 46gr, 47gr, 48gr, 49gr, and 50gr. After a fouling shot, starting at the start level I shoot for group and look for pressure signs. If any pressure signs are encountered before reaching maximum I stop and consider the previous load to be the maximum with the component mix I happen to be using. Any remaining rounds are then disassembled and the components get reused. Examining the groups determine which charge is the most accurate. At that point if I feel I can squeak out any more accuracy I'll load up the next batch with much smaller increments around that "best" load so far.
To me this seems like the most methodical method of finding the most accurate load.