Rounds to achieve a reasonably certain SD in load testing

[Choid]

I'm looking to finalize some loads I am working on. What is the general feeling here as to the minimum number of rounds necessary to establish a good standard deviation measure for comparison of two charges.

 

[spife7980]

Slippery slope as more is always better. I say 10 suffices for my intents and purposes.

 

[Sheldon N]

I get by fine with shooting 5 for SD evaluation. Shooting 10 rounds is not likely to change your SD by more than a FPS or two. You will see your ES range widen the more shots you take.

 

[Sig Marine]

I typically do 10, good enough for shooting steel and the occasional F-Class match for fun.

 

[JCCinOhio]

I believe my statistics teacher would tell you 30... but 10 is prob close enough

Edit.. Just found this... https://guns.ptosis.ch/guns/node/41 it's for group sizes, but should translate..

 

[_Raining]

IIRC the lower your SD, the less amount of rounds you need to not have overlapping confidence intervals (Somewhere in the 3-4-5 area you need much less than the 8-9-10 area).
SD 10 @ 30 shots, 90% confidence interval = 8.25 to 12.80
SD 8 @ 30 shots, 90% confidence interval = 6.60 to 10.24
So if you shoot 30 rounds of 2 different loads and the SD comes out to 8 and 10, you can't be sure that the SD of 8 is actually better than the 10. But there are people who shoot 5 rounds and pick the lowest SD of the bunch. Now this isn't a big deal because 8 and 10 are both good SD's so even if you pick the wrong one, you are still in good shape. Same goes for group size as JCCinOhio mentioned. It's just a little odd how much confidence people have when they say "I have found my load" and post a picture of like 10 groups of 5 shots each evenly distributed between .5 MOA and .6 MOA (the .5 MOA being "the one"). But hey, you do you... I ain't about to tell you how to live your life.

 

[Marksman10x]

Fifteen shots has always been my go to for numbers testing.