How far off the lands do you start?

[Harvey_NW]

As another data point, among many, is this comparison of 3 shot ladder tests to find nodes in a 308 hunting rifle backed up with larger groups that also confirm the same best and worst nodes. One set of tests was 800 rounds ago and it’s duplicated now for comparison. He does a good job of summarizing EC and what the Hornady boys have said, with what is seen in his rifle.

His takeaway is it made about 1/4 MOA difference, which may or may not be worth the time and money to shoot it for someone else.

This is a good video, but it's still drawing conclusions from groups without applying the variability. I'm not a statistician, so I'm probably messing this up. Maybe @solarshooter can chime in with corrections, but this is where my head goes when I watch this video-

Hornady says there's still 20-25% variability in 20 shot groups, and 15-20% in 30 shot groups, even 10% still in 50. The way they explain it is that percentage stat is +/- from an average. Since this is a single group and not an average, one would have to consider the possibility that it's on the low end of the expected spectrum, so it could have ~40+% increase (splitting the difference using 20%) if repeated, so ~1.2", very close to the other group.

Factor in shooter ability, weather, fatigue, and 50 rounds out of a bare muzzle 308, and there's just not enough there for me to conclude that one seating depth was actually better than the other.

 

[TaperPin]

This is a good video, but it's still drawing conclusions from groups without applying the variability. I'm not a statistician, so I'm probably messing this up. Maybe @solarshooter can chime in with corrections, but this is where my head goes when I watch this video-

Hornady says there's still 20-25% variability in 20 shot groups, and 15-20% in 30 shot groups, even 10% still in 50. The way they explain it is that percentage stat is +/- from an average. Since this is a single group and not an average, one would have to consider the possibility that it's on the low end of the expected spectrum, so it could have ~40+% increase (splitting the difference using 20%) if repeated, so ~1.2", very close to the other group.

Factor in shooter ability, weather, fatigue, and 50 rounds out of a bare muzzle 308, and there's just not enough there for me to conclude that one seating depth was actually better than the other.

You and I will just have to agree to disagree. He shoots that same 308 a lot, has plenty of 10 shot groups around 1/2 MOA in other tests so he seems more capable than the average guy, the wind conditions as stated were quite good, he doesn’t throw out fliers to alter the results, and all the shooting he did for this video was duplicated at an earlier date, which doubles the round count.

Beyond that, there was no doubt in my mind going back 30 years ago seating depth can fine tune a load based on extensive shooting by others, so what he did is more or less what I’d expect.

Not much in reloading is brand new never been discovered before. Powder charge makes a larger difference, but all the accuracy minded shooters I know do a seating depth test after settling on a powder charge. Not one out of a hundred, or twenty, but all of them.

The guys I know that reload for hunting don’t usually find the effort worth the results, partially because their rifles aren’t very accurate and seeing trends through the fog takes higher round counts, and partially because the accuracy achieved with varying the powder charge is good enough.

If long bullet jump didn’t have any effect on accuracy we’d all be shooting rifles with gigantic freebore built into the chamber like the Weatherby cartridges where 1/4” bullet jump isn’t uncommon. We’d be able to get a little extra velocity at the same pressure level and same COL.

The Hornady guys have come out and said what they focus on is information for average reloaders and does not fully translate to the accuracy sports. They are very much in advertising and while much of what is said sounds objective their comments are carefully crafted and often leave gray area in how it’s interpreted. At the end of the day their #1 job is to encourage the average reloader to be happy with their Hornady products and buy as much as possible.

 

[TaperPin]

That video is well done. For those not watching the video: he selected a seating depth using Cortina's method and then loaded 25 of those. He then picked the "worst" seating depth identified by Cortina's method and shot 25 of those 5, 5-shot groups of each he compiled into a composite group of 25 for each load. Very similar to what I challenged the OP to do seveal posts back. The "good node" group measured 0.861" with a mean radius of 0.287" while the "bad node" measured 1.129" with a mean radius of .332." Good shooting.

If I had gone to that trouble, I would definitely pick the seating depth that shot smaller in the test. Confidence is worth having. But . . . it's worth remembering that one should expect about 20% variability in size of even 25-shot groups. Is one group actually better than the other? Maybe? Probably a little? (I'm sure there's a statistical test that can actually asnwer that question). Is it worth your trouble? If I put 25 shots in 1.129" I'm rolling with it.

View attachment 947004

It shouldn’t get lost that generalized tables and expected extreme spreads can not describe a specific rifle and load, any more than statistics about average families describes a specific household, and shouldn’t be treated as such. Some rifle/load combinations are very consistent and some are very sporatic, so trying to estimate what a specific rifle will do based on computer simulation dispersions is a poor decision making tool. These are well meaning and do help people visualize the concept, but many barrels are burned up every year that don’t follow these simplistic percentages, not by a long shot. These remind me of something Brian Litz would put out.

 

[Harvey_NW]

You and I will just have to agree to disagree.

That's fine.

If long bullet jump didn’t have any effect on accuracy we’d all be shooting rifles with gigantic freebore built into the chamber like the Weatherby cartridges where 1/4” bullet jump isn’t uncommon. We’d be able to get a little extra velocity at the same pressure level and same COL.

But many of those Weatherby's were just as accurate as the competing Remingtons and Winchesters, so does it matter? Scott Satterlee says he cuts .300" of freebore into a certain chambering of his comp rifles because they just shoot great from there, and he's not an unknown competitor.

The Hornady guys have come out and said what they focus on is information for average reloaders and does not fully translate to the accuracy sports. They are very much in advertising and while much of what is said sounds objective their comments are carefully crafted and often leave gray area in how it’s interpreted. At the end of the day their #1 job is to encourage the average reloader to be happy with their Hornady products and buy as much as possible.

We'll have to disagree on this too, because in one episode they brought this up and laughed, Jayden said something like "they don't know what we use", and then alluded to their comp teams and how they've used and tested the same components as the top competitors, and still not seen the magic.

These are well meaning and do help people visualize the concept, but many barrels are burned up every year that don’t follow these simplistic percentages, not by a long shot. These remind me of something Brian Litz would put out.

And apparently they go completely undocumented, and the results are forbidden for the world to see. The closest example I've seen of something falling outside the statistical figures is the last video you posted, and there's quite a few variables in there. This is the type of stuff Litz and Hornady would put out, because they have unlimited funding, burn down hundreds of barrels, and use precision equipment like a rail gun to eliminate variables and get repeatable results, and curiously these trends haven't seemed to show up. Weird.

 

[TaperPin]

That's fine.


But many of those Weatherby's were just as accurate as the competing Remingtons and Winchesters, so does it matter? Scott Satterlee says he cuts .300" of freebore into a certain chambering of his comp rifles because they just shoot great from there, and he's not an unknown competitor.


We'll have to disagree on this too, because in one episode they brought this up and laughed, Jayden said something like "they don't know what we use", and then alluded to their comp teams and how they've used and tested the same components as the top competitors, and still not seen the magic.


And apparently they go completely undocumented, and the results are forbidden for the world to see. The closest example I've seen of something falling outside the statistical figures is the last video you posted, and there's quite a few variables in there. This is the type of stuff Litz and Hornady would put out, because they have unlimited funding, burn down hundreds of barrels, and use precision equipment like a rail gun to eliminate variables and get repeatable results, and curiously these trends haven't seemed to show up.

I think it’s great that you’re obviously interested in the topic and have spent time listening and digesting it - nothing wrong with that. I’ve even said the Hornady podcast has done more to educate shooters than any other source.

 

[solarshooter]

Maybe @solarshooter can chime in with corrections

Sorry to be late to the party, been busy!

The "good node" group measured 0.861" with a mean radius of 0.287" while the "bad node" measured 1.129" with a mean radius of .332."

A couple notes wrt the linked video.

1. Note the small difference in MR (14%) but the large difference in group size (24%). MR is a higher quality metric because it is informed by all the available data, ie every shot in the group, rather than just the extreme outliers. Use MR if you want to get to the "correct" answer "faster".
2. Simply comparing MR is not enough to determine if these loads are really different. What should be done is a statistical test, for instance a T-test (https://en.wikipedia.org/wiki/Student's_t-test), to more rigorously/quantifiably assess differences between the two groups. In order to do this you need to log the coordinates or MR of each individual shot, and compare the populations. Then you can calculate the confidence level with which the two populations of data are actually different.
3. Even with 25 shot groups, the error in the MR or group size can be significant. That's why a 14% difference may not actually mean the two loads are different. Could say much more on this but my time to post is short.

I haven't read through the entire thread, but I will say that I have not ever personally seen seating depth make a statistically significant difference in group size. Here's a test I ran not too long ago on which variables showed an effect on precision, using the data gathering and analysis methods I'm a fan of: https://rokslide.com/forums/threads/what-did-you-do-in-the-reloading-room-today.338563/post-4125326.

 

[49ereric]

Never had tools to measure so I start at recommended seating depth and shoot then I seat longer and repeat. Brother’s savage 110 though required shorter than recommended seating depth with Sierra 150PH bullets.
44 K98 using Sierra 175’s I had to barely seat the bullet but was made for 196-198 grain bullets.
I ended up at bullet seating depth is @17 thousands.
All my rifles show a difference in accuracy manipulating seating depth.

 

[Okie_Poke]

Sorry to be late to the party, been busy!


A couple notes wrt the linked video.

1. Note the small difference in MR (14%) but the large difference in group size (24%). MR is a higher quality metric because it is informed by all the available data, ie every shot in the group, rather than just the extreme outliers. Use MR if you want to get to the "correct" answer "faster".
2. Simply comparing MR is not enough to determine if these loads are really different. What should be done is a statistical test, for instance a T-test (https://en.wikipedia.org/wiki/Student's_t-test), to more rigorously/quantifiably assess differences between the two groups. In order to do this you need to log the coordinates or MR of each individual shot, and compare the populations. Then you can calculate the confidence level with which the two populations of data are actually different.
3. Even with 25 shot groups, the error in the MR or group size can be significant. That's why a 14% difference may not actually mean the two loads are different. Could say much more on this but my time to post is short.

I haven't read through the entire thread, but I will say that I have not ever personally seen seating depth make a statistically significant difference in group size. Here's a test I ran not too long ago on which variables showed an effect on precision, using the data gathering and analysis methods I'm a fan of: https://rokslide.com/forums/threads/what-did-you-do-in-the-reloading-room-today.338563/post-4125326.

Yeah, I subsequently looked up the appropriate test and ran a two-tail t-test on the five, five shot groups. One can pull the five-shot data from video. Even at 90% confidence interval, you can’t reject null hypothesis. So, statistically, you can’t conclude one group is better than the other. And as you point out, those mean radii are pretty close. I gave up on the thread and didn’t post it because it wouldn’t change anyone’s mind at this point.