Please don’t read this message is rude because I want to be direct about my thoughts here but I’m also very passionate about this topic so I may come off strong 
“Flat spot” where small variations in powder weight do not cause large swings in velocity? Does not exist.
For example, here’s a powder charge ladder test I did.
Looks like a flat spot at 38 to 38.5, and another at 39 to 39.3. If the test is valid, if I go through all the work of running it again, I’d get the same result, right? Here’s what happened.
Completely different result. The test is simply invalid because too few shots are being taken at each setting to be statistically significant.
So what happens if you do the entire test NINE TIMES? I did exactly that. Here is averaging all of the shots at each setting.
The result? An almost perfect straight line. Each time I add more tests, the line gets straighter. There’s almost not even a wiggle in that line.
I believe this testing proves my hypothesis. There is no such thing as a “flat spot.”
That frustrates reloaders. Frankly, this method is how 99% of guys have done reloading for decades.
Now, you mentioned the “Hornady method.” I’d definitely take issue with the AI summary, because that actually the opposite of their approach. I load the same way they do, and it’s following statistics
The goal is not to find a nonexistent “flat spot” where velocity, es and sd, vary less. There is no point that somehow magically becomes flat despite having more powder at that setting.
What DOES make a difference is tuning the load to jump the right amount, and have the right pressure signature for the bullet you are loading. But to know which is the best setting, you’ll need to make BIG changes and shoot LOTS of shots in order to have a statistically significant test.
Suppose I’m testing a 6.5 creed handload. First, I’m gonna just pick a known-reliable powder for accuracy in that cartridge. If you can’t get a good result with h4350, something is wrong. So that’s what I’m using. I won’t even try other powders unless I really get desperate.
Next, the powder charge. Depending on my velocity goals, I’m usually a half or full grain down from the listed max. Usually works.
The big one is the bullet. I usually try two or three. To get a valid result, it’ll probably take 20 shots PER BULLET. I wouldn’t be mad though if you just shot a ten shot group with each as long as we acknowledge we are just cutting out any obvious failure in bullet choice but recognizing we won’t have enough info to choose between the mean radius of .32 and .35” or other small differences in bullet accuracy.
Usually, that’s the full process. But it’s maddening, right?! How do you know if 41.y is more accurate, or 41.5?!? You simply can’t. It may take 100 shots to know how that small of a change will impact accuracy because they’ll be so close.
Even if you did shoot 100 shots accuracy testing those settings, the barrel literally wears out slightly in 200 shots so you wouldn’t know if it was the bullet or the barrel condition.
Then there’s bullet jump. It can definitely impact things. I like to test at least 10 and probably 20 shots at each setting, and I’ll test BIG jumps to see if it improves accuracy. often test 30, 60, and 90 though off the lands.
But that’s all we can really do. It just takes too many shots to test fine differences. For the best results, test the BIG levers that impact the load accuracy the most—which bullet, and jump. For powder charge, just pick one that will hit your velocity and will not go over pressure.
There are so many variables that most reloaders create invalid tests that make them feel better, but simply don’t pass scrutiny.
