LBT 309-200 mold cast with linotype. Weight varied from 190.0 grains to 191.9 grains. This is 1% weight variation. Bullets are sorted into 0 .1 grain lots. How do I set up a test to see if there is a difference on the target? Do I shoot all 190.0 grain bullets, then 190.1, etc? Or do I mix and match weights? We discuss this quite a bit, but I would like to find something that eases my mind.
Bullet Weight Variation and Testing for Accuracy Variation
- 666 Views
- Last Post 30 September 2024
Would help to know the distribution, but here's what I would do.
Pick the 25 heaviest bullets separating into groups of 5, all the bullets in each weighing the same.
Do the same with the 25 lightest bullets.
As you load each group of 5 put it in a ziplock bag. On a slip of paper write the weight of the bullets in that bag, fold it over and staple it shut then put it in the bag. Mix them up so you don't know which groups are heavy and which are light when you shoot them.
At the range pull out one bag of bullets and shoot that group. Write the group number on the slip of paper. Repeat.
When done shooting open each slip of paper and write the weight of that group of bullets on the corresponding target. You can then measure and sort the groups to determine if there is a relevant difference.
John Carlson. CBA Director of Military Competition.
Attached Files
I would do a combination of What John and Pat suggest. Blind random groups plus several control groups of mixed weight bullets. Find a method that you are not aware of what batch you are shooting at the bench.
Hopefully you will learn three things. Does weight matter? Do heavier bullets shoot better or to a different POI v POA? Is there a clear difference with weight sorting?
Attached Files
At some point we have to say "enough is enough" and go with a load. Otherwise, we would spend all our time and components "testing".
Again, I don't think it is the amount of groups that necessarily count but rather the number of shots in the group that count. How many have tested a load with five 5 shot groups at 100 yards and thought it was good? Then when shooting 10 shot groups, especially at 200 yards, it went south? I've done it and seen others do it too many times to believe it an anomaly.
I have found shooting 10 shot tests at 100 yards and then confirming with three 10 shot groups at 100 yards, preferably 200 yards, to be "enough". If the load doesn't shoot well with those 40 tests shots then any further amount of shots to find a 95% confidence level that the load doesn't shoot well isn't exactly productive in my view. Conversely, after four 10 shot groups if the load shoots well then that's good "enough" to give it a go in competition or hunting. My opinion of course.
LMG
Concealment is not cover.........
Attached Files
I'd pick enough 190 grain bullets for a few groups. Then some of the the 192 grainers and shoot them. Then I'd take some of the heaviest and the lightest bullets and mix them. Trying to see anything shooting 19 groups separated by a tenth grain I think would be pretty hard
Attached Files
I can remember “way back when” where I used the blind test process that John Carlson describes. I was crazy enough to spend time weighing bullets, sorting, shooting, evaluating, etc. Hobby time became a challenge so I resorted to testing loads at CBA BR matches instead of making many trips to the range in search of perfection.
However, I shoot two Scheutzen matches every summer. I use a Sharps Model 74, black powder, iron sights, from the bench. I used to weigh and sort bullets. Sometimes the match results were OK, sometimes not so OK.
The mold has always been a Paul Jones “Money” bullet design, 416-420 grains. I cast up (500) early this year. This time no weighing, sorting involved.
At our most recent match, my first target was not very good. I decided to switch to a larger aperture opening in the front sight for the second target…best target I have ever shot….random bullets and all! Maybe for me the moral of this brief experience is….it’s bench and set-up technique more than bullet sorting.
YMMV
Tom
Attached Files
The "blind test", if one is honest with himself during the testing is unnecessary. An honestly conducted test procedure with correct testing protocols will reveal the same data.
I have posted several times on this forum the benefits of proper/correct weight sorting. I have also documented the results based on numerous tests. With quality cast bullets to begin with, weight sorting many times does not appear to have any beneficial effect on accuracy at 100 yards. Especially at the 1400 - 2000 fps range most are shot at. However, when the velocity is increased [the RPM increases as velocity increases with a given barrel twist] and/or the range is increased beyond 100 yards we most often see a benefit when the cast bullets have been correctly weight sorted.
I would add one thing to the testing of weight sorted cast bullets: if you are going to compete with groups at 200 yards, then conduct the testing at 200 yards not 50 or 100 yards.
LMG
Concealment is not cover.........
Attached Files
"The "blind test", if one is honest with himself during the testing is unnecessary. An honestly conducted test procedure with correct testing protocols will reveal the same data. "
That is true of course if a human could be absolutely honest both consciously and sub-consciously. Scientists and engineers doing serious research recognize that this isn't likely, even for someone thinking they are being completely honest, thus they do blind testing if possible. It probably isn't true of us gun loonys either.
"A man's gotta know his limitations." Detective Callahan in Dirty Harry
John
Attached Files
but when factors totally out of our control such as wind, lighting, humidity, and temp. reaching a conclusion from just a few groups on a single day may or may not in fact accurately reflect the mean of the population. If sampling were perfect and predictable those that have set group records would be earning screamers on a regular basis. I do not believe that is the case.
Couldn't agree more. Unless you have access to a Houston Warehouse you're test results could be skewed by numerous outside influences you have no control over. I don't have the patience or ambition to shoot hundreds of groups to see if a 175 grain bullets shoots better than a 176 grain bullet when I already KNOW the results might be different the next time out. Then when you have what you think might be the best weight bullet there's powder, powder charge, primer, and God knows what else. By the time you're done with all these tests you have a shelf full of powder you'll never use. An empty pocket from buying everything needed for a serious test. And no closer to what you were looking to find out in the first place.
Attached Files
"So we probably should be arguing about how many groups are needed instead of on "blind tests" which admittedly is a refinement worth doing but much less important than shooting enough groups."
I also concur with Johns statement along with Marks explanation.
I will add that, in the case of testing, all too often it is the size of the sample [number of shots in the test group] that really is what leads us to "self deception". Unless 10 shot test groups are used at least at 100 yards the end result of "best accuracy" can indeed be very deceiving. In addition to using 10 shot tests once a load shows potential the potential of that load, either in testing for terminal ballistics (group size), external ballistics (velocity, ES, SD and BC) or internal ballistics (MAP, ES, SD, pressure trace and rise) then a series of additional 10 shot tests should be conducted. Some of us recommend at least three such additional 10 shot tests. Only then can you have any assurance or confidence the tested load will most likely perform the same in future shooting.
After such complete testing we must also keep in mind is testing for accuracy/group size potential the end result is not the "average" nor is it the smallest group. The end result of the tests will be a high confidence level of the largest group of the four ten shot groups [the first test group plus the three confirmation test groups. A higher level of confidence can be measured by using the same aiming point w/o sight adjustment and over laying the targets to get a composite 40 shot group.
I don't see how a "blind test" could be done with such a method or using the large amount of test sample shots Litz used or what Mark suggests is needed. The only true way to have a 100% confidence in group size, velocity, BC, SD/ES, etc. would be to shoot the entire lot of ammunition. Thus, we are stuck with a conundrum of just how much is enough? SAAMI says and uses 10 shot tests. Since I began measuring internal and external ballistics 15+ years ago and began adhering to SAAMI methods I've come to believe 10 shots in a test string is enough for initial testing.
My exception is if testing a known accurate load [as in small group size] for use on a scored target. Then I use the number of sighters and scored shots allowed in a match on the real target at the farthest distance it is to be used in a match. In such a test I do not shoot for score nor make any sigh adjustments Other than to get on target. What I am looking for is the groups size/cone of fire of the allotted shots for the match. I want the cone of fire to be as small if not smaller than the 10 ring of the 200 yard CBA score target for example. If I have a cone of fire that hold the 10 ring at 200 yards then I am confident the load will hold the 10 ring at 100 yards. I'm sure many of us have found finding an accurate load at 50 or 100 yards does not mean it will be accurate ate 200 yards.
LMG
Concealment is not cover.........
Attached Files
Paul
If the question you are trying to answer is " does sorting shrink my group aggregate" (is sorting a waste of time) The most direct way to find out is to shoot groups with the lots that have been sorted and an equal number of groups from the same population (casting session) before sorting (190 to 191.9).
Shoot them all in the same shooting session and in the same manner as possible using the plastic bag method described by John Carson above. This will ensure that the two types are shot in random order and also "blind".
Let's say you have 120 bullets (from one casting session. Divide into two batches of 60. Sort on group of 60. This will allow you to shoot 12- five shot groups (or 6- ten shot groups) of sorted and the same number of unsorted. This is enough groups of sorted versus unsorted to give a fairly good first try.
If the group average of the sorted bullets is more than 15% better than the unsorted groups, then you can say shooting 12 groups was enough to claim that the sorting improved group size with a confidence level of 95 %.
if the group average of the sorted bullets is less than 15% better, or if the average of the unsorted groups is smaller (don't laugh, this happens more often than you would guess even with 12 groups of each.) then, you need to shoot more comparison groups if you want to know an honest answer.
If you have already sorted the bullets a more severe test would be to shoot a mixture of the light and heavy bullets against groups of identical weight bullets from those in the middle. You will still have to shoot lots of groups of each to find a reliable answer.
John
PS: The estimates of confidence level were by a table provided by Joe Brennan. I could send a copy if interested.
Attached Files
Larry,
NO. I am not saying that. I am also not saying that many shooters are dishonest, Although, as in any group of humans, some probably are.
What I am saying is, if you can do a test "blind" by some simple method similar to what John Carlson suggests above, it eliminates one possible source of bias. You also get the side benefit of firing the competing loads in random order. This is perhaps more important than achieving "blindness" since condition (wind, mirage, cloud cove) can change quickly.
Next best to random order is firing alternate groups of the two loads. However, it is more convenient to fire all of one load and then all of the competing load, and I think that's what most shooters do.
Far more important than the issues above as far as getting invalid results is deciding which load is best by shooting one or two groups of each and thinking you have the truth. All tests such as ladder tests that depend on one group at each powder charge,"tuning" that depends on one group, etc. are exercises in self deception. This has been proven by Bryan Litz (Vol.3 of his books) by extensive range tests.
So we probably should be arguing about how many groups are needed instead of on "blind tests" which admittedly is a refinement worth doing but much less important than shooting enough groups.
John
Attached Files
Mark,
Well, my last stat class was in 1967 so I probably have forgotten more than you. I also think it is 95% certain thatI would flunk such a class today.
I have no objection to your numbers but you seem to be considering a different question but I may not understand.
The 95% confidence level I quoted was the confidence that the sorted bullet were really better than the unsorted IF THE AVERAGE WAS 15% OR MORE SMALLER THAN THE AVERAGE OF THE UNSORTED AVERAGE and 12 groups of each were fired.
if the sorted average had been only 8% smaller it would taken 30 groups of each bullet to reach 95% confidence
If the sorted average had been 25% smaller then it would take only 5 groups of each bullet to reach 95% confidence.
I think it is obvious that if the unsorted groups were three times as big (300%) then one or two groups would give 95% confidence.
So, as with a lot of question, the correct answer to how many groups are needed is -- it depends.
Another thought: Statisticians get hung up on 95 or 99 confidence levels as if any thing less was worthless. However, if you could predict the outcomes of the NFL games with a confidence level of say 60% you could be a multi millionaire before half the season was over -- or maybe have a contract on your head by the casino owners.
That applies to my example. If you thought an 85% confidence level was close enough to be useful. When the average of the sorted groups was10% or more smaller you would only need 4 or 5 groups in each average to achieve 85% (and have a pretty good hint that sorting was worthwhile).
John
Attached Files
Some amount of allowance in weight variance would seem practical. Even using a single cavity mold can produce a spread. Smaller weight bullets such as 22 caliber vary less than say heavy end 30 caliber molds. The user of course has some control. Using 22 Caliber molds might be a waste of time weighing. I seldom see any real difference.If I do I usually can see a defect without weighing. As bullet caliber and weight go up that’s a different story. How much is to much? Each time I do these test, shooting light bullets then average weight bullets and still heavy or above average weight bullets is just a test. Do this test on several different days over a period of time with your best to date loads. Chances are your results may change each time you do this. What works one day might not the next. But finding some sort of consistency over a period of time gives you more confidence that it should shoot. Table manners and conditions will always have you wondering. Some days you just shoot better than others. One of my best score match came from a mix match of weight And separating Bullet runout. It should have been one of my worst. The more firearms you fool with the more you find what works in one doesn’t in another. How do you explain a proven load that goes south.
Attached Files
I would start by shooting a 5shot group of the lightest and then a 5 shot group of the heaviest and go from there.
Attached Files
John,
Please send me a copy of the confidence level numbers.
Attached Files
Paul,
Will do.
BTW if anyone had trouble reading my post above, I have now cleaned up the grammar and typos and it may be easier to read.
John
Attached Files
forward shoulder pressure on butt of rifle eliminates the jittery reticle and groups will improve and be more consistent. this technique takes away most all shooter error. i use a bipod with forward stop. free floated barrels work best.
bullet fit is way more important than miner weight difference. over thinking can be fun though.
Attached Files
So we probably should be arguing about how many groups are needed instead of on "blind tests" which admittedly is a refinement worth doing but much less important than shooting enough groups.
John
I can agree with this. I disagree with Joe Brennan's confidence interval of 95% with a sampling of 12 groups. This is based on my recollection of the stats classes I took in college in the early 80's so I may be off base. 12 groups is 12 populations. When we are looking to predict how close to the mean of 12 populations a single population is, we seek a confidence interval. 12 simply is not enough samples to generate a 95% confidence interval. However if you are looking at 12-5 shot groups with a mean size of "b" all sharing the same center of group it is easy to establish a confidence interval for how close to center of group a single shot will be, and with 60 samples the confidence interval would be high. To really have useful results you need to establish the total number of the population. For example if you want a 95% confidence interval for the mean of 100 groups you need to sample 60 groups per a calculator that I just used. Bottom line is the more groups the better but when factors totally out of our control such as wind, lighting, humidity, and temp. reaching a conclusion from just a few groups on a single day may or may not in fact accurately reflect the mean of the population. If sampling were perfect and predictable those that have set group records would be earning screamers on a regular basis. I do not believe that is the case.
Attached Files
Categories
- All Categories
- General Polls
- Contact Us w/ Forum Issues
- Welcome to The Cast Bullet Association Forum
- General
- Bullet Casting
-
Guns and Shooting
- AR Platform
- TC Contenders & Other Single Shot Handguns
- Shotguns
- Informal Matches & Other Shooting Events
- Gunsmithing Tips
- Gun Cleaning & Maintenance
- Optics
- Benchrest Cast Bullet Shooting
- Military Bench Rest Cast Bullet Shooting
- Silhouette Shooting
- Postal Match Cast Bullet Shooting
- Factory Guns
- Black Powder Cartridge
- Hand Guns
- Lever Guns
- Single Shot Rifles
- Bolt Action Rifles
- Military Surplus Rifles
- Plinkers Hollow
- Muzzleloaders
- Hunting
- Reloading
- Buy, Sell or Trade
- Other Information & Reference
Search
This Weeks High Earners
- Ed Harris 27
- Shuz 16
- OU812 15
- Tom Acheson 14
- Larry Gibson 7
- Ken Campbell Iowa 6
- linoww 6
- M1fuzz 5
- Ross Smith 5
- SkinnerD 5