Standard Deviation

Joined
Aug 14, 2013
Messages
16
Location
utah
Can I test standard deviation on a warm barrel or am I better off using a cold barrel because in most hunting situations the first shot will be a cold barrel??? How many shots to get an accurate standard deviation? I use a Lab Radar.

Can I test different loads standard deviation with just a few minutes between firings??
Thanks
 
SD only matters for large sample sizes. It you're going to put less than 10 data points into the population don't even bother with it at all. If you want to truly care about it you need to chrono every shot over multiple sessions and and conditions to get a true load SD.

Or in your specific case, multiple cold bore shots over multiple range sessions.

Anyone can get a 3-shot SD of 1fps in a single range session. A correctly tuned load will be sub-10fps for dozens of shots over time in different conditions.
 
Before worrying about measuring SD, do you know what SD you need to accomplish what you intend to do? What cartridge are you shooting? What bullet? What distance? What Target?
Big Game. Deer and Elk. Shooting steel and just want sub 1/2 moa groups. 7 SAUM. 168 Berger VLD Hunting. Out to 1200 yards. 24" Proof barrel on Tikka Action, and Mesa Precision stock.

Been trying out ladder testing at 800 yards and also looking for SD of 10 FPS or less. Currently using 64 grains RL-16, Fed Mag Match primer, ADG Brass.
 
By my estimates you are looking at 95% of your rounds deviating +/- 6" due to velocity (I assumed 2000 ft elev). In comparison at 1200yds that's equivalent to misreading the wind by 1mph. Its likely you will miss the vital zone due to wind in the horizontal. You are also looking at 1-1/2 seconds of fligh time. A lot can happen in that time and you are losing a lot of energy at that range. I personally don't think I would be willing to risk that shot.

Depending on your location and terrain vertical wind and temperature variations may be more of issue that SD.

But to your question about number of rounds, let me answer it this way:
Shoot 10 rounds measure 3100 fps average 10fps SD, 95% confidence interval 3093-3107 fps SD 6.9-18.3 fps
Shoot 20 rounds measure 3100 fps average 10fps SD, 95% confidence interval 3095-3105 fps SD 7.6-14.6 fps

Another way to look at it is you would need to have a 5.5 SD over 10 rounds or a 7 SD over 20 rounds to have a 95% confidence that you have a true 10 SD.
 
An interesting stat (SD) useful for repetitive shots like in NRA NM (600 yd), volume steel shooting, or colony rodents (100's). Good for loading up moderate to large volumes of ammo for a certain application. More shots, better SD indication. Same spread for 5 & 40 shots, 58 fps, but SD's of 20.165 (5) & 15.373 (40). A moderate velocity disparity would screw a shot up & a nice SD value for the entire string would have little value. My personal & self-imposed limit for shooting deers & whatever is 400 yards. Should I be shooting pop can size or bread loaf size rodents at 600 plus ammo with a low SD will result in a higher % of hits.

Screenshot (488).png


Screenshot (491).png
 
By my estimates you are looking at 95% of your rounds deviating +/- 6" due to velocity (I assumed 2000 ft elev). In comparison at 1200yds that's equivalent to misreading the wind by 1mph. Its likely you will miss the vital zone due to wind in the horizontal. You are also looking at 1-1/2 seconds of fligh time. A lot can happen in that time and you are losing a lot of energy at that range. I personally don't think I would be willing to risk that shot.

Depending on your location and terrain vertical wind and temperature variations may be more of issue that SD.

But to your question about number of rounds, let me answer it this way:
Shoot 10 rounds measure 3100 fps average 10fps SD, 95% confidence interval 3093-3107 fps SD 6.9-18.3 fps
Shoot 20 rounds measure 3100 fps average 10fps SD, 95% confidence interval 3095-3105 fps SD 7.6-14.6 fps

Another way to look at it is you would need to have a 5.5 SD over 10 rounds or a 7 SD over 20 rounds to have a 95% confidence that you have a true 10 SD.
You nailed it Doom - the biggest factor in LR shooting is your ability to estimate wind. I don't think anyone can estimate it within 1 mph all the way out to 1200, and when you do the math, hitting a 10" target at 1200 yards 100% requires an estimate that is less than 1/2 mph. To your point, it can change that much during the TOF.
 
SD - Standard Deviation:

"In statistics, the standard deviation is a measure of the amount of variation or dispersion of a set of values. A low standard deviation indicates that the values tend to be close to the mean (also called the expected value) of the set, while a high standard deviation indicates that the values are spread out over a wider range."

I use SD as a general indicator for ammo consistency. This assumes a bell-shaped curve (normal distribution) with low to high values and a mean midway.

Imagine, shooting a 5 shot string, producing a SD of less than 5 fps, but round #6, if tested, would produce a SD of 25. Shooter confidence for making that 1,000-yard would go from nice at a 5 round SD, to terrible at a 6 round SD.

Playing around with this stuff will also provide average low and high range values at a certain % interval or "confidence", like 2 1/2 % at each end or 95% confidence. This will provide low & high averages. I produce moderate amount of ammo lots, like 250-500 rounds per lot, and try to get maximum consistency and like to see both tail ends of the curve show minimum differences.

Shooting deers & such at 1,000 yards, is way beyond my ability. I usually rely on 2 sighter shots to get within 4 inches of target center at 1,000 and can't do sighter shots for shooting deers. Having nice ammo with a low mean velocity spread at each tail end assures an elevated hit rate on pop can size rodents at ranges like 450 - 600 yards provided the bullets are pointy and have zippy velocities.
 
The 1200 yard stuff is at the range. Most of our hunting situations are 600 or under. However on occasion I have shot game out to 1000 yards conditions permitting.

My SD question was more for my understanding to better load development as it seems useless to get a good average SD after barrel starts heating up without quite a bit of time between shot's.

I maybe need to work on "tuning my load's more" to get what im after. Might even start turning necks and using a mandrel to make neck tension more consistent. Have a buddy that says the mandrel really helped his bullet runout when he started using them.

Not a benchrest guy just love shooting and learning!

Thanks
 
My SD question was more for my understanding to better load development as it seems useless to get a good average SD after barrel starts heating up without quite a bit of time between shot's.
Direct answer to the question - it shouldn't matter. Barrel temp and chamber temp shouldn't have such an impact on a load that it changes anything.

If your cold bore shots are going to be in an extreme conditions you should verify your POA/POI in those conditions. But the load itself shouldn't be affected unless the temp is on the too hot side of the spectrum, 90* or over.

I maybe need to work on "tuning my load's more" to get what im after.
In this case tuning means finding a wide enough powder node that differences in POI from shot-to-shot velocity are minimized. There are several ways to get there, but getting shots to waterline consistently at under 1000 yards isn't terribly difficult, and SD/ES can be pretty awful on paper but still print good targets.

There are different solution sets to meet differing ultimate end goals - smallest possible group, widest possible nodes, furthest possible minimum velocity. Defining what you need a load to do determines what all you need to do. I've hunted with God-awful 100+FPS ES on loads before because the rifle printed bugholes at 100 yards, and guess how far the feeder was? Flip side I was pulling my hair out over a 10/30 SD/ES load because my rounds were impacting +/-10" at 500 yards. FPS was phenomenal, stats were nice and tight over 60 rounds, but there was zero precision at all in the load, mainly because it was a volatile powder and it was 104* outside. The tightness of the velocity range was probably because I was so overpressure that I should have been scared 🤣

Decent components shot through a decent barrel should net you 1MOA groups out to 600 yards their own, shrinking that MOA down is where the work is. (224s/smaller and sub 90gn bullets excepted due to lateral drift).
 
Last edited:
Can I test standard deviation on a warm barrel or am I better off using a cold barrel because in most hunting situations the first shot will be a cold barrel??? How many shots to get an accurate standard deviation? I use a Lab Radar.

Can I test different loads standard deviation with just a few minutes between firings??
Thanks
TLDR: Lets stay real about single digit SD and shooting stats. Setting a realistic baseline expectation is difficult for a beginner because the internet is full of anonymous liars and not enough objective honest posters willing to share real stats.

This won't be a tutorial on statistics, but one or two points on your original list of questions quoted above.

You can test on a warm barrel, but that will represent that condition and the history on the barrel. However...

If you want cold bore, then no, you will need to test cold bore and include the cleaning and storage history conditions that matter to you between shots. Do you store your gun dirty or clean every time for example?

Does the gun get bounced around in a scabbard? Temp cycled into and out of hotel rooms or trucks? Lint or dirt in the barrel between shots?

When we are lucky, a hunting rig shoots the cold bore shot into the same group. Some rigs do not and there can be several reasons like the cleaning and storage conditions, or like in the case of rimfire if the ammo is wax lubed, etc..

If folks leave too much cleaning or preservative residue in the chamber or bore, it can cause a cold bore shift.

"How many shots?" On the order of 15 to 30 shots or more. The higher the scatter, the more shots it will take. The more process control problems you have, the more batch breaks you introduce, the more shots it will take...

Highly disciplined loading with a tight scatter can show a lower risk answer sooner.

"Can I test different loads standard deviation with just a few minutes between firings??" This is another "it depends" question. I would avoid it, but you can try. The issue is different powders and bullets versus your cleaning regimen. If those loads leave a residue that the next load doesn't like, then your data is affected.

I generally find jumping between copper solids and regular bullets means cleaning in between. Same issue with jumping between ball powders and stick powders in some instances. You will spend some shots to learn the ones you need to plan to clean between changes.

The cleaning between testing different bullets or recipes can have an effect. There is basic (chemical solvent) cleaning, and then there is scrub to bare barrel cleaning, and then there is also how many shots to stabilize the barrel after those points. Even the shot count on the barrel can change the issue comparing a fresh barrel to a worn barrel.

Being methodical with your testing includes being realistic about the type of gun and the type of shooter you are. A low experienced shooter, with a light carry gun is one extreme, and an accomplished high level shooter with a heavy accuracy rig is another. We should not expect similar baseline results.

I will take a risk of attracting flak for saying this, but unless you and your equipment are capable of tight scatter, then the shot count required to even detect a change can be so high that many folks mistakenly make the claim that some factors make no difference when in fact they do. You can often find that their guns and their shooting are the issue and that means taking more shots to get the stats to resolve the answer.

Shooters who boast or cherry pick their data don't do us any favors. Don't be unrealistic about your expectations when folks show you their "filtered" results instead of all their data.

Without being rude and uncivilized on the forums... when we take the extreme of a low experienced shooter with a low quality light carry gun, who makes claims that their baseline results are equal to those of a very accomplished shooter using state of the art gear, then we take it all with a grain of salt and don't fall for it. Too many folks are disappointed in their own stats but they are being unrealistic about their expectations when they baseline themselves at the wrong level.

I hear a lot of folks comment on shooting game at 1000 yards or more. I wasn't there and I would not call them out... however, I have pulled pits or spotted targets of all kinds, and worked with DoD and LEO folks of all kinds, and the first round hit percentages at 1000 yards on a cold bore without other shots with other guns in those same conditions... leave me with a lot of doubts.

A good rangefinder, and climate data from things like a Kestrel help improve the percentages, but even the most accomplished shooters are not close to 100% out in the field at those distances from a cold bore.

No wind flags, no sighters, no way to know cross canyon winds, etc., etc., means that it takes bluebird conditions, state of the art gear, and an accomplished shooter, to call it ethical. "Know thyself" is ancient wisdom that applies to shooting and hunting ethics. YMMV.
 
I like to play around with stat stuff. Should I be responsible for producing vast quantities of ammo for certain endeavors like military or demanding commercial use I would insist on preforming a stat analysis on each ammo lot. Ammo lots that exceeded reasonable stats for the intended use would undergo additional testing.

How that ammo lot was used would be beyond my control and any value prediction based on exemplary SD data could be negated by limitless use variations.

For my use, knowing environmental conditions, bore condition, component variations and target demands having an ammo lot of 250 - 500 rounds with SD values between 20-35 fps is more than adequate. I usually pull 10-20 rounds at random from the ammo lot for chrony data, get SD then get low/high estimates at each 2 1/2 % tail (95% probability) of normal distribution. T .05, two tails.

Talking about glop in bores, wind, and non-stat stuff is another subject.

Should I go out to make that first shot count, if possible, I fire 2-3 fouling shots thru a clean bore. Stat data collection is from cool (ambient temperatures) but fouled bores. Like drive out to shoot that trophy rodent with a slightly fouled bore. Clean the bore at the range then shoot 2-3 more.Screenshot (492).png
 
Last edited:
Top