SD and ES inconsistencies
- Thread starter Gd33333
- Start date
I do have a bore scope so I will take a look for any carbon buildup. As far as the velocity node, I graph out 5-6 five shot groups shooting round robin looking for velocity plateaus only. Not worrying about small groups. Will tighten groups with seating depth later.
I’m shooting sub 1/2moa consistently which has me scratching my head on the chronograph. Will definitely be looking for a radar to compare.
As BuildingConceptsIIc says, after cases are fire formed to your chamber (takes ~2 firings to really get that), the cartridges do become more consistent and gives better results. And at that point, it's actually good not to remove the powder residue from the inside of the necks as it acts as a lube.
The neck wall thickness will vary on a single case. . . like it's not unusual to see .0015 and on poorer cases >.002". You can see this with a caliper as you turn the neck while measuring with the caliper. So, you can just see what the average of the change on a given case and to with that number. Better quality cases, like Lapua, have some variation, but it's not much over .001". That issue is only important if the variation is large as the smaller variations won't make any noticeable difference, especially out of an RPR.
I have to RPR's, one is .308 and the other is 6.5 PRC, and though I use Lapua brass that have small variations, I'd just like to make my brass as perfect as possible. Therefore, I turn my necks to make them without such variations.
"still hits within a half MOA 69% of the time"
That's why I still get 9's
I don't play that game, but isn't the 10 ring 1 MOA with a .5 MOA X ring? That would mean 20-14X, roughly?
I put a .22 group SD (about a half MOA gun) in the calc.In a tunnel, with a Laser
I'll see myself out.I don’t have a neck turning device but have contemplated getting one.The neck wall thickness will vary on a single case. . . like it's not unusual to see .0015 and on poorer cases >.002". You can see this with a caliper as you turn the neck while measuring with the caliper. So, you can just see what the average of the change on a given case and to with that number. Better quality cases, like Lapua, have some variation, but it's not much over .001". That issue is only important if the variation is large as the smaller variations won't make any noticeable difference, especially out of an RPR.
I have to RPR's, one is .308 and the other is 6.5 PRC, and though I use Lapua brass that have small variations, I'd just like to make my brass as perfect as possible. Therefore, I turn my necks to make them without such variations.
To drive this point home, I was out at the range today validating some loads before heading out to the hills this weekend. One such load was for my 300, which was still gaining velocity during my last trip, which took it out of the nice zone for SDs. I backed the load off a bit to get the speed back in "the zone" - which is in the ~2900 fps range. I loaded 10 rounds, ditched the cold bore shot from the group, and got the following:
Twice-fired
2897.8
2904.5
2902.8
2904.6
2898.4
2897.6
2899.8
2904.5
2904.0
Average 2901.56 fps
Std Dev 2.9 fps
I also loaded 6 rounds with the same load in virgin brass:
Virgin
2916.0
2910.6
2887.1
2921.1
2917.0
2905.3
Average 2909.5 fps
Std Dev 11.2 fps
Granted, the sample sizes are small, but there is a clear and substantial difference between the two.
^^^ THIS IS THE ESSENCE of small sample sizes! Things that look different aren't!
we're going to see a lot more of this now that guys are measuring every shot with the Garmin.
These small samples, selectively reported, showing SD of 3.4 over 5 shots are going to magically disappear like unicorn semen
we're going to see a lot more of this now that guys are measuring every shot with the Garmin.
These small samples, selectively reported, showing SD of 3.4 over 5 shots are going to magically disappear like unicorn semen
Sort of, but not really. When you have a substantial difference, even in a small sample size, the chances of two matching up with each other after a statistically valid sample size is run diminish over time.
As an example, for the two groups of rounds to match up at, say, an SD of 8 after 20 shots, the next 11 shots in the twice-fired group would have to drop from the 2.9 over the first nine shots to somewhere around 10 over the next 11 shots. Possible, but unlikely. For the virgin brass to match up at that same 8 SD over the next 14 shots, it would have to magically go from 11.2 over 6 shots to somewhere in the neighborhood of 4 for the next 14. Exceedingly unlikely.
I cannot from the data above tell you that twice-fired brass is exactly 11.2/2.9 = 3.86x better SDs than virgin - but I can tell you with good confidence that one set is measurably better than the other. You don't need sample size = 30 to tell you that.
yes, if P< alpha, you reject null, even if the sample size is way underpowered. However, if you're a Bayesian, the posterior probability is lower if you're underpowered (small sample size).
The problem is this does not account for selective reporting, which is a practical/philosophical matter more than a statistical matter. Discussed here: https://www.cambridge.org/core/book...vere-testing/D9DF409EF568090F3F60407FF2B973B2
Small samples, in my estimation, are not "severe tests" of the null hypothesis and are too easy to selectively report.
If he repeats them again, with larger sample and the results are replicable, we will have the bonus of a narrower confidence interval for the true difference, too.
The problem is this does not account for selective reporting, which is a practical/philosophical matter more than a statistical matter. Discussed here: https://www.cambridge.org/core/book...vere-testing/D9DF409EF568090F3F60407FF2B973B2
Small samples, in my estimation, are not "severe tests" of the null hypothesis and are too easy to selectively report.
If he repeats them again, with larger sample and the results are replicable, we will have the bonus of a narrower confidence interval for the true difference, too.
Last edited:
Update: I spoke with Caldwell regarding my chronograph and was told the erratic velocities may be because my muzzle was too close to the chrono (~10ft). I ran to the range today and tried 15ft and even 20ft. The numbers were worse than before. This group pictured is the once fired SL brass with small magnum primers. I tried the same load with once fired brass and non-magnum primers and only took one shot as the firing pin damn near poked through the primer. I also tried a box of Hornady Precision Hunter as suggested to see if my loading was the problem. Same results. I seen velocities from 3700 down to 2100. With those numbers and this group at 100, I think it’s fairly safe to say it’s the chronograph. I haven’t been able to locate a lab radar or Garmin to verify yet. May have to just bite the bullet and get one. Thanks everyone for your input. I appreciate your time.
I'd say the chrono is definitely a problem if you got those kind of readings
I wouldn’t hump a labgaydar with your dick, if building concepts were pushing
Just get a Garmin and never look back. Maybe a used magnet to speed, but not a labgaydar.
Just get a Garmin and never look back. Maybe a used magnet to speed, but not a labgaydar.
Yeah , there's really no point in even messing with it without a good chrono. There's plenty of magneto speeds available used for cheap. With a mount (though I modified my mk mount to make it a lot better), they are extremely reliable.
That said, the Garmin makes it all SOOOOOOO much easier and better. It's probably the best $600 I've ever spent. Check out active junky and then buy from Cabelas or bass pro and get either 5 or 8% more off. Definitely worth the money
Update: so I purchased a MagneetoSpeed V3 and ran the same loads through. Ended with a SD of 2.4 with a ES of 7. What a waste of time and money a cheap chronograph can cause. Thanks for everyone’s input and help on the matter.
