What do Extreme Spread and Standard Deviation really tell us?

Tim

Tim

Checked Out
Staff member
2A Bourbon Hound 2024
2A Bourbon Hound OG
Charter Life Member
Benefactor
Vendor
Multi-Factor Enabled
Joined
Dec 17, 2016
Messages
16,465
Location
A Glass Cage of Emotion
Rating - 100%
85   0   0
I'm new to reloading for precision (vs. volume) and am trying to figure out what's actually important.

I ran 13 different loads - 5 shots each - through my 6.5CM today looking for a good starting point to fine tune from. Since "group size" is more a measure of my skills versus empirical data, I decided to concentrate on ES and SD values.

Factory Match Ammo (Hornady 140gr ELD-Match #81500)
SD 10.9
ES 26fps
2727 average fps

I figured since that's "Match Ammo", and has generally excellent reviews, that this would be my benchmark.

All hand loads were factory fresh Hornady brass, loaded with the same 140g ELD bullet as the factory load. No crimp.

Accurate 4350 with charge weights between 39.0 --> 41.5g in 0.5g increments.
Best of this bunch was
SD 12
ES 32
2545 average fps

RL-17 with charge weights between 40.5 -->42.0g had much better results. With such a small sample size, it's hard to really be sure where to go from here.

40.5g
SD 6
ES 13
2710 avg fps

41.0g
SD 11
ES 27
2723 fps avg

41.5g
SD 11.7
ES30
2769 avg fps

42.0 g
SD 8.1
ES 18
2783 avg fps

These are all reasonably close to the factory "Match" ammo (SD 10.9, ES 26, 2727 fps).

None of the loads showed any pressure signs at all.

So, I'm thinking I need to focus on the 42.0g RL-17 load to start refinement? SD/ES both below the factory match with highest velocity. "Refinement" means OAL +/-, crimp, +/- a couple tenths of a grain, different primers, etc.

Does that make sense or am I over analyzing things?
 
Last edited:
http://precisionrifleblog.com/2015/04/18/how-much-does-sd-matter/

interesting read regarding point of diminishing returns when chasing SD. Short story is, once you get into single digit SD, the benefit of eeking out more SD improvement isn't a good use of energy.

just a 1% improvement in hit percentage going from an SD of 10 fps to 3 fps!
Click to expand...

“Regarding available ammunition types, a Standard Deviation (SD) of 20 fps is considered relatively poor consistency, and is generally representative of mass produced factory ammo. 15 fps is considered better than average for factory produced ammunition, but still substandard for those who handload their own ammunition. 10 fps or less SD is typically the goal of most handloaders, and very few commercially available ammo suppliers are capable of producing ammo with SD’s under 10 fps.”
Click to expand...

Interesting and backs up my thought that the 8.1 SD load is a good place for me to start tinkering.
 
You gave us the stats on each load but not the group size. I have had test batches with excellent numbers and groups that looked like a shotgun pattern and vice versa.
I'm not saying that the numbers are unimportant but they don't always tell the entire tale.
What is the end goal?? Hunting load or target load?? If hunting you need to be concerned with velocity so the bullet works as advertised but if it's a target load velocity is somewhat unimportant......
 
Bailey Boat said:
You gave us the stats on each load but not the group size. I have had test batches with excellent numbers and groups that looked like a shotgun pattern and vice versa.
I'm not saying that the numbers are unimportant but they don't always tell the entire tale.
What is the end goal?? Hunting load or target load?? If hunting you need to be concerned with velocity so the bullet works as advertised but if it's a target load velocity is somewhat unimportant......
Click to expand...

The goal is too reach out to 1,000 yards. Purely target/competition environment.

I did not post group sizes because...
- I had a Magnetospeed chronograph hanging on the end of the barrel messing with harmonics.
- group size is more a measure of my skill than the quality of the load

I can't say I ignored group size though. The best groups were the ones with relatively low SD. ~.7" @ 100yds.

Now that I have a starting point based on empirical data, I can better focus on refinement and skills.
 
Tim said:
The goal is too reach out to 1,000 yards. Purely target/competition environment.

I did not post group sizes because...
- I had a Magnetospeed chronograph hanging on the end of the barrel messing with harmonics.
- group size is more a measure of my skill than the quality of the load

I can't say I ignored group size though. The best groups were the ones with relatively low SD. ~.7" @ 100yds.

Now that I have a starting point based on empirical data, I can better focus on refinement and skills.
Click to expand...

Ahhhhh... yet another plus for the old fashioned Chrono's... data AND group size at the same time.
If you're shooting for 1000 then velocity is going to play a large part in your drop tables.
 
Since I usually don't pull out the chrono until I get a consistent group (at least 5 shots) I went back and checked some of my final loads and they all have an SD of less than 10, so I think there is a correlation between accuracy, precision, and SD. When it comes to simply shooting the best "group" I shoot at 200 yards or more. One hundred yard groups just seem to hide a lot of inconsistencies (shooter errors being the major one) and don't tell you much.
 
Again, I wasn't shooting for groups. I was shooting for data. Yes, group size is the ultimate data point. My goal of the exercise was to develop a starting point load that has a relatively high chance for success. Short of locking the rifle in a vise there's no way to take "me" out of the group size equation.

With a good starting point based on the data, now I can tinker with +/- charge weight, OAL, etc. and other variables as I also build skills.
 
I'm working up a load from scratch for one of my rifles. I may just forgo shooting for groups and collect ballistic data before hand, find the most consistent velocity parameters, and then see what happens. Shooting simply for groups is a royal PITA!
 
Last edited:
  • Like
Reactions: Tim
I agree with both trains of thought but speaking for my self I shoot for the accuracy first and then increase velocity until the load begins to lose the accuracy, back off a little and then stretch the range. I do that with any and all rifles that I load for. Let it be known that I have never loaded a 6.5..... It could be a different animal....
 
Its tells me how shitty some foreign made ammo really is. SDs on some 7.62x39 are in the 50s and higher
 
Back when I had a chrono and I learned not to let it drive me bonkers.....

I shot a load to see if it was accurate....say 20 shots or so in groups.

I checked velocity to see if it was where I wanted it and copied down the other numbers.....

I put one target up, fired it, saved it.

Each time I shot the load I put the saved target under the new one, saving the aggregate target and the new ones.

Later I had an aggregate target of up to 50-100 rounds to compare to all the individual ones and the story told me the true accuracy of the selected load regardless of the chrono numbers.

If I was happy with the accuracy, I kept the load and damn the chrono numbers. The velocity was only captured in order to calculate energy and trajectory for the hunt. The targets were captured to generate confidence it was accurate.
 
Last edited:
The spread, deviation, etc doesnt mean mean much. What matters it the consistency of the groups...... Usually with a load that groups well, you can note the average velocity with the bullet used. Later on if you change brass or powder for some reason, try to match the velocity and your groups will usually be as good.
 
just because a load will not shoot "good" at 100 doesn't means nothing. Take it out to 300 at least. You'll be surprised. Every load for my rifles has had the lowest SD (what I really care about) and ES out of the group testing. There's a correlation weather one wants to believe it or not..
 
Downeast said:
But if you had just shot 10 rounds and checked your SD you wouldn't have had to shoot 100 rounds to figure that out. :D
Click to expand...


AD, SD, ES, and all the other numbers don't mean much. Its all in the shooting. And anyway, the 5 rounds lots of folks shoot on a chrono and say, wow, good load or wow bad load.....statistically insignificant. It takes quite a bit more shooting to have statistical confidence that a particular combination is actually doing what your 5 little shots say it is. Basically, if yer not willing to spend time/money to shoot up the ammo, you'll never know for sure.
 
I think a lot of folks are missing the point I made about using SD to find a reasonable starting point for load refinement. No one technique or process is infallible.
 
Catfish said:
I don't think it's a good starting point. Theres no reason low SD should correlate with accuracy nodes of your rifle
Click to expand...

4 match rifles, SBR and a Garand. Accuracy node and lowest CONSISTENT SD appeared together. Not a 3 or 5 shot groups, 10 min.

If you graph your charge weight vs velocity you'll see the flat spots (node) and around those nodes have the lowest SD's. I had a wide node at 42.1 and 42.3 for on my 6.5 creedmoor. Loaded up 10x at 42.2 confirm SD's and did another load with seating depth.. done.
DVyPyTH.jpg
 
Last edited:
Djstorm100 said:
4 match rifles, SBR and a Garand. Accuracy node and lowest CONSISTENT SD appeared together. Not a 3 or 5 shot groups, 10 min.

If you graph your charge weight vs velocity you'll see the flat spots (node) and around those nodes have
DVyPyTH.jpg
Click to expand...

Did some your comment get cutoff? I'm very interested to hear the rest of that sentence.

By flat spot, are you referring to Loads 4&5 and 7&8 being accuracy nodes?
 
Last edited:
Tim said:
Did some your comment get cutoff? I'm very interested to hear the rest of that sentence.

By flat spot, are you referring to Loads 4&5 and 7&8 being accuracy nodes?
Click to expand...
I did it my accident with I inserted the image.

4,5 and 7 yes. 8 is questionable because of the 2836 because it was first time the brass being form. Neck tenison was to high on that round. As you go up the node gets smallers, so to speak. compare 4,5 to 7 and 9

Listen to Scott



 
Sharps40 said:
AD, SD, ES, and all the other numbers don't mean much. Its all in the shooting. And anyway, the 5 rounds lots of folks shoot on a chrono and say, wow, good load or wow bad load.....statistically insignificant. It takes quite a bit more shooting to have statistical confidence that a particular combination is actually doing what your 5 little shots say it is. Basically, if yer not willing to spend time/money to shoot up the ammo, you'll never know for sure.
Click to expand...

You can actually get statistical confidence in relatively few shots (once you have the right load and control the variables). Everything beyond is simply increasing the confidence and decreasing the SD.

I knew my graduate stats course would come in handy....
 
Just looking back and reading this thread. I had an old college stat professor tell me that in every day situations a sample size of less than 30 is the realm of the boogeyman and you need to to avoid his enticements and pitfalls. That would be a lot of shooting but chances are you would have a high degree of confidence (check your p values!!). I believe now the tacticool term for larger sample sizes in the geek/nerd/stat world is referred to as the "power value".
 
You must log in or register to reply here.
Back
Top Bottom