Originally Posted by Rman
Originally Posted by Jordan Smith
SD numbers that most people quote are based on a statistically invalid number of sample shots anyway, so unless you're shooting 50 rounds over the chrono to get a SD value, it's easy for a guy to think he's shooting the load with the lowest SD, but if he ran a few more shots over the chrono he may be surprised when the SD goes up.



Wouldn't that all depend on the original lot of ammunition loaded?
Most don't load 1000's of rounds at a time.
So, out of a batch of say 200, what would an appropriate sample be?
I am thinking less than 50...

R.


It does definitely depend on the number of rounds in the population, meaning that if you plan on re-developing your load and re-testing for SD/ES for every batch of 200 rounds of ammo you load, then yes the calcs would be based on a population of 200. But if you develop a load and then produce 1000 rounds of it in batches of 200, then you're working with a population of 1000, not 200.

When you calculate the required sample size to have a 95% confidence interval and a sub-10 SD, it's a lot larger than most people think. Suffice it to say that a sample of 10 or even 20 shots is fairly meaningless when it comes to getting an accurate SD number.