Zaydok Allen
Member
- Joined
- Feb 12, 2011
- Messages
- 13,274
This past weekend I got to use my Labradar for the first time. It worked perfectly and I found it very intuitive with a quick skim of the directions.
Frequently I see 5 shot groups reported online with averages, highs, lows, extreme spreads, and standard deviations.
While any sample is better than no sample, I question the 5 shot data set. I'm no mathematician or statistician, but interpreting stats is part of my job, and I'm somewhat familiar.
My point is that a 5 shot series is a very small sample. You can infer a lot from SD, ES, and average but why not 25 shot series? I realize that when looking at the data the variance must be quite small in the grand scheme, so maybe 5 shots are as good as 25 shots because they just don't fluctuate enough to be relevant.
I tend to load in batches of 50 so I'm thinking about going bigger on my number of shots in a series when developing loads.
Just curious what others think and if anyone has a reference on this. Should I just assume industry testers have checked this and found 5 shot series to be statistically sound due to low variance, and collecting larger data sets to be a waste of time?
Frequently I see 5 shot groups reported online with averages, highs, lows, extreme spreads, and standard deviations.
While any sample is better than no sample, I question the 5 shot data set. I'm no mathematician or statistician, but interpreting stats is part of my job, and I'm somewhat familiar.
My point is that a 5 shot series is a very small sample. You can infer a lot from SD, ES, and average but why not 25 shot series? I realize that when looking at the data the variance must be quite small in the grand scheme, so maybe 5 shots are as good as 25 shots because they just don't fluctuate enough to be relevant.
I tend to load in batches of 50 so I'm thinking about going bigger on my number of shots in a series when developing loads.
Just curious what others think and if anyone has a reference on this. Should I just assume industry testers have checked this and found 5 shot series to be statistically sound due to low variance, and collecting larger data sets to be a waste of time?