Range Report: Primer test

[Jim]

In my last report, I was not able to produce any acceptable groups in the SLYM with Bullseye or Unique. I ran a primer test to see if that made any difference. It most certainly did and I was surprised at the results.

[frnkeore]

Are these groups also 223? What brand of primers did you use?

Frank

[waksupi]

I found a big difference in several firearms when changing the primer. Gotta try EVERYTHING!

[Jim]

Frank, yes, 223. Primers are CCI.

Rick, yes sir, can't lock anything down until you've been through every possible variable.

Testing continues.....

[Larry Gibson]

Based on 5 shot groups a repeat of the same test may very well result in different group sizes. It takes 7 shots for minimal assurance and the industry standard is 10 shots. What is called "random dispersion" can easily skew the results if less than that is used. Yes I know "everyone" uses 5 shot or less groups for accuracy or strings for velocity but the results from an insufficient sample size can be very misleading.

Larry Gibson

[atr]

thanks for your posting this information...enlightning to say the least....
I am now beginning to wonder if changing primers in my 22 Hornet loads will get them more consistant.

[Jim]

Based on 5 shot groups a repeat of the same test may very well result in different group sizes. It takes 7 shots for minimal assurance and the industry standard is 10 shots. What is called "random dispersion" can easily skew the results if less than that is used. Yes I know "everyone" uses 5 shot or less groups for accuracy or strings for velocity but the results from an insufficient sample size can be very misleading.

Larry Gibson

I agree, Larry. However, at the rate I'm testing, I'd be going through a lot of primers if I went to even 7 shot groups. I cut my group shot count to 5 to economize.

I start by shooting 5 shots per load, 5 loads to the grain. E.G.: 3.2, 3.4, 3.6, 3.8 & 4.0.
I find the best group in that test and then load 5 shots in 3 groups surrounding the best group. E.G.: If 3.6 yeilded the best group, I would load 5 shots at 3.5, 3.6 & 3.7. I would then load 10 shots of the best group and shoot that for the final test.


So, in one complete test, I burn 50 primers. Sometimes I shoot 3 or 4 tests at a time. That can get expensive.

Atr, I go through 3 to 5 different powders and bullets and 4 different primers. I try using dacron as a locator and without. I change lubes. I've been doing this since the late summer of 2011 and there's no end in sight.

[WILCO]

I've been doing this since the late summer of 2011 and there's no end in sight.

That's the truth Jim. Thanks for sharing your work.

[Larry Gibson]

Jim

Your additional testing is spot on and I understand your method now as I use something similar. It works for me. Well done.

Larry Gibson

[bangerjim]

........the site is doing double/triple posts......sorry

banger

[bangerjim]

Jim.....good info. I agree it would almost cost prohibitive to do 10 test each. Mabe those that insist on industry standards will send you some money! Unless your money tree is thriving this time of year.

Keep up the reports.

bangerjim

[Larry Gibson]

Bangerjim

Might do the math. Using jim's example test he uses 65 primers. Using "industry standard" of 10 shots with the first 5 test strings would have given him the answer using 50 primers. Would have saved 15 primers. However, to be sure another 10 shot confirmation group or 3 five shot groups would be called for so it comes out fairly even.

Works for me.

Larry Gibson

[bangerjim]

I am usually happy with 5 on my shoots for testing.

Heck......I'm jes happy to hit the target!!!!!! HA....ha

banger

[Larry Gibson]

Yes many do use 5 shot groups. Let me ask this; have you ever worked up a load with 5 shot groups and found "the one" that puts all 5 shots in a nice small little group? Eureka you yell and rush home and load up another 50 of them only to find the next 5 group runs twice as large....so you try again and darned you shoot up all 50 with ten 5 shot groups of which only one comes even close to the "eureka mother load" group you found when working up the load.........

So let's say the "eureka" group was 3/4", the closest to it was just under an inch and the largest was right at 2". You could make yourself feel better and say; "not so bad, the average was 1 3/8" (that's a hypothetical average)....not so bad at all....." But you would just be fooling yourself again as the "average simply means that half the shot will be that good but it also means half the shots will be worse. The real accuracy capability of that ammo in that rifle with you shooting it is 2", the largest group shot. Not what most of us want to believe but that is the truth.

Ten shot groups give you statistical assurance that the accuracy will be what you see. Five shot groups do not. Shoot a 10 shot group sometime and then see just how many different 5 shot groups you can get out of that single 10 shot group. Each of those groups (good and bad) are examples of random selection/dispersion. Had you selected those 5 shots for a 5 shot group that is what it would look like.

Probably why you see most "tests" in gun rags are three or five 5 shot groups. They do that to get a meaningful sample for statistical assurance that the groups sizes are representative. Unfortunately they still want to use the "average" group size as the measure of accuracy which it really isn't. You, the rifleman, are only as accurate as the widest shot that may be fired.

Larry Gibson

[popper]

I agree with Jim & Larry but my method is different. I use CCI LR, a few powders & a mould for each rifle. My goal is repeatable MOA for each. We talk statistical curves and numbers but that doesn't work. Use the average or mean or sigma of a 'normal' curve and you will allow gross errant shots. A sample of ten is not large enough to make a curve. To work up a load, yes, to verify a load, no. Are you happy that the shot on the monster buck goes 6" the wrong way?