The big issue is trying to make a decision with a sample size of 1. Tests like this look good because our brains are really really good at picking out patterns, to the point we'll find patterns in noise. There's a lot of math and statistics behind why tests like this don't work, but the easiest way to understand the problem is to do a test more than once and see how inconsistent the results can be. You'd need so many rounds for a ladder test to give you good data you'd burn out a barrel just testing a few bullets.I'm picking up what everyone is putting down. The method Scott Satterlee uses is not popular here
Lesson learned. Thanks for the input.
Velocity nodes don't really exist with the exception of some powders that need a high case fill to give minimal velocity variation. You control SD by how consistent everything is from round to round. Ladders are handy for picking a charge that'll give you the velocity you want, but don't do much good beyond that.