After giving this some more thought, I remembered that I had done a 5x5 with new Sig cases. The results were about the same as the 3x loaded Lapua cases, so that maybe blows my theory. But, according to my buddies, non-Lapua cases can have wide volume/weight variations so...
The theory was that there it is volume differences in some cases that there were enough in numbers to be randomly, but consistently distributed throughout the cases. Looking at the above charts, it looks as if the outliers are significantly slower than the other 4 in the same group. The theory can even explain the large and small discrepancies, and even the good number groups.
I'll be as brief as I can. The nature of the velocity spread is essentially one round out of five is an outlier significantly slower in velocity than the other 4. Further, it seems that this spread is due to something I'm doing. That is, my buddies load the same way and while they do get spreads they are much tighter than mine. I have eliminated the gun as the problem because I've seen the same thing in two bolt guns and three ARs. This just almost has to be something I'm doing or not doing that's different than my buddies.
What I came up with "checks all the boxes" so to speak. Here it is. I typically load a 5x5 set. Five charges of five shots per charge. I push the pressure limits. My buddy and I closely watch for pressure signs and see nothing of consequence. I changed presses and along with that I changed to a RCBS shell holder. I started noticing a lot of fired cases would not slide into the RCBS shell holder. To see what would happen if my buddy loaded my components, I gave him 25 cases, primers, bullets, and the powder I was using. He too found some of the cases very difficult to slide into his RCBS shell holder. And, his reloads produced about the same results as mine. So it isn't me! It almost has to be the cases. Bullet weight variations of +/- 0.2 would not cause the wide velocity variations, nor would seating depth variations of +/- 0.002" - according to QuickLoad.
That's when it hit me - it almost has to be the cases. I believe those problematic cases are the cases that are exposed to the higher charge levels. Force exerted by pressure is equal in all directions. We know that the pressure expands the cases or why would they have to be resized? What we may not realize is that pressure can expand the case in another directions - it's length. What if the higher pressures are forcing the case to expand into the rim groove of the case? That would make the case not want to slide in the shell holder. Although I have not confirmed this and don't know that I can (I've thrown all those cases out), I suspect that also increases the volume of the case because the dimensions have changed. A given charge in a given area will produce a consistent pressure which will produce a consistent force on the bullets, which will accelerate the bullets to very near the same velocity. That's exactly what's happening with my buddies's loads - they are seeing fantastic numbers/consistency.
Further, they are not pushing pressure limits like I am. None of their cases resist sliding into the shell holder!
So does this explain things - yes! But at this point it's just hope and a theory. Here's what it answers:
1- I'm seeing this problem and my buddies are not.
2- My buddy loads using all my components, the same I'm using, we get the same results I get.
3- I have problems getting some cases to slide in the shell holder and so did my buddy
4- This one needs some explanation. Why are some 5 shot strings showing excellent results, some are fair, and some are just awful?
I almost always shoot a 5x5 array, i.e. 5 incremental charge weights of 5 shots each. Once shot, all the brass is dumped into a container. When I get a hundred or more, I clean them and reload them. So let's realize that in this mix of brass on a 1 to 5 ratio, these cases have been exposed to five different charge levels. Some levels high enough to elongate the case. IF that increases the internal volume, it would lower the pressure and hence the volume. Since these elongated cases would be present throughout the cases, there is a high probability of one showing up in every charge level of the next reloads. It's also highly probable that I pick 5 cases in one group that does not have an elongated case, and that's when I see the very low numbers.
Further all combinations of charge levels can show up in any given 5 shot group. That would explain why some groups can be excellent, some fair, and some awful.
It has to be something. The fact that my buddy and I see the same results when reloading my with my components indicates it isn't me or my reloading method. I'm hoping, since I/we have eliminated everything else, I can substainiate the theory, but my confidence level is quite low as nothing I've tried changed anything.
This does look promising though - it does fit the strange results.
I happen to have 20 new Lapua cases and many more new Sig cases. I am going to pick one charge where I got good numbers and load 20 rounds with Lapua cases and 20 rounds with Sig cases and measure them. The Sig is just a tack on. My buddy said when he switched to Lapua cases his numbers went down. So I'll see how the two match up.
The nagging thing here is I've shot a LOT of new Lapua brass trying to figure this out, and it too gave similar results to what I'm seeing. But, hope springs eternal they say, so maybe knowing what to look for will be the key.