At what distance does standard deviation start to really matter?

[Varminterror]

I really can't fathom how LR could ever be considered a suitable replacement for Varget. One of the HIGHEST temp sensitivities vs. one of the lowest...

Big SD, the first stones I turn over are neck tension, case fill ratio, and ignition.

A few things of note which I bring up whenever folks ask this particular question...

• First, recognizing that ES for valid sample sets typically will be ~4-4.5x the SD, a 31fps SD likely corresponds to an ES of ~125-140fps. So we're talking about potential for a LOT more than just 31fps shift in the ballistic engine. We may not realize all of that potential in our group, since we have other error factors within the group which do more to cancel each other than they do to compound, but our potential error is a LOT bigger than just the difference in drop calculated with 31fps difference.

• Elaborating that a bit: recall that SD is "Standard Deviation," meaning that ONE SD is the statistical expectation for any given shot to fall away from the arithmetic average. It's not a maximum, it's a STANDARD. We expect only ~2/3 of shots to fall within +/-1 SD above and below the average, with ~40% of those within +/-1/2 SD (~30% of shots between +/-1/2 and +/-1), and ~15% of those to fall outside of +/-1.5 SD - this means, for an SD of 31, we expect 4 out of 10 shots to fall within a 31fps span above and below the average (+/-15.5fps), and then expect 1 out of 10 shots, and ~50/50 to expect 2 out of 10 shots to be outside of ~46fps from the average... Pile that together and you're expecting one shot MORE than 46fps slow (or fast) and a bunch spread up to 31fps faster than average (or slower), so we expect to experience at least 77fps spread in a group, even when we trim off the outliers, knowing we should experience 125-140fps spread if we include the worst outliers.

• Tracking Varget in Dashers and H4350 in 6 Creeds, under 105's, every kernel accounts for ~1fps/kernel (.92-.97fps/kernel measured), with 62-64 kernels per grain. Experiencing an Extreme Spread of 125-140fps is around 135-150 kernels, which is equivalent to more than 2 FULL GRAINS of powder measurement error. Would you tolerate a powder drop or dispenser throwing charges so inaccurate that they were only +/- 1 FULL GRAIN?? I wouldn't, so I REALLY don't want to tolerate any other part of my handloading process which yields the same potential error.

• For folks who subscribe to the ideas of either velocity nodes or optimal barrel time, or really, ANY kind of node, can't ignore the fact nodes won't be 125-140fps wide, or 2 FULL GRAINS wide. Most of us are happy to see nodes 0.3-0.4grn wide, that's ~18-25 kernels, which is "deferred increase" of ~16-23fps... Do we expect a node to swallow a difference of 125-140fps?

 

[Ronnie milford]

For what it’s worth I can actually shoot within 40 feet of where my reloading room is located that allows me the ability to shoot tune etc in same conditions. I have actually seen as little of 2 flakes of powder or .02 grains change group size’s. .001 seating depth change groups drastically but the game changer for me is consistent seating pressure all kinds of different ideas out there and think they all will work. But ultimately you the shooter must find what works for you. For example a lot of the rifles I build the shooters feed back valuable information but that info is for their barrel doesn’t mean it will work in mine Find what works best for you and never get caught up in what your hearing at the range it’s most likely BS
just saying
Ronnie

btw
my best groups to date have less than 5 fps extreme spread again just works for me

 

[Dave Way]

I haven't shot paper past 2,075 yards, so I would answer somewhere 2,076 or more. That said, 31 on SD is quite a lot. My experiences referred to above are more like trying to see 20 vs 5 on the ES, and more often than not the paper will print opposite of the crony.....but only out to 2,075 lol.

Tom

Agreed. I have had higher SD loads print better than low SD loads at 1K…more than once. If you’re loading practices are sound you should have SD’s in the low double digits, if not single digits.

I’m not trying to be critical but I would recommend examining your reloading processes. You should be able to improve your SD’s but remember the paper is all that matters in the end.

Dave.