Standard Deviation...?

[Sir Rollo]

I've been using the Chrony, and finding differences in velocity with my handloads. Now this past weekend I made a few bullets to try, and was careful to select brass based on number of loadings, and tried to keep each step consistent. I still ended up with differences in velocity, (2806, to 2863). The rest were in the middle, with a mean of 2829 ft/s. Is that good results? What is a good "standard deviation" for a 10 shot string? What does it even mean, I looked it up and can't really figure it out, sorry to be dim.

[kayaker]

SD is essentialy the average of the deviations from your mean. So its the amount that any deviation from the average is most likely to be...in a nutshell

[sdeviation]

i have a few rifles that SD under 10fps
and they do shoot great....
SD is important if u where making thousands and thousand of rounds of ammo and u need em to fall within a certain tolerance ,,,they use the SD in conjuction with a bell curve...
but for a shooter what more important is the xtreme spread ...which is the distance (fps) between the lowest and the highest shot ,,, this can sumtimes (most times)contribute to vertical in the shot string .....

[WLT]

Standard deviation (SD) is the term used to describe where 95% of your data/measurements fall on either side of your mean or average. If most of your measurements fall close to the mean (+ or -), the amount of variability in your data is small - which is what you want. Averages or means do not really mean anything unless you have some measurement of variability, which the SD does measure.

[gitrdun]

I've had some shooting sessions with not so great SD and good groups. However, if you want to improve that a bit weight and classify your brass.

[greylynx]

I would not base any statistical analysis on a sample of ten.

Unless you shoot hundreds of rounds under identical conditions, don't worry about standard deviation.

[kayaker]

Originally Posted by WLT

Standard deviation (SD) is the term used to describe where 95% of your data/measurements fall on either side of your mean or average. If most of your measurements fall close to the mean (+ or -), the amount of variability in your data is small - which is what you want. Averages or means do not really mean anything unless you have some measurement of variability, which the SD does measure.

Much better description than my botched job..

[Erik]

Originally Posted by WLT

Standard deviation (SD) is the term used to describe where 95% of your data/measurements fall on either side of your mean or average. If most of your measurements fall close to the mean (+ or -), the amount of variability in your data is small - which is what you want. Averages or means do not really mean anything unless you have some measurement of variability, which the SD does measure.

Your data will fall within plus or minus two standard deviations 95% of the time. This assumes a normal distribution (aka a bell curve with the mean in the middle of the bell). The amount of data that should fall within plus or minus one standard deviation should be roughly 68%.

Originally Posted by greylynx

I would not base any statistical analysis on a sample of ten.

Unless you shoot hundreds of rounds under identical conditions, don't worry about standard deviation.

This is true. Statistical samples can be misleading where sample sizes are less than 30.

[Sir Rollo]

Thanks, I was worried about getting all the shots as close to the same as possible. Extreme spread of 50 fps is about normal then, for a 2800 fps round?
I chronied some 7.62 x 39 mil surplus stuff, and the extreme spread was around 10, amazingly consistent; from an SKS no less!