Seating depth - does it even matter?

[TaperPin]

It's the greatest variance I've seen as well. I've done a decent amount of testing recently and I believe seating depth makes a measurable difference in at least some gun/ammo combinations.

The main test rifle is a .223 Howa with the factory Stocky stock topped off with a Nightforce NXS 2.5-10x42 firing 73 gr. ELD-M with H4895.

I've done 10 shot strings at 100 yards and 825 yards and I'm generally seeing about a 25 percent improvement between a good cbto vs. a bad one. If I get time I might post some more results.

Thanks for posting those groups - that’s really interesting stuff.

 

[huntsd]

Observe the difference in groups. The only change is seating depth.

View attachment 557613

View attachment 557614

View attachment 557615


My process is to start by just kissing the lands, then work backwards. It’s the Berger method. Works well for me.





P

I fully agree with larger sample sizes, but I also think small samples like are good first steps and can help weed out areas that you dont need to focus on. For example, with these results i feel like you can throw away the 2 loads and focus time and components on/around the last load to confirm that its a good load for you

 

[solarshooter]

I fully agree with larger sample sizes, but I also think small samples like are good first steps and can help weed out areas that you dont need to focus on. For example, with these results i feel like you can throw away the 2 loads and focus time and components on/around the last load to confirm that its a good load for you

I thought this for a while too. But small samples (<10 shots), are literally a waste of time and components. You think you are saving both compared to 10 shot groups, but you really should just do the full test or not bother and switch components if you don't like what you're seeing (this is the real best answer).

I know this can be interpreted harshly but please just try and test it yourself by doing a small sample test, say 3-5rd groups, and then repeating that test a few times and seeing if the conclusions from the first test hold. Another way would be to take the same exact load, and shoot multiple 3-5rd groups, and see if those group sizes don't vary by at least 100%. Or run your small sample test, take the best and the worst performers, and subject both to bigger sample testing. Anyone who has actually taken the time to objectively test this for themselves has reached the same conclusion, including this guy right here writing these words.

Here's an example test I personally ran:

My dad has a benchrest gun, 13lbs, Benchmark 24" M24 chambered in 6BR, Pierce single shot action, Nightforce 12-42x56 benchrest scope. I borrowed it from him to see if I could improve his load. His best load that he has shot in it for years is Lapua brass, CCI BR4s, 29.5gr Varget, Sierra 107 Matchking jumping 0.010". When I picked the gun up he also gave me 25rds of his ammo. I started by shooting all this ammo to establish a baseline for precision. Results were a 5x5 agg of 0.46" at 100yds, mean radius 0.21", SD radius 0.12".

I loaded a powder sweep, 3 shots each from 29.2 to 30.4 in 0.3gr increments, which included his original powder charge. Group sizes were as follows:

29.2, 0.32"
29.5, 0.26"
29.8, 0.26"
30.1, 0.16"
30.4, 0.30"

I of course picked 30.1 as the new "node" of interest, since I could see a "pattern" of larger then smaller groups, and it was significantly better (~60%) than the loads on either side. So then to apply the same level of rigor to it as my Dad's load, I loaded up 25 and shot 5x5. Average group was 0.50, mean radius 0.21", SD radius 0.12". Literally IDENTICAL performance to his load which was 0.6gr less powder.

 

[huntsd]

I thought this for a while too. But small samples (<10 shots), are literally a waste of time and components. You think you are saving both compared to 10 shot groups, but you really should just do the full test or not bother and switch components if you don't like what you're seeing (this is the real best answer).

I know this can be interpreted harshly but please just try and test it yourself by doing a small sample test, say 3-5rd groups, and then repeating that test a few times and seeing if the conclusions from the first test hold. Another way would be to take the same exact load, and shoot multiple 3-5rd groups, and see if those group sizes don't vary by at least 100%. Or run your small sample test, take the best and the worst performers, and subject both to bigger sample testing. Anyone who has actually taken the time to objectively test this for themselves has reached the same conclusion, including this guy right here writing these words.

Here's an example test I personally ran:

good points. I too am really starting to lean towards the @Formidilosus way. Pick a bullet and powered. Shoot just under pressure. It either shoots or you move on.... no dinking around

 

[solarshooter]

good points. I too am really starting to lean towards the @Formidilosus way. Pick a bullet and powered. Shoot just under pressure. It either shoots or you move on.... no dinking around

@Formidilosus certainly preached this early on this forum, but I wouldn't call it the "Form way". This has been well known in science for many decades if not hundreds of years, since the field of statistics first came about.

What was missing was educating the everyday reloader on the reality of the situation. Litz, Hornady, and many others (including @Formidilosus) have been able to do this in the last few years. I think the proliferation of chronographs and Quickload in the last ~10 years led to a real explosion in pseudoscientific theories regarding precision - the "Satterlee load dev method" was one of the more infamous and egregious ones. The new emphasis on theory and statistical rigor is a natural reaction to that wave.

But, there are still lots of folks out the "tuning" to "nodes" with small samples, and there probably always will be. It's fundamentally frustrating to accept that you have little to no control over something other than buying more shit. But you can still have fun experimenting, you just have to learn to be a lot more objective and methodical when trying to extract conclusions.

 

[doc holiday13]

my 7 PRC changed how I view CBTO. It could just be based on how seekins chambers their rifles but I found setting my CTBO to match factory ammo was better than seating longer. Oddly enough I'm kind of okay with it because I'll have lots of mag length left over once the throat erodes a little. So I'm only going to load 90 rounds at a time and save 10 rounds to be .005 longer

 

[solarshooter]

my 7 PRC changed how I view CBTO. It could just be based on how seekins chambers their rifles but I found setting my CTBO to match factory ammo was better than seating longer. Oddly enough I'm kind of okay with it because I'll have lots of mag length left over once the throat erodes a little. So I'm only going to load 90 rounds at a time and save 10 rounds to be .005 longer

Can you explain what testing or results you have that led you to this conclusion?

 

[JGRaider]

@Formidilosus certainly preached this early on this forum, but I wouldn't call it the "Form way". This has been well known in science for many decades if not hundreds of years, since the field of statistics first came about.

What was missing was educating the everyday reloader on the reality of the situation. Litz, Hornady, and many others (including @Formidilosus) have been able to do this in the last few years. I think the proliferation of chronographs and Quickload in the last ~10 years led to a real explosion in pseudoscientific theories regarding precision - the "Satterlee load dev method" was one of the more infamous and egregious ones. The new emphasis on theory and statistical rigor is a natural reaction to that wave.

But, there are still lots of folks out the "tuning" to "nodes" with small samples, and there probably always will be. It's fundamentally frustrating to accept that you have little to no control over something other than buying more shit. But you can still have fun experimenting, you just have to learn to be a lot more objective and methodical when trying to extract conclusions.

Agreed. I either load OAL's to max mag length or off lands .020 to start, load in increments of .3 grains starting .06 below book max, and end when I see pressure. I seldom have to screw around with OAL with a few exceptions.......Berger, Accubond LR, and some Barns TSX bullets.

 

[doc holiday13]

Can you explain what testing or results you have that led you to this conclusion?

I find powder charge by taking max charge minus 2g and going up in .5g increments. I keep adding .5g until i see pressure. Then I pick the biggest charge that produced an acceptable consistency of grouping but not showing pressure. In initial load development I set seating dept to mimic factory ammo that shot consistenly well. Once i got my powder charge i did 10 rounds of .025 and 10 of .05 longer. Groups went from 1/2" to over 1" then 2.25" with those jump changes

with a high pressure/magnum cartridge you'll start to see throat erosion around roughly 100-120 rounds and you'll start falling out of your accuracy node because your bullet jump changes. So having your last few rounds in a batch loaded longer are meant to be saved for the range to see if those longer bullets fall within the accuracy node. If you're shooting anything that isn't an overbore/magnum then I do not recommend changing CTBTO. I don't check my 308 till i hit ~900 rounds

 

[solarshooter]

Groups went from 1/2" to over 1" then 2.25" with those jump changes

Are you saying your 7PRC was grouping 10 shots in 0.5"?

Then I pick the biggest charge that produced an acceptable consistency of grouping but not showing pressure.

How many shots per charge weight when you're doing this test?

Once i got my powder charge i did 10 rounds of .025 and 10 of .05 longer.

These are quite large adjustments, are you sure you weren't jamming into the lands with these? Typically 7PRC and a lot of the modern Hornady cartridges are designed such that common bullets at factory COAL will have pretty ideal jump, let's say 0.030-0.050", which is why I ask. How do you measure jump?

 

[TaperPin]

More rounds will never shrink a bad group. If testing something and even the first two rounds are larger than the group they are being tested against I stop, mark them for use as foulers or offhand practice.

Seating depth testing does have value, but new reloaders get overwhelmed and don’t know where to start or stop, and if it is worth messing with. It helps to know if you are trying to reach a group size goal, or if the end result is playing with variables to minimize the absolute most accurate loads. They are two very different goals.

I don’t really care about absolute best accuracy, like benchrest guys do. If I’m happy with a 1 MOA goal and the very first load shoots that, I’m done. Folks get in trouble when they start confusing goals. They want a XX MOA rifle, it won’t shoot that so they waste time and money trying to get a turd to do something it can’t, or it does shoot better than expected and they keep going so the goal didn’t really mean much from the get go.

In 50 rounds of break in and 50 rounds of tinkering, how much does the last 10 tell us that’s different from the first 10? Not much.

If the load was just shy of the goal group size I’d swap bullets and/or powder before fine tuning seating depth, neck tension, powder lots, sorting components by weight, or different brands/types of primers. These last five should not even be considered anything other than low odds, last ditch efforts for small gains, but mostly they are time wasting.

 

[solarshooter]

More rounds will never shrink a bad group.

But more rounds can reduce the Mean Radius. This is why it's such a better/more powerful metric, you are getting information from all shots in the group rather than just the extreme-most two.

Which group is more precise (imagine they have the same extreme size, despite my shitty artwork)?

Stated differently, which one would yield a higher hit rate on a 0.5" target (let's say the group extreme size is 1" for both pictured above, and you adjust your 0 for both)?

 

[doc holiday13]

Are you saying your 7PRC was grouping 10 shots in 0.5"?


How many shots per charge weight when you're doing this test?


These are quite large adjustments, are you sure you weren't jamming into the lands with these? Typically 7PRC and a lot of the modern Hornady cartridges are designed such that common bullets at factory COAL will have pretty ideal jump, let's say 0.030-0.050", which is why I ask. How do you measure jump?

I measured what my max load-to length was. Factory ammo to the lands .125. So I had plenty of distance to mess with. when I got closer to the lands and my groups opened waay up I stopped chasing more accuracy.

I was only doing 5 shots per charge because as much as you want to argue you need a bigger data set, you really don't. I was looking for grouping consistency not smallest groups. You want consistency and then close the gap for accuracy with seating depth

 

[solarshooter]

I measured what my max load-to length was. Factory ammo to the lands .125. So I had plenty of distance to mess with. when I got closer to the lands and my groups opened waay up I stopped chasing more accuracy.

I was only doing 5 shots per charge because as much as you want to argue you need a bigger data set, you really don't. I was looking for grouping consistency not smallest groups. You want consistency and then close the gap for accuracy with seating depth

For a pressure test I tend to agree that fewer shots are acceptable, since the goal is just to find a limit, rather than define or differentiate two populations.

So the 0.5" group was with 5 shots, and the 1" and 2.25" groups were with 10 shots each? I can almost guarantee that's your issue. That 0.5" 5 shot group could easily be a 1-2" group if more samples were taken. And that 1" 10 shot group could also vary significantly, maybe up to 2" as well. Without looking at your data and testing, I can't really say more. Has the 0.5" 5 shot grouping held? And I don't mean your best 5 shot group, I mean what's the worst 5 shot group over the last 10 groups?

 

[Harvey_NW]

I was only doing 5 shots per charge because as much as you want to argue you need a bigger data set, you really don't.

You don't if it's bad, you do if you want to prove it's good. That's the crux of the sample size debate.

 

[doc holiday13]

For a pressure test I tend to agree that fewer shots are acceptable, since the goal is just to find a limit, rather than define or differentiate two populations.

So the 0.5" group was with 5 shots, and the 1" and 2.25" groups were with 10 shots each? I can almost guarantee that's your issue. That 0.5" 5 shot group could easily be a 1-2" group if more samples were taken. And that 1" 10 shot group could also vary significantly, maybe up to 2" as well. Without looking at your data and testing, I can't really say more. Has the 0.5" 5 shot grouping held? And I don't mean your best 5 shot group, I mean what's the worst 5 shot group over the last 10 groups?

Rifle shoots .5 with 3 different bullets all set to the same jump. I'm not the best shooter but I'm not a 1.75"+ shooter

 

[solarshooter]

You don't if it's bad, you do if you want to prove it's good. That's the crux of the sample size debate.

Even this isn't right. You need more regardless. Both the "bad" 5 shot group and the "good" 5 shot group should be expected to vary at least 100-200% if repeated. Considering those variations, the bigger "real" groups that you will start to see with more samples MOSTLY overlap. You aren't proving anything with 5 shots, one way or another. That's why it's a waste of time.

 

[solarshooter]

Rifle shoots .5 with 3 different bullets all set to the same jump. I'm not the best shooter but I'm not a 1.75"+ shooter

To be clear, I'm not saying anything about your shooting abilities or implying any offense. I'm talking purely about demonstrated precision on target as you change certain variables in your load.

I think it's very unlikely that your gun is shooting ONLY 0.5" 5 shot groups, never worse. Again, I'm saying ALL groups for a given load, not excluding any flyers or shots. And I think if you increase that shot count to at least 10, you will see larger and more consistent group size numbers. And you might see less sensitivity to seating depth than you originally stated. Your case would be by far the most sensitive/extreme reaction to a seating depth change that I've ever seen or heard of.

And it is POSSIBLE that you are seeing a real signal. But based on the evidence you provided, I don't think you've PROVEN that yet.

Here's 15 10-shot groups I just shot this weekend with my T1X. Top 2 rows are testing different lots of ammo, bottom row is same lot as top row, but with my suppressor added. You can see how different the 10 shot groups can be, both in size, shape, and shot distribution within the group. Imagine how much more extreme that would be with only 5 shot groups.

 

[Harvey_NW]

Even this isn't right. You need more regardless. Both the "bad" 5 shot group and the "good" 5 shot group should be expected to vary at least 100-200% if repeated. Considering those variations, the bigger "real" groups that you will start to see with more samples MOSTLY overlap. You aren't proving anything with 5 shots, one way or another. That's why it's a waste of time.

I disagree. If my goal is sub-MOA for 10 shots and the first 3 are 1.5", that doesn't fit my criteria and based on the testing I've done and the subject at hand, no amount of tweaking seating depth or likely changing powder charge without sacrificing an unwilling amount of velocity will make that load consistently sub-MOA. So I change a component, and start over.

 

[Q_Sertorius]

I disagree. If my goal is sub-MOA for 10 shots and the first 3 are 1.5", that doesn't fit my criteria and based on the testing I've done and the subject at hand, no amount of tweaking seating depth or likely changing powder charge without sacrificing and unwilling amount of velocity will make that load consistently sub-MOA. So I change a component, and start over.

In this situation, I would keep shooting because I want to establish “the best I can do with this combination.” Because the shooter is always the largest source of error. Unless you are eliminating shooter error and environmental factors, I think it still makes sense to finish the test.

But when I do load development I am at the farm with the range next to my cabin. So I might as well finish the string and go back into the house and try something else. Every shot is a practice shot.

I could go in and use the kinetic puller to offload the remaining seven, but I prefer to have the complete data set.


____________________
“Keep on keepin’ on…”