the valid point is that in real life the small difference is less than noise in the real numbers we deal with.
we deal in .01 or .02 grains, NOT IN .00X.
we deal in .01 or .02 grains, NOT IN .00X.
Rifle Brass Sorting Experiment - Part One
- Thread starter Ammolytics
- Start date
I AM JUST TRYING TO POINT OUT that writing and publishing does not mean it has useful value. this is a shooting site. where are the targets that go with all those shots ? where is the relationship between charts and graphs and on target results ?.
the sample size is too small, there is not enough spread from end to end of the curve. we need both ends of the curve and the middle, not a bunch just in the middle. the problem with too small of a sample.
the sample size is too small, there is not enough spread from end to end of the curve. we need both ends of the curve and the middle, not a bunch just in the middle. the problem with too small of a sample.
Thank you very much for your research and sharing very valuable information!
I find real testing and research very refreshing. I like to share my opinion and experiences, but real data is not very plentiful. I appreciate your efforts to contribute in a way that provides real data. Is any test perfect, will it satisfy everyone hell no. But i still appreciate someones efforts!!!!
I stand by that statement. Maybe your definition of "guarantee" differs from mine? Outliers and exceptions will translate to fliers, which is what we're trying to avoid, isn't it?
It may just be a matter of deciding what level of consistency is right for you, actually! If your goal is case volumes that are consistently within +/- 0.01 gr H₂O for example, then weight-sorting may not be the right approach.
I appreciate the feedback, thank you!
Ned Ludd
Silver $$ Contributor
Certainly getting rid of as many outliers as possible is one goal of sorting brass. However, even minimizing the number of outliers, if not completely eliminating them, will still generally improve (reduce) variance about the mean.
In a typical scatter plot of case weight versus volume for which the best straight line has a negative slope, the very highest and lowest y values often are not at the same place in the plot (i.e. where high and low y values both have the exact same, or very close to the same x values). For that reason, you will typically find that extreme high/low volume outliers for their volume, for the most part do not reside in the exact same weight range. In other words, they can usually be segregated into different weight sub-groups by sorting, thus minimizing the overall variance within a single sorted sub-group. Of course, that depends in part on how good the linear correlation coefficient (R) for the data set is. The more frequent and extreme the outliers (points not directly on the trend line), the lower the correlation coefficient will be. If I'm very careful with the volume determination process, which is not a trivial exercise, I generally expect to see R values of ~ 0.70 or better for a typical brass prep. In that event, segregating brass into several distinct weight groups such as low, medium, and high, will usually result not only in smaller overall volume variance within each subgroup as compared to not sorting at all (i.e. lower ES), it will often also result in a reduction in the SD within each subgroup.
Below is an illustration of what what I'm describing with respect to how both ES and SD can be improved by wight sorting. I may have posted this particular illustration here before; it was made using a case volume versus case weight data set posted by someone else, which is the opposite of the way I usually graph out the data. Nonetheless, it still clearly shows a strong negatively sloped trend line and I merely used it to illustrate how sorting can substantially decrease the volume variance within a sorted weight sub-group of cases as opposed to doing no sorting at all.
In this example, sorting cases into three different weight groups would have resulted in a reduction of overall volume variance in any given sub-group by 30 to 50%, which is significant. What I was referring to above is that it is unusual to find both the very highest and lowest case volume samples falling within the same weight sort sub-group. Is it possible to find them in the same sub-group? Yes, it certainly is. However, I do these sorts of brass analyses and prepare graphs regularly and in my hands they rarely do, largely due to the negative slope of the trend line. The very highest weight outliers (or those with the lowest volume) tend to fall at one end of the line, the lightest weight/lowest volume outliers tend to fall at the other end. In the event the two highest/lowest outliers do fall within a single selected weight sub-group, it is still probable that weight-sorted cases within that single sub-group will have lower variance about the mean than the total sample of cases that have not been sorted at all.
Whether someone will directly and noticeably benefit from weight sorting cases in this manner can depend on a number of variables, such as what kind of shooting they're doing, the number of cases to be sorted, the average case volume of the cartridge they're using, and the overall effort involved. For disciplines that utilize relatively short strings of fire or that overall fire relatively small total numbers of rounds, sorting cases directly by water volume would probably be the better and more precise approach. In disciplines such as F-Class, where long strings of fire and large numbers of loaded rounds are the norm, sorting cases by weight can definitely make a difference and can be done much faster than direct volume measurement. The best news is that sorting cases by weight will never cause the volume variance within a single sub-group to become any larger than the total variance within the unsorted parental brass. In other words, it will never make them worse. Further, sorting cases by weight requires only a relatively minor effort.
In any event, I'll add my thanks to that of others above for starting this post. It's refreshing to see someone make the effort to post their results, especially on a topic such as sorting brass by weight, which often seems to elicit overly antagonistic and negative responses. Keep up the good work!
Last edited:
J F Johannes
AKA "RatRifle"
And that is why you just earned a spot on Jim's "IGNORE" list!
Jim
P.S. Thank Goodness for that site feature!!
This is excellent and insightful -- thank you for sharing!!
I'll be sure to reference this when I get through the next part of this series. As I stated in the Next Steps, I have 100 pieces each of Lapua, Peterson, and Hornady brass that I'll use to continue this work. Before that though, I'm pretty sure I'm due for a new barrel!
Dusty Stevens
Shiner
There are those who said Earth was the center of the solar system until scientists did actual testing to prove otherwise. I’m sure their initial sample size was only a few planets and not as many as they wanted, but that smaller sample size of planets gave them a roadmap to more testing and expanding to more planets as well as inspiring others to do their own testing.
The more I hang out and shoot with guys who actually test and use real data to load with, the more I’ve learned to test and use data myself. It’s inspiring to say the least, has improved my shooting 10x over and makes all the ‘emotional’ processes in loading seem almost crazy.
But I also don’t fault or fight those who load differently than me. I’ve seen guys do everything opposite of convention and still outshoot me and others, but I view shooting as a long term endeavor and am convinced that if viewed as an overall experience spanning years, the data based reloaders will prevail in the majority of matches. The more we know the more we improve.
Thanks to Eric for at least doing the testing and making us think.
The more I hang out and shoot with guys who actually test and use real data to load with, the more I’ve learned to test and use data myself. It’s inspiring to say the least, has improved my shooting 10x over and makes all the ‘emotional’ processes in loading seem almost crazy.
But I also don’t fault or fight those who load differently than me. I’ve seen guys do everything opposite of convention and still outshoot me and others, but I view shooting as a long term endeavor and am convinced that if viewed as an overall experience spanning years, the data based reloaders will prevail in the majority of matches. The more we know the more we improve.
Thanks to Eric for at least doing the testing and making us think.
This didn't tell me much but did give me a headache. I weighed these cases dry and with a water/alcohol mix on an FX scale as carefully as I could. All the cases were full length resized with the shoulders within .001", trimmed to length, and neck turned. The smallest volumes tend to be the heaviest but not always. All these have been fired, and are from several lots. Hornady small/large means the stamping was a different side and were not stamped MATCH. This would be more meaningful if it were all from the same lot number. I plan to load and test fire these keeping each batch within 1% of the same volume
I once got curious and did my own experiment. I measured 100 pieces of 5.56 lake city brass after sizing, neck turning and uniforming pockets. I then weighed each piece on an A&D scale empty and after using water/alcohol with the kit that plugs the flash hole. Below are my results. Each cell represents a case, top rows are the weight of each empty case (rounded to nearest .1) and bottom rows are the volumes of the same cases. Was this scientifcally done? no. Did it show some correlation between wieght and vloume? I think it shows that the outliers on either side definitely have more/less volume. Did it make any difference on target after sorting out the cases to keep from 91.5 to 92.6 and use the rest as siters? That remains to be seen, my shooting isnt good enough to tell the difference, yet.
A couple of things to consider is the precision of the measuring devices used. These uncertainties build up so that you can lose precision pretty rapidly. In the case of weighing a case on a scale with resolution (display) of .02 grain you have an uncertainty of at least .01 grains. A reading of 29.00 grains might be anywhere between 28.99 and 29.01, perhaps more. Generally speaking, if I can resolve only .02 grains I’d tend to drop the 4th digit as less than significant. Or at least go to the nearest .05.
This is the crux of uncertainty.
now couple that with human interface and visual resolution, the uncertainties begin to add up. Things like, how much meniscus is present? Did a micro drop spill? Is the wind affecting the scale reading? Has it drifted, electronically?
So the measuring method and the inherent uncertainties have to be accounted.
This is the crux of uncertainty.
now couple that with human interface and visual resolution, the uncertainties begin to add up. Things like, how much meniscus is present? Did a micro drop spill? Is the wind affecting the scale reading? Has it drifted, electronically?
So the measuring method and the inherent uncertainties have to be accounted.
Well don't associate your 'facts' with truths, because truths actually pass all tests.
Rick300
Gold $$ Contributor
I agree! He recently joined and has been nothing but negative the whole time.
Ammolytics - Ignore him. I am very glad you went to all the trouble you did trying to provide useful info. THANK YOU!
Gentlemen, if you don't have something constructive to contribute to the subject in this thread, please refrain from posting.
The originator of this post has made a substantial contribution of time, money and effort to produce some analytical information for us. Whether you agree or not, let's look at it as another resource to ad to our war bag.
Otherwise, consider your comments, would you make them face to face? If not, please refrain from non constructive comments.
The originator of this post has made a substantial contribution of time, money and effort to produce some analytical information for us. Whether you agree or not, let's look at it as another resource to ad to our war bag.
Otherwise, consider your comments, would you make them face to face? If not, please refrain from non constructive comments.
Last edited:
ronemus
Gold $$ Contributor
I did an experiment several years ago to determine just how much effect brass weight has on .223 loads. The effect is small but measurable. Under the right conditions the measurement is straightforward.
I used WW brass (sized, trimmed and deburred, primer pockets uniformed, flash holes deburred, and neck turned) , WSR primers, charges of RL-15 or N-550 powder weighed to 0.1 gr, and 75 gr A-Max bullets.
Using the lightest and heaviest cases (sorted from 1000 once-fired I had on hand), I had two lots of 10 cases with a 3 gr difference in weight. The average muzzle velocity difference was 16 fps, just a bit more than the 12 fps due to 0.1 gr of powder.
I choose to sort 0.5 gr lots of .223 brass for my long range loads, but the effect will only matter at 800-1000 yards - the vertical displacement on the target from such a small velocity change is negligible at shorter distances. Unless you control all other sources of variation, the effect of brass weight is negligible.
I also shoot .284, and because the brass is twice as heavy (and twice the internal volume) I batch in 1 gr lots.
Upgrades & Donations
This Forum's expenses are primarily paid by member contributions. You can upgrade your Forum membership in seconds. Gold and Silver members get unlimited FREE classifieds for one year. Gold members can upload custom avatars.
Click Upgrade Membership Button ABOVE to get Gold or Silver Status.
You can also donate any amount, large or small, with the button below. Include your Forum Name in the PayPal Notes field.
To DONATE by CHECK, or make a recurring donation, CLICK HERE to learn how.
Click Upgrade Membership Button ABOVE to get Gold or Silver Status.
You can also donate any amount, large or small, with the button below. Include your Forum Name in the PayPal Notes field.
To DONATE by CHECK, or make a recurring donation, CLICK HERE to learn how.