Was wondering what is a reasonable figure for standard deviation? My average velocity matched that in the book almost perfectly, but the standard deviation seemed a little high around 40ish. Am I OK with these stats. Tahnks for any help.

7mm. guy

How do they shoot is the real question? Quite a few guys on this forum have gotten very accurate groups nothwithstanding a std dev that is higher than other loads that are less accurate. Let the group size be your guide. If the load is very accurate consistently, don't worry about the std dev. Velocity is only one factor in the accuracy of the load.

LWD

Guy

LWD gave you the correct answer.

The very best SD is the one that shoots teeny tiny groups.

Ray

I have a serious question about standard deviation .
Why worry about it ?. I know what it's for . I know that serious bench guys are driven for it .

I have had my best groupings ( # 3 - # 5 Shot averages when my STD. Deviation is all over the place . 65 - 80 !. YEA . I've also had them down around 9 -17 and groups went to shit !.?.

Go figure , I've never been able to figure that out it goes against all reason !.

Shoot Straight Know Your target . ...

In my experience I have found low standard deviation to go hand in hand with accuracy aggregate. I prefer a single digit SD but will answer your question by saying under 20 fps.

Let me temper my initial comment a little bit.

If you have a choice between a large std dev and a small one, always take the small one all else being equal (which is the tricky part).

At longer ranges, variations in velocity do become important. Run some numbers on a ballistics calculator to see the difference in drop at 1000 yards a 100 fps change makes. That .2" at 100 yards gets magnified a bit.

All that said, for hunting, plinking and many other things it's more statistical than anything.

LWD

quote:

will answer your question by saying under 20 fps.

Good reply!

http://www.shootingchrony.com/manual_F1M1.htm#formula
here is alittle more information

Standard deviation matters more the further you shoot. Standard deviations of 40 fps may not make any noticeable difference in 100 yard groups. Shoot that same SD at 400 yards and you'll see the vertical stringing.

IMHO This is probably one of the most common reasons (second only to poor shooting) that people's loads groups at 1" at 100 yards but at 300 yards they're shooting 6-8 inch groups.

At 1,000 yards you will be talking about grouping in feet instead of inches.


Standard deviation is the square root square root of the variance.

So a Standard Deviation of 40fps means that the extreme spread of the velocity was 1,600 fps! That is a huge difference in bullet speed and it will affect the grouping ability of the load.

quote:

So a Standard Deviation of 40fps means that the extreme spread of the velocity was 1,600 fps! That is a huge difference in bullet speed and it will affect the grouping ability of the load.

Actually the theoretical extreme spread of a standard deviation of 40 is 6.2 times 40 or about 250'/sec.

quote:

Originally posted by LWD:
How do they shoot is the real question? Quite a few guys on this forum have gotten very accurate groups nothwithstanding a std dev that is higher than other loads that are less accurate. Let the group size be your guide. If the load is very accurate consistently, don't worry about the std dev. Velocity is only one factor in the accuracy of the load.

LWD

+1

out of 15 powders tested by me for a 6mm rem, the one that shot the best ( one hole groups at 100 yds) was the powder with the worst SD of all tested, and that was H 414... Deviation was as high as 150 fps...

and on targets it still hangs tight right out to 500 meters...

so I'll just have to live with it I guess..

Sea,

The way I read your post I think you're saying your "Extreme Spread" was 150 fps? Which translates into a small "Standard Deviation".

quote:

Originally posted by Buliwyf:
Sea,

The way I read your post I think you're saying your "Extreme Spread" was 150 fps? Which translates into a small "Standard Deviation".

Seafire reported an "OBSERVED" deviation in velocity of 150'/sec and this is quite a large SD depending on how many shots were taken.

One might have to shoot 100 rounds under a chronograph in order to actually "OBSERVE" the calculated SD.....and as a matter of fact probably wouldn' then as the SD assumes infinite number of shots.

If one gets ten shots over the chrony and the device reports a SD of (for example) 25, then it's a fair chance he has observed 1/3rd the actual calculated variation of velocity.

For observations of about ten to twenty rounds one will likely never come close to observing anything close to the calculated SD.

The theoretical velocity range will always be 6.2 times the SD!

It is likely that seafire's actual SD was closer to 75 depending on the number of observations he had......indeed a very poor SD!

quote:

Originally posted by LWD:
...At longer ranges, variations in velocity do become important. ...

quote:

Originally posted by Fjold:
...Standard deviation matters more the further you shoot. ...

Gotta agree with those thoughts.

The nice thing about it is you still don't have to use a chronograph to see the effects of a High SD on the paper at distance. But, a chronograph can help determine what you should be focusing your attention on - reducing the variability in the Load.

If you do see the High SD on a chronograph you can be sure you will have l-o-n-g distance accuracy problems.
----

If your Hunting is inside 300yds - just go by the groups and don't waste "Brain Time" thinking about it - leave the chronograph at home and take the batteries out of it.

Best of luck to all you folks.

Vapo,

Briefly, without asking for a Phd in statistics, where does the 6.2 multiplier come from?

Also, are you saying 1/3 of the shots will be wothout chronogaph error; or, are you saying 1/3 of the shots will be within one standard deviation?

I believe in uniformity for accuracy and see SD has a measurement of velocity uniformity so I want to be sure I understand your comments.

OK guys, I need to be educated. I'm just an ignorant farm boy but I thought that the SD was the "mean of the mean" or how closely distributed the individuals were around the average. So the SD can never be bigger than the ES. Roughly, the square root of the variance.

Where did I go wrong?

And if you are working with an average 1000 yard cartridge, a velocity spread of 100 fps results in a vertical of only a little more than 1 MOA, not the feet or yards that some would suggest.

But no long range shooter worthy of the name would compete with a cartridge that showed an ES of 100fps, unless it grouped like the dickens. I know cause I R 1.

Pay attention ONLY to how close together the holes are in the target, not how close together the numbers are.

Ray

quote:

where does the 6.2 multiplier come from?

The mathamatics of statistics is set up such that when calculated the entire population is represented by +/- 3.1 standard deviations. (It's a complicated mathamatical proof)

Any sample will yield a standard deviation that can be multiplied then by +/-3.1 (or 6.2) to get the total population.....even though you have only taken a sample!

quote:

Also, are you saying 1/3 of the shots will be wothout chronogaph error; or, are you saying 1/3 of the shots will be within one standard deviation?

About 67% of the entire population will be found in the first standard deviation (+/- 1)....this is where we see the most results! So about 2/3 of all the witnessed events fall in the first SD!

+/- 2 SD yields about 95 % of the population and it's rare to see events outside this.

99.5% of the events are within +/- 3 standard deviations and while we never reach 100% because we can go into 20 standard deviations if we want with diminishing returns....it's mathamatically conceded that for practical purposes +/- 3.1 SD represents the entire population.

If you've ever seen a "normal distribution" curve (bell shaped curve) it is represented by the standard deviation which is a measure of the width of the curve. The larger the SD....the wider the bell shaped curve.

The SD should not be too much of a concern as stated by others except that smaller is (usually) better.....but it seems groups are not necessarily related to SD.....

Ok. The 6.2 multiplier addresses the 3.1 standard deviations above the normal curve plus the 3.1 standard deviations below the normal curve (3.1 + 3.1 = 6.2 standard deviations)which encompases 99.5% of the events.

Thanks for your response.

quote:

Originally posted by Cheechako:
I thought that the SD was the "mean of the mean" or how closely distributed the individuals were around the average.
Ray

You are correct here in that the SD is a measure of how closely the individuals are about the mean! (average)

quote:

So the SD can never be bigger than the ES. Roughly, the square root of the variance.

Where did I go wrong?

When we speak of ES (extreme spread) we are referring to +/- 3.1 SD.....definitely not the observed spread we saw when we shot groups over the chrony......and yes.....the ES is absolutely always larger than the SD.....(by a factor of 6.2)

As to "variance".....(It's been since 1977 that I took these courses) IIRC it's the square root of the SD...or is it the square?....but it's being misused as it's not the same as ES and I believe we're trying to assign ES to variance!!!

Lets forget this value for the shooting sports!

For my purposes, I do not use SD at all. I look first at ES, an entirely different thing.

Then I calculate the average variance of each shot from the mean of the shots actually fired. Again, that is not the same thing as the SD as I have seen its calculation explained.

But ultimately, when tossed in with such things as the mean harmonics of both barrel vibration and bullet flight, muzzle "departure angle", and a zillion other things, I find both the easiest and and most meaningful indicator to be group size at the actual ranges I intend to use the load.

I save ALL MY TARGETS and use their average size as a very rough guide as to what I MIGHT luck into for a future average group size.

Of course, even there I am misleading myself to some extent. Its exactly like the fella who shoots a .019" group with his off the rack .338 Win Mag. He cuts the group out, puts it in his wallet, and forever brags that his rifle will shoot like that "if he does his part".

Yes it will, but he never knows how many thousand, or million, groups he'll have to fire to see it do it again. Every other rifle might, at some time, for one group, too.

Statisics, after all, are hypothetical expressions of predictive values, based on an infinite number of shots during controlled (or at least "known") variables.

Although in theory, a perfectly balanced coin tossed precisely the same forever will result in 50% heads and 50% tails, in actual fact it could very well land heads for 210,000 consecutive tosses. Likewise, there's always the possibility that 5 consecutive shots, or even 10, may all go through exactly the same hole.

If we shoot enough groups, we can calculate how accurately a rifle has shot so far, and therefore have an educated guess as to how it may shoot the same loads in the future, on averqe. But we still don't know about the NEXT group. IT MAY WELL BE THE WORST GROUP THE RIFLE WILL EVER FIRE, IN AN INFINITE NUMBER OF GROUPS SHOT.

"Variance" is the "lack of uniformity" which is another view of "Standard Deviation" which is "velocity uniformity".

"Extreme Spread" is the highest velocity minus the lowest velocity.

I don't take my eyes off SD in load development because I believe low SD keeps the odds on my side of developing the load I'm looking for. I have never experienced the high SD / excellent group scenario. When I see high SD and reasonable groups I look at the primer for ignition uniformity thinking I have a good average velocity for my barrel. I look for aggregate or in other words repeatability of the load. High SD means the rifleman is starting each bullet on a different trajectory which puts the odds against shooting repeatable consistent groups.