SD question...

[IdahoSharpshooter]

SD kills me.

If a load has an SD of 10(fps) is that 10fps total, or +/- 10fps? I am once again immersed in Brian Litz's series of books. Everything seems fairly straightforward, and then SD and how it fits in the quest for accuracy (read: consistency of MV).

thanks,

Rich

 

[memilanuk]

SD is used to estimate i.e. guess forward, the future spread of the MV. Assuming the samples follow something approximating the normal / Gaussian distribution, then you can (generally) assume that approx. 68% of the subsequent values will fall between +10 fps and -10 fps (+/- 1 SD) from the mean (average) muzzle velocity. Approx. 95% will fall between +20 and -20 fps (+/- 2 SD). 95% works out to about 19 out of 20 shots, btw. If you really, really want to be sure - or just really depressed - you can go for +/- 3 SD, or approx. 99.7% ... so *all* your shots would (theoretically) be within +30 fps and -30 fps from your mean MV.

Now do you see why its important to get that SD value as low as possible? If I have an SD of 5, that +/- 2 SD would be a spread of 20 fps, vs 40 fps.

 

[IdahoSharpshooter]

I shot Hunter Class BR for several years. 308 Win, 150gr Sierra MK's. I also had access to a seed ware house from about 11pm to 5am one or two nights a week. Over 3100 rounds, enough to be statistically valid from one rifle, same lot of 4895, and MK's. Tested with a 36X Leupold, never had more than 14fps ES for 10-shot groups. Biggest group over that 3 years was .347".
SD did not match target results, the bell curve only exists, compared to my personal target results, in somebody's imagination.
SD is theory, recorded results are reality. IMHO, +/- one or more SD's is an cop out. It is always, in the data I have read, speculative, and relies on the quaint notion that one cannot maintain the initial accuracy or velocity spread level over time.

SD assumes (and we all know about ASS/U/ME) deterioration over time.

 

[memilanuk]

Sounds like you have it all figured out, then. Have a good one.

 

[shortthroat]

Keep in mind that SD, or ES for that matter, is a usually measure of consistency for the powder's ignition, but it can also be affected by other factors in the rifle. Accuracy requires low ES and SD, but low values do not guarantee anything other than if they are large values, then the accuracy is likely to be poor. Accuracy is a big combination of firearm response to internal ballistics, as well as just powder ignition. It involves the chamber, brass, bullet performance in wind, throat, bullet jump, rifle, muzzle crown, etc. and especially the shooter skills. All these variables keep the powder manufacturers and gunsmiths in business, and most shooter's bank accounts small.

 

[BoydAllen]

Most shooters who speak of SDs could not tell you what one is, or how to calculate one if their life depended on it. The real reason that they are spoken of at all is that their cronographs automatically calculate it for any string of shots. The thing that is missing in the discussion is sample size. I don't know how many times someone has shot a string of 3-5 shots and then told me the SD. Right. Most of us should stick with ES because that is what our targets see. Divide the installed cost of one of your barrels by the number of accurate shots that you expect of it, and add the answer to the cost of a round's worth of powder, bullet, primer, and cost of a fraction of a case's life. Now tell me how many of those shots, how much of that investment you want to spend so that your sample sizes can be large enough to justify the use of SD.

 

[Ned Ludd]

I shot Hunter Class BR for several years. 308 Win, 150gr Sierra MK's. I also had access to a seed ware house from about 11pm to 5am one or two nights a week. Over 3100 rounds, enough to be statistically valid from one rifle, same lot of 4895, and MK's. Tested with a 36X Leupold, never had more than 14fps ES for 10-shot groups. Biggest group over that 3 years was .347".
SD did not match target results, the bell curve only exists, compared to my personal target results, in somebody's imagination.
SD is theory, recorded results are reality. IMHO, +/- one or more SD's is an cop out. It is always, in the data I have read, speculative, and relies on the quaint notion that one cannot maintain the initial accuracy or velocity spread level over time.

SD assumes (and we all know about ASS/U/ME) deterioration over time.

Not sure where you're getting some of this stuff, but it clearly shows you have a limited grasp of statistical analysis. If you really want to get the most out of Bryan's books, you should probably first do some research on statistical analysis; what it can and cannot tell you about real (not predicted) data, and how it can also be used to estimate future results. I'll give you a hint for starters - SD is relatively meaningless at short distances. Just how long was the seed warehouse? If you don't have a solid grasp on even the simple aspects of how statistical analysis can be used in ballistics and what its limitations are, you will not be getting everything you could out of Bryan's books.

 

[Mozella]

Here's something contemplate. It seems to me that consistency ought to be the key to accuracy (precision). It follows that shooting groups with low ES and low SD numbers should produce improved MOA and Mean Radius when the bullets impact on the target. It just makes sense, or at least it ought to make sense............ right?

After reading this thread, I thought it might be fun to look at a representative batch of data. I dug up the Excel spread sheet data for Berger 80gr VLD's shot from a .223 Shilen Match barrel at targets which were scanned and carefully measured using On Target Software. Each of the data points consist of five-shot groups at 100 yards. I have data from 560 rounds of the 80 VLD's shot from this particular barrel, but of course many of them are expected to exhibit diminished precision since during load development the highest and lowest charge weights, for example, often aren't the best. So I picked the best 175 rounds (35 groups) of this particular bullet and examined the data.

The average MOA is .334 and the average Mean Radius is .117, so while these aren't world record groups, they're not bad.

SD's ranged from a low of 5.07 to a high of 36.49.

This plot shows both MOA and Mean Radius of these 35 five-shot groups vs SD. If good groups are correlated with low SD's, then one would expect the trend line to rise from left to right. In fact, it decreases slightly. That means that the low SD values that we all try to achieve are no guarantee of precision when the five bullets get to the target.

In fact the smallest group had an MOA of .167 with an SD of 25.91. I'm happy to brag about any 5 shot group with an MOA in the "ones" but I'm certainly not proud of the corresponding SD.


I have lots more data on various bullets from several guns and when I study SD's for my 5 shot groups the results are essentially the same. That is to say I don't see any correlation between small groups and low SD values. Often the results show a tiny trend in the opposite direction; i.e. better groups with worse SD's. Nevertheless, I regularly place in the top three in 600yd F-Class and BR matches using loads identical to the ones which produced this and other sets of data.

I continue to strive for low SD's but unlike some, (especially those who measure their group size with a coin and keep their records by scribbling in a dog-eared note book and saving a pile of old targets) I don't put too much emphasis on SD when it comes to selecting a good load recipe.

Where I shoot, the trophy is given to the guy with the highest score or smallest group; not the guy with the best chronograph data.

 

[CharlieNC]

There are 10,000 posts regarding load development, harmonics, tuners, positive compensation, etc which are aimed at minimizing the effect of velocity variability on point of impact. And there are many factors which lead to inaccuracy. BUT if chrono variability does not matter then how are the effects of charge weight, case volume, etc manifested upon the bullet? Some proper balance of load tuning and uniformity is necessary.

 

[BoydAllen]

As I understand it, a reasonably low ES is required to minimize vertical when shooting long range, ,BUT that is not the same as being in tune. The two are related but not the same. For short range (don't ask me where the dividing line is on this.) ES is not a factor. We tune by group size, with almost no reference to velocity.

 

[IdahoSharpshooter]

So, you are telling me that SD is not a universal standard...? IE, it could not be taken to a court of law and proven.

We'll just call it a theory and everyone can define its' effect to suit their own reality. For me, the other major issue with SD is the transonic zone.

Warehouse shooting lane was 400yds long.

Rich

Note to qstaylorq: condescending attitude is a primary characteristic of narcissism. Next time, tell me how many ELR matches you have won this year. That will impress me.

 

[savagedasher]

SD I found out is seating depth . lol
Larry

 

[JPeelen]

SD (standard deviation) is simply a measure of consistency that is used worldwide in engineering. It can be applied to muzzle velocity, group size, variation in bullet mass, aiming errors, and many, many other things.
For example, group size, expressed as extreme spread (ES), grows a lot with the number of shots, although the underlying dispersion remains constant (same shooter, same ammo, tec.) SD is much less sensitive to the number of shots, because it takes into account all shots in the group, not just those 2 farthest apart. ES of a 5 shot group is on the average [!] a little more than 3 times SD.

A low SD of muzzle velocity cannot guarantee a tight group, because many other factors are at work. But good consistency of everything is important for tight groups. SD is simply one measure of consistency that has proven to be a most useful tool. In our days of spreadsheets and pocket calculators it is immediately available to everyone.

 

[Ned Ludd]

So, you are telling me that SD is not a universal standard...? IE, it could not be taken to a court of law and proven.

We'll just call it a theory and everyone can define its' effect to suit their own reality. For me, the other major issue with SD is the transonic zone.

Warehouse shooting lane was 400yds long.

Rich

Note to qstaylorq: condescending attitude is a primary characteristic of narcissism. Next time, tell me how many ELR matches you have won this year. That will impress me.

I would say you are the one with the poor attitude, and your response is simply further evidence of it. You started a thread entitled, "SD Question", then jumped all over Milanuk's response, who was merely trying to provide you with a helpful answer to your "question" and has probably forgotten more than you will ever know about statistical analysis. Why didn't you simply entitle the thread, "SD Rant", as that seems to be all you're interested in doing. If you're not interested in understanding what information statistics can provide to precision shooters, and merely wish to attack those trying to help you, that's your choice. Otherwise, I'll again suggest that it might benefit you to take the time to learn a little more about it.

 

[IdahoSharpshooter]

My opinion is based on the lack of proof positive, predictability, and repeatability. If I fire a thousand rounds of match ammunition, indoors again, the Bell Curve would suggest my results in the warehouse are invalid. IMHO, what makes SD attractive in theory are the nondeterminables of outdoor testing.

Proof, not a belief system. My limited testing tells me one thing that is "known", you are merely telling me what you think. You have college, or job experience to back your theory? Your second sentence supports my theory about narcissism. Memory loss is a terrible thing to have happen.

I attended Brian's seminar this spring, he only thought it was the best SWAG with the available data.

You have, in fact, not answered my question. Take a paragraph and prove the theory. I would like to know.

regards,

Rich

 

[CharlieNC]

Rich, you are shooting a 1000yd Fclass match. After 19 rounds you have shot all in the 10 ring, and the electronic scoring system indicates the velocity has ranged from 1625 to 1630. For that last shot to achieve a perfect score are you hoping to stay within that velocity range, or would a 1600 not bother you?

 

[BoydAllen]

SD (standard deviation) is simply a measure of consistency that is used worldwide in engineering. It can be applied to muzzle velocity, group size, variation in bullet mass, aiming errors, and many, many other things.
For example, group size, expressed as extreme spread (ES), grows a lot with the number of shots, although the underlying dispersion remains constant (same shooter, same ammo, tec.) SD is much less sensitive to the number of shots, because it takes into account all shots in the group, not just those 2 farthest apart. ES of a 5 shot group is on the average [!] a little more than 3 times SD.

A low SD of muzzle velocity cannot guarantee a tight group, because many other factors are at work. But good consistency of everything is important for tight groups. SD is simply one measure of consistency that has proven to be a most useful tool. In our days of spreadsheets and pocket calculators it is immediately available to everyone.

So....tell us about sample size as it relates to SD and percentage of confidence.

 

[CharlieNC]

So....tell us about sample size as it relates to SD and percentage of confidence.

Boyd I posted a lengthy discussion about SD and ES a few months ago, response was crickets. I believe few are inclined to invest the effort to understand such abstract topics when they could be having fun touching and feeling at the loading bench.

 

[BoydAllen]

Agree.

 

[IdahoSharpshooter]

1. If I were on the underground range (you can shoot over 2 miles, iirc) at Picatinny, I adjust a click. Outdoors, there are too many variables to gauge. Same hold I had on the first 19 and remind The Lord I have only missed Church twice this year. By time 1600 comes up, that bullet has already impacted the target. Pinwheel! That said, I see your point.
2. It would depend on who I was having a "fun touching and feeling at my reloading bench" scenario with.

Back to business. I thank both of you, Boyd Allen and CharlieNC very much for the shared information.

Rich

CharlieNC if you could post a link to that thread I would very much appreciate it.