Don't you have some seating depth testing to be doing? Don't forget to test the neck tension when you finish the seating depth test. Next on the list is primer seating depth. Then install your new barrel.
-
Having trouble using the site?
Contact support
Seating depth test
- Thread starter Rickyd1
- Start date
I meant to tell you that I liked that 3rd group. It looks like you are very near finding your sweet spot. We should shoot some seating depth tests together sometime. Maybe on a beach somewhere? Definitely not any place snowy and windy. But I also don't want to be shooting seating depth tests when I can feel the sweat rolling down the middle of my back into my ass crack...it's just too hard to concentrate on holding the rifle still when I feel that tickle trickle.
I would do it in jump variations: 0.010'', 0.040'', 0.080''. 20 shots each, on 1000y, on Shotmarker in order: 1.=0.010'', 2.=0.040'', 3.=0.080'', 4.=0.010'', 5.=0.040'', 6.=0.080'', ...
I think that way everybody will see what they want and must see...
Keith has done this in 3 different jumps in 11×3-shot groups; in total jump range of 0.024'', which is quite small, but he saw the difference:
I will say that with my own anecdotal evidence from my fat ass walking 30 feet to my shop/reloading bench/shooting range; That seating depth variation has made an impact on my group sizes.
How much depends on the individual barrel and bullet mostly. Some combinations have been much more sensitive than others.
I've watched a video or three made by extremely experienced shooters (who win). The one that stood out most was (3) 30-shot tests done by Keith Glasscock with one of his F-Class rifles. In it, he explored some of his loads based off seating depth and he - as an engineer IIRC - put a lot of statistical emphasis and data into the presentation - even to include that his samples were too small. I was almost completely sold on the idea that I was just the victim of 'random sample' chasing, until I realized that his three sample groups covered a grand total of .............. .015" seating depth difference. Then it made a lot more sense. I would love to see his results if he did that test again covering another .030".
We often fail to account for 1) Shooter error/skill level, and 2) The amount of ammunition expended by each individual during a range session (some guys may only take 50 rounds to the range to shoot from 100 - 1000 yards in a day etc...). The first point is easier for us to understand, though we often overlook it. The second seems to tie us up in arguments more...
...If I'm going to only shoot 5-10 shot groups at any particular distance for any of my recreational shooting...then by the 3rd or 4th time I've done so, I ought to have a pretty good idea of performance potential based off of my results. If I am shooting 20 round strings at an F-class match, then perhaps it would be wiser to shoot 20 shot strings during load development/proofing. Conversely, if I'm shooting a lightweight hunting rifle then I ought to have a pretty solid idea of where that bullet is going to hit after my 4th or 5th 3-shot group... and another 4-5 cold bore attempts.
Expecting one discipline to base their performance testing protocols off of a very different discipline isn't realistic. But we'll argue about this and throw poop at each other for the next decade doing so anyway.
I'm happy to live in my little bubble, and do things my own way. When I find a better way, I switch. I'm never going to be married to any one idea...that's how you never better yourself.
How much depends on the individual barrel and bullet mostly. Some combinations have been much more sensitive than others.
I've watched a video or three made by extremely experienced shooters (who win). The one that stood out most was (3) 30-shot tests done by Keith Glasscock with one of his F-Class rifles. In it, he explored some of his loads based off seating depth and he - as an engineer IIRC - put a lot of statistical emphasis and data into the presentation - even to include that his samples were too small. I was almost completely sold on the idea that I was just the victim of 'random sample' chasing, until I realized that his three sample groups covered a grand total of .............. .015" seating depth difference. Then it made a lot more sense. I would love to see his results if he did that test again covering another .030".
We often fail to account for 1) Shooter error/skill level, and 2) The amount of ammunition expended by each individual during a range session (some guys may only take 50 rounds to the range to shoot from 100 - 1000 yards in a day etc...). The first point is easier for us to understand, though we often overlook it. The second seems to tie us up in arguments more...
...If I'm going to only shoot 5-10 shot groups at any particular distance for any of my recreational shooting...then by the 3rd or 4th time I've done so, I ought to have a pretty good idea of performance potential based off of my results. If I am shooting 20 round strings at an F-class match, then perhaps it would be wiser to shoot 20 shot strings during load development/proofing. Conversely, if I'm shooting a lightweight hunting rifle then I ought to have a pretty solid idea of where that bullet is going to hit after my 4th or 5th 3-shot group... and another 4-5 cold bore attempts.
Expecting one discipline to base their performance testing protocols off of a very different discipline isn't realistic. But we'll argue about this and throw poop at each other for the next decade doing so anyway.
I'm happy to live in my little bubble, and do things my own way. When I find a better way, I switch. I'm never going to be married to any one idea...that's how you never better yourself.
I think it was only .015", as that is how he labeled his data in quotations. But I've been wrong before.
I think maybe people need to take a step back and remember there's a difference between exploratory testing i.e. screening, and testing for publication/scholarly review/commercial production. Maybe you want/need that level of confidence in all aspects of your shooting and loading... maybe not.
The main point was made several posts ago.
If you can determine that what you are seeing is good enough in x rounds, good deal. I do it all the time and so do most of us.
Because...
If you want a statistically viable determination, plan on shooting out a barrel in the process. You are going to need to shoot upwards of 4000 rounds. Anything less holds very little statistical value.
However, to ignore that one cannot determine with one single shot fired that something is not right is pure ignorance.
If you can determine that what you are seeing is good enough in x rounds, good deal. I do it all the time and so do most of us.
Because...
If you want a statistically viable determination, plan on shooting out a barrel in the process. You are going to need to shoot upwards of 4000 rounds. Anything less holds very little statistical value.
However, to ignore that one cannot determine with one single shot fired that something is not right is pure ignorance.
Last edited:
You used the wrong equations to derive a sample size (I mean your online calculator). The equation you used was meant for the sampling distribution of the proportion. Assuming p=0.5 (which is normally assumed for a worse case scenario when p is unknown), N=5000, and ME=0.01, that comes to n=3842 rounded. This makes zero sense based off how you asked the question. “How many rounds should be fired (sampled) to determine with a 99% confidence level that the SD of the MV is within 1% of the measured value?”
You elaborated the question to pertain sampling distribution of the mean of the standard deviation, which btw is something that is not taught in high school stats and rarely taught if at all in graduate school. Even still, a reformulated question using your elaboration makes zero sense to use the equation you indirectly suggested.
But let’s pretend nothing was incorrect. You use an edge case for a 99% CI for a ME of 1%. There is absolutely zero reasons for a reloader to need such a precise estimate.
What the fuck ever.
Point is made and made plain enough.
One cannot claim that 20 or even 50 is enough STATISTICALLY (for any measurement you want to discuss) and then turn around and claim 1 or 3 or 5 or 10 is not.
diggler 1833: thank you for clarifying the total range of jump. I only find Keith's statement: ''Load #1 was .009" from load #2 and .015" from load #3.'', so I thought that maximum total difference jump can be 0.024'', but in reality is probably 0.015''.
yes, this it the first parameter. we must be in the same discipline to determine number of shots and expected accuracy. benchrest shooters must see overlapping holes, hunters not so much.
but Hornady said that after 50 shots group in not expanding anymore, and after 30 shots we can clearly see what we can expect.
so if we are a little lazy and economical, 20 shots may be enough to see what is good and what is bad...
imho
yes, this it the first parameter. we must be in the same discipline to determine number of shots and expected accuracy. benchrest shooters must see overlapping holes, hunters not so much.
but Hornady said that after 50 shots group in not expanding anymore, and after 30 shots we can clearly see what we can expect.
so if we are a little lazy and economical, 20 shots may be enough to see what is good and what is bad...
imho
Square root of the sum of the squares of each Variable. Protocol might address bore condition at start of any test, ie cleaned plus a predetermined number of fouling shots. Barrel temp across the test thus firing interval gives atmospherics. Set-up. “Delivery” ie consistency of shooter interface and trigger control. Change in conditions. Flags plus mirage reading skill thus knowledge of where a given shot should have gone , thus the question of holding for that vs holding a single point of aim. Physics lab, one variable isolated best as possible, is ideal, yet the results of seating depth testing are often obvious. Working at “just kiss” is unforgiving. Shaping the group and its POI at 100 yd to achieve best 1000 yd results is a more interesting discussion.
Not really true. Every additional piece of data gathered (shot fired) raises the confidence. Chances of getting a correct assessment from a 3-shot group is significantly less than doing so from a 20-shot group.
What is true is that you can disqualify certain loads or seating depths from low sample sizes. As an example, if I'm going for 1/4-minute groups and I fire two shots that are an inch apart (and I know that I didn't pull, flinch, whatever), then that configuration is disqualified - no need to fire more. What I do next is load larger amounts for configurations that haven't been disqualified and disqualify the losers there as well. Usually I'm left with one winner at that point that I might or might not decide to validate further.
