Hornady on tuners.

Status
Not open for further replies.
The problem is that just like Litz, they keep using the same flawed method of testing by having a device that has as many 50 or more settings and only using 3 or 4 of those settings in their test and then proclaiming the device doesn't work....That's the same logic as using seating depth variations of every .020" and being baffled why you can't seem to find any nodes that group well.....
What you just said not only hurts and helps both causes of for or against the use of tuners.

When someone frames the experiment, they have to frame it from the aspect of the group size given the tuner setting. So we need to define what a group is? Is it 5 shots, 10 shots, 30 shots, etc.? To get an accurate representation of the true group size, it has to be much larger 5 shots. 30 shots has been tossed around so let’s use that.

Moving on from the definition of the number of shots per group, a single group is only a sample of 1. Meaning, the group shot is a sample size of one. So, at any particular setting we need more than one group. We’ll pretend we did the analysis and it is 5 different groups per tuner setting.
How many tuner settings? 5, 10, 20?? Let’s say 10 different tuner settings.

So, 30 shot per group x 5 groups per tuner setting x 10 tuner settings. That’s 1500 rounds for a single rifle. That’s assuming my assumptions are correct and it could be much worse like 2000+ rounds needed for a single rifle.

Now, what about a rifle with this barrel or that barrel, or this tuner or that tuner, a barrel this long or that long, or a BR rifle or a PRS rifle, etc.

The test becomes so damn expensive that no one will want to sink the money into it.

Edit:
Anyone who wants to conduct this test, I will freely offer statistical consulting and provide a detail report of the experiment and its design. If you want it publication ready, I can do that as well.
 
Last edited:
The problem is that just like Litz, they keep using the same flawed method of testing by having a device that has as many 50 or more settings and only using 3 or 4 of those settings in their test and then proclaiming the device doesn't work....That's the same logic as using seating depth variations of every .020" and being baffled why you can't seem to find any nodes that group well.....
Sounds like a relocation of goal posts. There will never be enough factors to include. So, you have to go for the average, hence the statistical universe they create.
 
  • Like
Reactions: FredHammer
What you just said not only hurts and helps both causes of for or against the use of tuners.

When someone frames the experiment, they have to frame it from the aspect of the group size given the tuner setting. So we need to define what a group is? Is it 5 shots, 10 shots, 30 shots, etc.? To get an accurate representation of the true group size, it has to be much larger 5 shots. 30 shots has been tossed around so let’s use that.

Moving on from the definition of the number of shots per group, a single group is only a sample of 1. Meaning, the group shot is a sample size of one. So, at any particular setting we need more than one group. We’ll pretend we did the analysis and it is 5 different groups per tuner setting.
How many tuner settings? 5, 10, 20?? Let’s say 10 different tuner settings.

So, 30 shot per group x 5 groups per tuner setting x 10 tuner settings. That’s 1500 rounds for a single rifle. That’s assuming my assumptions are correct and it could be much worse like 2000+ rounds needed for a single rifle.

Now, what about a rifle with this barrel or that barrel, or this tuner or that tuner, a barrel this long or that long, or a BR rifle or a PRS rifle, etc.

The test becomes so damn expensive that no one will want to sink the money into it.

Edit:
Anyone who wants to conduct this test, I will freely offer statistical consulting and provide a detail report of the experiment and its design. If you want it publication ready, I can do that as well.
If we could do that and the barrel held up to 1500 rounds , what would you think the confidence interval would be with this test design?
 
If we could do that and the barrel held up to 1500 rounds , what would you think the confidence interval would be with this test design?
I have no idea. The intervals would be easy to construct since I’d let the statistical software handle that. The more things controlled and accounted for will construct better intervals.

I personally found similar findings that Hornady did. The act of putting a tuner (without adjusting the settings) turned my 300PRC into a nice shooting rifle. I didn’t bother doing a big test because I’m not throwing money at a magnum rifle. I couldn’t find any difference in tuner setting I used but the groups did shrink with the tuner installed. I went on to win two different 1MOA LR matches with it so I guess it technically worked for me.
 
So you mean physics doesn’t magically stop working lol
justsaynotonodes.jpg



ES and SD don't really appear to change with powder charge in my testing, either. Unless you get into the realm of sneezing cases (high pressure) or on the low end with really stupid low case fill, recorded 20-50 shot string SD's are going to live within about a 1-2fps window. If 39gr of H4350 is 10fps SD, 40gr and 41gr and everything in between is going to be 9-11fps with regularity. The idea that there's a "happy spot" that can be differentiated by 0.3gr of powder is silly.
 
I have no idea. The intervals would be easy to construct since I’d let the statistical software handle that. The more things controlled and accounted for will construct better intervals.

I personally found similar findings that Hornady did. The act of putting a tuner (without adjusting the settings) turned my 300PRC into a nice shooting rifle. I didn’t bother doing a big test because I’m not throwing money at a magnum rifle. I couldn’t find any difference in tuner setting I used but the groups did shrink with the tuner installed. I went on to win two different 1MOA LR matches with it so I guess it technically worked for me.

Lol I understand and have been there myself. I sure would be interested in seeing a 6 grain graph on your rifle with and without a tuner to see what pattern it produced which could help to explain why your rifle shoots better and why the tuner showed no change by adjustments.
 
  • Like
Reactions: JB.IC
Heres an interesting thought...

Bullets. Do the bullets themselves cause variations in ES SD. ??
Jacket density, or lead core density. What it one bullet is marginally harder or softer than the next ? One engraves easier in the rifling, altering velo and pressures ?
What if one has a longer driving surface than the next ?

Get a brand new bullet, roll it on paper a bit, and you get a duller surface, showing exactly how long the parallel drive band is. Do that to 10 bullets.. are they all identical length, or do they vary by 0.1mm ? Does that have an effect on engraving force, or engraving force duration ?

It gets to a point that variables are out of control, and we will never see a 25 shot string of 3 SD, 5 ES and a 0.3moa / 0.1mil group.

There.. i said it. Yes i expect to be wrong.. or right. No idea.
 
The problem is that just like Litz, they keep using the same flawed method of testing by having a device that has as many 50 or more settings and only using 3 or 4 of those settings in their test and then proclaiming the device doesn't work....That's the same logic as using seating depth variations of every .020" and being baffled why you can't seem to find any nodes that group well.....
The thing is they aren't testing if it's possible to use the tuner to make groups better. They are testing if it makes any difference at all and thus far the 2 tests that were conducted scientifically have shown that moving the tuner has made no difference in group size.
 
The thing is they aren't testing if it's possible to use the tuner to make groups better. They are testing if it makes any difference at all and thus far the 2 tests that were conducted scientifically have shown that moving the tuner has made no difference in group size.
That's true because they are skipping 90% of the tuner settings that can have numerous settings in that range that can potentially make improvements in group sizes. Erik has numerous videos on how to correctly test and validate his tuners function, but so far the testing done by Litz and Hornady ignore his simple procedure to do it and continue to miss the boat.
 
That's true because they are skipping 90% of the tuner settings that can have numerous settings in that range that can potentially make improvements in group sizes. Erik has numerous videos on how to correctly test and validate his tuners function, but so far the testing done by Litz and Hornady ignore his simple procedure to do it and continue to miss the boat.
Go back and read my post again.
 
That's true because they are skipping 90% of the tuner settings that can have numerous settings in that range that can potentially make improvements in group sizes. Erik has numerous videos on how to correctly test and validate his tuners function, but so far the testing done by Litz and Hornady ignore his simple procedure to do it and continue to miss the boat.

So if I go a full revolution of the tuner, and there's no change in group size, it's because I skipped all of the settings that would have produced a change in group size?
 
Wait, so we're now saying that the Rocket Scientist owned research company......and all the ballisticians and such at Hornady.....are just "skipping 90%"??
To be fair, it’s hard to capture a non-linear relationship with only a couple settings. Say someone tried only two settings and the group sizes are exactly the same, then that’s a flat line. But if you try three settings, the the first and last setting is the same result but the middle setting is either better or worse than the other two, the response is non-linear (assuming all else was equal). Now let’s pretend that many different settings produce results that are not the same. Some better, some worse.

Let’s pretend that the tuner is as sensitive as people say they are where very small adjustments have an impact. The number of combinations that would be required to test all those settings to get a statistically valid conclusion would be so damn large and expensive, it would not be worth even testing it. At worst, it wouldn’t be repeatable from one rifle to the next. So, you would have to spend all that time and money on the new rifle because your results from the last test would not apply.

I like testing things, but I’m not about to dump $1000+ just to get a statistically valid test just to maybe find the right tuner setting just in time for the barrel to be shot out.
 
I like testing things, but I’m not about to dump $1000+ just to get a statistically valid test just to maybe find the right tuner setting just in time for the barrel to be shot out.
This would argue for using a long-lasting barrel-preserving cartridge for testing, like 308 or 223.
 
Last edited:
The thing is they aren't testing if it's possible to use the tuner to make groups better. They are testing if it makes any difference at all and thus far the 2 tests that were conducted scientifically have shown that moving the tuner has made no difference in group size.


My understanding of Brian’s tests is that they would better the groups verses without and changes with the tuner made groups worse therefore the tuner effected the group size . Their whole premise was that since tuners did not better the groups then they did not work . But they did change with adjustments .
 
Heres an interesting thought...

Bullets. Do the bullets themselves cause variations in ES SD. ??
Jacket density, or lead core density. What it one bullet is marginally harder or softer than the next ? One engraves easier in the rifling, altering velo and pressures ?
What if one has a longer driving surface than the next ?

Get a brand new bullet, roll it on paper a bit, and you get a duller surface, showing exactly how long the parallel drive band is. Do that to 10 bullets.. are they all identical length, or do they vary by 0.1mm ? Does that have an effect on engraving force, or engraving force duration ?

It gets to a point that variables are out of control, and we will never see a 25 shot string of 3 SD, 5 ES and a 0.3moa / 0.1mil group.

There.. i said it. Yes i expect to be wrong.. or right. No idea.


So here's a fun one I did recently. I sorted through roughly 1500 6mm bullets to finally get 30x of each on the high and low end of weight variation (about a 0.5gr total spread). And if you figure my average velocity in my 6mm ARC is ~2750 and that because the KE equation is 1/2mv^2 the mass term should be pretty linear and you do KE1 = KE2 (assuming the powder charge [chemical potential energy] being the same will result in the same final KE) with the two different masses (108.7 and 109.2, for example), you should see about 7-8fps change in the average of those weight sorted extremes.

30 shot strings:
108.7- 2735.6fps
109.2- 2735.2fps

ES within 3fps, SD well within 1 fps between the two.

So big picture takeaway, the transition of chemical potential energy into KE isn't 100% efficient, and there are many factors included as to what makes the peak pressure happen, the duration of pressure, and the sum of all of it that finally results in the muzzle exit velocity of a particular shot. At least in this test, the effect of weight variation is being grossly out-weighed by some other factor(s). Unfortunately I didn't think to measure base-to-ogive on these before I shot them but I still have a bunch of the rest of them I could go through and measure to see if there's any trend there.

At some point, however, things happen that are out of our control on a scale that's not feasible to control. At some point the scale to which a tuner "must need to be" to be "in tune" slips inside of variation of the forcing function we see from P&V results and that's where my level of skepticism about the whole thing comes from.

On aliasing...
The testing that Jayden did puts a pretty big damper on the whole "shoot 2, if they're not touching adjust the tuner...." until all 3 touch. We kind of suspected that going into it. The 3-shot strings do not correlate to the 20-50 shot strings (shocker). So maybe we aliased and skipped over the best settings, but what are the odds that we did it exactly 3-5 times per barrel to achieve pretty much the same results within the error brackets of what a 20-50 shot test will repeat to with/without/at different settings?

As someone else pointed out, to hit every setting with statistically confident results would take someone weeks of dedicated range use and several thousand rounds. To anyone saying we did it wrong, the world is just lying there waiting for you to do it better and share your results.

Again, I'm not saying that muzzle devices don't play a part in dispersion. I have been conducting a series of tests with this and in bolt guns, gas guns, the accuracy fixture I can document different levels of dispersion with different muzzle devices and that's repeatable. I just don't know that the reason why that happens if because of the weight/vibrations/mode shapes or muzzle exit gas flow, both, neither, something else.... Still digging into it.
 
So here's a fun one I did recently. I sorted through roughly 1500 6mm bullets to finally get 30x of each on the high and low end of weight variation (about a 0.5gr total spread). And if you figure my average velocity in my 6mm ARC is ~2750 and that because the KE equation is 1/2mv^2 the mass term should be pretty linear and you do KE1 = KE2 (assuming the powder charge [chemical potential energy] being the same will result in the same final KE) with the two different masses (108.7 and 109.2, for example), you should see about 7-8fps change in the average of those weight sorted extremes.

30 shot strings:
108.7- 2735.6fps
109.2- 2735.2fps

ES within 3fps, SD well within 1 fps between the two.

So big picture takeaway, the transition of chemical potential energy into KE isn't 100% efficient, and there are many factors included as to what makes the peak pressure happen, the duration of pressure, and the sum of all of it that finally results in the muzzle exit velocity of a particular shot. At least in this test, the effect of weight variation is being grossly out-weighed by some other factor(s). Unfortunately I didn't think to measure base-to-ogive on these before I shot them but I still have a bunch of the rest of them I could go through and measure to see if there's any trend there.

At some point, however, things happen that are out of our control on a scale that's not feasible to control. At some point the scale to which a tuner "must need to be" to be "in tune" slips inside of variation of the forcing function we see from P&V results and that's where my level of skepticism about the whole thing comes from.

On aliasing...
The testing that Jayden did puts a pretty big damper on the whole "shoot 2, if they're not touching adjust the tuner...." until all 3 touch. We kind of suspected that going into it. The 3-shot strings do not correlate to the 20-50 shot strings (shocker). So maybe we aliased and skipped over the best settings, but what are the odds that we did it exactly 3-5 times per barrel to achieve pretty much the same results within the error brackets of what a 20-50 shot test will repeat to with/without/at different settings?

As someone else pointed out, to hit every setting with statistically confident results would take someone weeks of dedicated range use and several thousand rounds. To anyone saying we did it wrong, the world is just lying there waiting for you to do it better and share your results.

Again, I'm not saying that muzzle devices don't play a part in dispersion. I have been conducting a series of tests with this and in bolt guns, gas guns, the accuracy fixture I can document different levels of dispersion with different muzzle devices and that's repeatable. I just don't know that the reason why that happens if because of the weight/vibrations/mode shapes or muzzle exit gas flow, both, neither, something else.... Still digging into it.

I did some testing for bullet weight and other dimensions with MV being the response with two different rifles in two different cartridges. The test was just north of 300 rounds each.

One of the rifles, which had a much variable bullet in weight (I chose that on purpose because without variation we cannot learn it’s impact), did not have any statistical impact on MV at any point in the test.

The other had much less variable weight. Within a hundred rounds, bullet weight did have a statistical impact on MV. Then after another 100 rounds or so it completely disappeared and was not statistically significant.

I was disappointed because I wanted it to have a measurable affect on MV while ruling a lot of things out. As you pointed out, we can only rule so much out because some variation we cannot control which we are left with a blended variation.

You being an engineer I think you can appreciate this: the things we are asking these rifles to do at small variations is incredibly surprising. I do not think others have been exposed to other phenomenons to appreciate that someone can load a cartridge and obtain a MV SD around 10 FPS or less. Or a rifle that can produce MOA or less. There are very few things in this world that have variations that small outside of special case manufacturing machines. A single cartridge and rifle has so many stacking tolerances that it’s still kind of crazy how well these rounds and rifles perform.
 
So here's a fun one I did recently. I sorted through roughly 1500 6mm bullets to finally get 30x of each on the high and low end of weight variation (about a 0.5gr total spread). And if you figure my average velocity in my 6mm ARC is ~2750 and that because the KE equation is 1/2mv^2 the mass term should be pretty linear and you do KE1 = KE2 (assuming the powder charge [chemical potential energy] being the same will result in the same final KE) with the two different masses (108.7 and 109.2, for example), you should see about 7-8fps change in the average of those weight sorted extremes.

30 shot strings:
108.7- 2735.6fps
109.2- 2735.2fps

ES within 3fps, SD well within 1 fps between the two.

So big picture takeaway, the transition of chemical potential energy into KE isn't 100% efficient, and there are many factors included as to what makes the peak pressure happen, the duration of pressure, and the sum of all of it that finally results in the muzzle exit velocity of a particular shot. At least in this test, the effect of weight variation is being grossly out-weighed by some other factor(s). Unfortunately I didn't think to measure base-to-ogive on these before I shot them but I still have a bunch of the rest of them I could go through and measure to see if there's any trend there.

At some point, however, things happen that are out of our control on a scale that's not feasible to control. At some point the scale to which a tuner "must need to be" to be "in tune" slips inside of variation of the forcing function we see from P&V results and that's where my level of skepticism about the whole thing comes from.

On aliasing...
The testing that Jayden did puts a pretty big damper on the whole "shoot 2, if they're not touching adjust the tuner...." until all 3 touch. We kind of suspected that going into it. The 3-shot strings do not correlate to the 20-50 shot strings (shocker). So maybe we aliased and skipped over the best settings, but what are the odds that we did it exactly 3-5 times per barrel to achieve pretty much the same results within the error brackets of what a 20-50 shot test will repeat to with/without/at different settings?

As someone else pointed out, to hit every setting with statistically confident results would take someone weeks of dedicated range use and several thousand rounds. To anyone saying we did it wrong, the world is just lying there waiting for you to do it better and share your results.

Again, I'm not saying that muzzle devices don't play a part in dispersion. I have been conducting a series of tests with this and in bolt guns, gas guns, the accuracy fixture I can document different levels of dispersion with different muzzle devices and that's repeatable. I just don't know that the reason why that happens if because of the weight/vibrations/mode shapes or muzzle exit gas flow, both, neither, something else.... Still digging into it.


One thing I would add to my method of 2 shots with 1 grain difference on powder charge at 1000 was more to confirm tune with positive compensation. In the morning they hit level but in the afternoon they did not so a readjust was needed in a large temp swing . I kind if did something similar years ago while trying to get level velocities but what I was able to correlate was bearing surface variation and measure poi drop according to that variation . With .016 difference equaled 10 inches of vertical dispersion at 1000 yds . Did not seem to effect velocities though.
 
Last edited:
  • Like
  • Haha
Reactions: Tokay444 and Ledzep
You being an engineer I think you can appreciate this: the things we are asking these rifles to do at small variations is incredibly surprising. I do not think others have been exposed to other phenomenons to appreciate that someone can load a cartridge and obtain a MV SD around 10 FPS or less. Or a rifle that can produce MOA or less. There are very few things in this world that have variations that small outside of special case manufacturing machines. A single cartridge and rifle has so many stacking tolerances that it’s still kind of crazy how well these rounds and rifles perform.

Blows me away routinely.
 
Obviously the engineers, statisticians, and scientists in the thread already know this......but if you ever want to see how nutty small sample size can be (going to use velocity SD in this case.....but dispersion is more or less same idea)........

Go into excel and use the data analysis tool. Use the random number generator tab. Put in the average velocity you want in the "mean" and whatever standard deviation you want. For this example, I'll use 7sd. Not terribly too low and obviously not high. Make sure you're using "normal" distribution.

Now, run it for 3,000 samples (or whatever. just picked 3000 as a generic barrel life for a 6br). It'll auto fill in 3000 random velocities that you would see in a rifle with that average FPS with that SD. Now, make a formula for SD for the first 3 or 5 velocities. Auto fill that all the way down. And take a look how often your SD for those small samples is significantly higher or lower than 7.


I just ran a 3,000 rnd sample and use 5rnd groups. Almost 13% had a double digit or higher SD. That's more than 1/10 five round group you fire. And people are culling those powder charge weights using that data. There was also a 4sd or less about 17% of the groups. Which most loaders would also focus on because it's so "good."

So now, you're making decisions to either cull a powder charge or to keep a powder charge using numbers that come up very regularly which do not actually tell the story of what's going on.



Obviously the above mentioned professions have better tools for the job than excel. It's just a really cool way for the average person to run their own half ass sims and see just how much variance/randomness happens with this stuff.
 
Obviously the engineers, statisticians, and scientists in the thread already know this......but if you ever want to see how nutty small sample size can be (going to use velocity SD in this case.....but dispersion is more or less same idea)........

Go into excel and use the data analysis tool. Use the random number generator tab. Put in the average velocity you want in the "mean" and whatever standard deviation you want. For this example, I'll use 7sd. Not terribly too low and obviously not high. Make sure you're using "normal" distribution.

Now, run it for 3,000 samples (or whatever. just picked 3000 as a generic barrel life for a 6br). It'll auto fill in 3000 random velocities that you would see in a rifle with that average FPS with that SD. Now, make a formula for SD for the first 3 or 5 velocities. Auto fill that all the way down. And take a look how often your SD for those small samples is significantly higher or lower than 7.


I just ran a 3,000 rnd sample and use 5rnd groups. Almost 13% had a double digit or higher SD. That's more than 1/10 five round group you fire. And people are culling those powder charge weights using that data. There was also a 4sd or less about 17% of the groups. Which most loaders would also focus on because it's so "good."

So now, you're making decisions to either cull a powder charge or to keep a powder charge using numbers that come up very regularly which do not actually tell the story of what's going on.



Obviously the above mentioned professions have better tools for the job than excel. It's just a really cool way for the average person to run their own half ass sims and see just how much variance/randomness happens with this stuff.

EDIT:
I made an error and own up to it. I'll try update this. Probability was completely off the mark.

EDIT2:
A more concrete example for the variability of the sample STD is given by the following 95% confidence interval.

- The true STD is 12
- The sample STD is 3
- The sample size is 2
95% CI of the sample STD is 1.33 to 95.73 FPS. Meaning, 19 out of 20 sample STDs will fall within that interval. As we can all see, the interval is smaller than the true STD on the low end of the CI.
Probability of sample STD being less than or equal to 6: 38.3%

Bump it up to n = 5
- The true STD is 12
- The sample STD is 3
- The sample size is 5
95% CI of the sample STD is 1.79 to 8.62 FPS.
Probability of sample STD being less than or equal to 6: 9.02%

What makes this worse is that this interval will change depending on the sample STD. Lets do this again with a larger sample size and a closer sample STD to the true STD.
- The true STD is 12
- The sample STD is 10
- The sample size is 15
95% CI of the sample STD is 7.32 to 15.77 FPS.
Probability of sample STD being less than or equal to 6: 0.22%

With a closer sample STD.
- The true STD is 12
- The sample STD is 11.5
- The sample size is 30
95% CI of the sample STD is 9.16 to 15.45 FPS.
Probability of sample STD being less than or equal to 6: 0.0013%

Finally, an unrealistic example:
- The true STD is 12
- The sample STD is 11.9
- The sample size is 100
95% CI of the sample STD is 10.44 to 13.82 FPS.
Probability of sample STD being less than or equal to 6: ~0.0%

In all cases, small sample sizes favored a small STD on the lower end of the CI. In all cases, seeing a 6 FPS sample STD dropped quickly as the sample size grew. I chose 6 because it was half of 12. Otherwise, there was no real reason other than making a comparison decision.
 
Last edited:
Heres an interesting thought...

Bullets. Do the bullets themselves cause variations in ES SD. ??
Jacket density, or lead core density. What it one bullet is marginally harder or softer than the next ? One engraves easier in the rifling, altering velo and pressures ?
What if one has a longer driving surface than the next ?

Get a brand new bullet, roll it on paper a bit, and you get a duller surface, showing exactly how long the parallel drive band is. Do that to 10 bullets.. are they all identical length, or do they vary by 0.1mm ? Does that have an effect on engraving force, or engraving force duration ?

It gets to a point that variables are out of control, and we will never see a 25 shot string of 3 SD, 5 ES and a 0.3moa / 0.1mil group.

There.. i said it. Yes i expect to be wrong.. or right. No idea.
That was another factor that Hornady was considering that I have not seen addressed elsewhere but I don't know everything, by quite a long shot. Essentially, the center of gravity of a bullet. I call it mass distribution. If the material of the bullet is not uniform throughout, then, the moment arm would seem "lop-sided." That would introduce some wobble, I think.
 
Last edited:
Thats above my pay grade.
However from a physics and engineering stand point, its quite possible that bullet construction variation, as small as it is, could be a strong factor in ES and SD variations.

How a tuner would effect velo, i have no idea. If it even does or its snake oil...
 
That was another factor that Hornady was considering that I have not seen addressed elsewhere but I don't everything, by quite a long shot. Essentially, the center of gravity of a bullet. I call it mass distribution. If the material of the bullet is not uniform throughout, then, the moment arm would seem "lop-sided." That would introduce some wobble, I think.
Welcome to 1909. Franklin Mann's The Bullet's Flight from Powder to Target is an excellent read. He wasn't 100% correct on everything, but what he did with the resources available was incredible.
 
  • Like
Reactions: Ronws
To be fair, it’s hard to capture a non-linear relationship with only a couple settings. Say someone tried only two settings and the group sizes are exactly the same, then that’s a flat line. But if you try three settings, the the first and last setting is the same result but the middle setting is either better or worse than the other two, the response is non-linear (assuming all else was equal). Now let’s pretend that many different settings produce results that are not the same. Some better, some worse.

Let’s pretend that the tuner is as sensitive as people say they are where very small adjustments have an impact. The number of combinations that would be required to test all those settings to get a statistically valid conclusion would be so damn large and expensive, it would not be worth even testing it. At worst, it wouldn’t be repeatable from one rifle to the next. So, you would have to spend all that time and money on the new rifle because your results from the last test would not apply.

I like testing things, but I’m not about to dump $1000+ just to get a statistically valid test just to maybe find the right tuner setting just in time for the barrel to be shot out.

And if it's impossible to adequately test a tuner without large sample sizes, wouldn't it also be safe to assume that the same is true for an individuals ability to find an optimal tuner setting with same said tuner?

This is an argument that I don't understand. That tuners are so impossibly hard to adequately test to determine if/how they work, but are also simultaneously so easy to use, that one only needs to do 2 shot test at each setting to determine their optimal setting.

I can't reconcile those two thought processes.
 
What you just said not only hurts and helps both causes of for or against the use of tuners.

When someone frames the experiment, they have to frame it from the aspect of the group size given the tuner setting. So we need to define what a group is? Is it 5 shots, 10 shots, 30 shots, etc.? To get an accurate representation of the true group size, it has to be much larger 5 shots. 30 shots has been tossed around so let’s use that.

Moving on from the definition of the number of shots per group, a single group is only a sample of 1. Meaning, the group shot is a sample size of one. So, at any particular setting we need more than one group. We’ll pretend we did the analysis and it is 5 different groups per tuner setting.
How many tuner settings? 5, 10, 20?? Let’s say 10 different tuner settings.

So, 30 shot per group x 5 groups per tuner setting x 10 tuner settings. That’s 1500 rounds for a single rifle. That’s assuming my assumptions are correct and it could be much worse like 2000+ rounds needed for a single rifle.

Now, what about a rifle with this barrel or that barrel, or this tuner or that tuner, a barrel this long or that long, or a BR rifle or a PRS rifle, etc.

The test becomes so damn expensive that no one will want to sink the money into it.

Edit:
Anyone who wants to conduct this test, I will freely offer statistical consulting and provide a detail report of the experiment and its design. If you want it publication ready, I can do that as well.
Someone was willing to pay someone else to do all that testing. I wonder what ever happened with that…?
 
  • Like
Reactions: kthomas
And if it's impossible to adequately test a tuner without large sample sizes, wouldn't it also be safe to assume that the same is true for an individuals ability to find an optimal tuner setting with same said tuner?

This is an argument that I don't understand. That tuners are so impossibly hard to adequately test to determine if/how they work, but are also simultaneously so easy to use, that one only needs to do 2 shot test at each setting to determine their optimal setting.

I can't reconcile those two thought processes.
I would think so. I reserve to change my opinion in light of evidence. But, just trying to design an experiment to account for all the different settings and any factors that need controlled or blocked is going to take a lot of observations.
 
  • Like
Reactions: Ronws and kthomas
Yep. It’s a linear relationship. The only thing I’m not entirely certain of is if the powder and primer behave differently at different temperature extremes. That’s important because the variance might change across charge weights in an uncontrollable manner. I have experience near frozen ammunition cause wide dispersions and I assume it’s because the powder isn’t burning up completely.
 
  • Like
Reactions: Tokay444
Someone was willing to pay someone else to do all that testing. I wonder what ever happened with that…?
I’m not sure. A member reached out to me. We’re in the brainstorming phase but if they’re serious, something will be happening and as far as I can tell, they’re taking the human out of the element which will help a lot; even if it’s not practical.
 
  • Like
Reactions: Ronws and Tokay444
I would think so. I reserve to change my opinion in light of evidence. But, just trying to design an experiment to account for all the different settings and any factors that need controlled or blocked is going to take a lot of observations.

I agree.

Which is why I always thought the quick tuner tests (and stated advantages of a tuner - such as skipping load development) was pure snake oil.

I'm not saying they don't "work". There's just an incredible amount of inconsistencies from the tuner advocates, many of which aren't really reconcilable.
 
My understanding of Brian’s tests is that they would better the groups verses without and changes with the tuner made groups worse therefore the tuner effected the group size . Their whole premise was that since tuners did not better the groups then they did not work . But they did change with adjustments .
I don’t think you understood their results.
 
All we need is for someone to come forward with a gun that benefits from a tuner to shoot ten 3 shot groups so they can be plotted x & y and then do the same from the worst setting they identified during their two shot tuner adjustment test. Figure the mean radius of best and worst from the POA and we should have an answer, correct?
 
EDIT:

I made an error and own up to it. I'll try update this.

Earlier when I mentioned the probabilities of seeing a specific sample STD or less, I edited the comment to remove it. I made a mistake in the formula. I forgot to square a variable..... Totally wrong probability was told. I am owning up to it. I'll try to give a better follow up with say a confidence interval or something.
 
All we need is for someone to come forward with a gun that benefits from a tuner to shoot ten 3 shot groups so they can be plotted x & y and then do the same from the worst setting they identified during their two shot tuner adjustment test. Figure the mean radius of best and worst from the POA and we should have an answer, correct?
Except, someone else, or a machine rest needs to do the shooting.
 
Earlier when I mentioned the probabilities of seeing a specific sample STD or less, I edited the comment to remove it. I made a mistake in the formula. I forgot to square a variable..... Totally wrong probability was told. I am owning up to it. I'll try to give a better follow up with say a confidence interval or something.
Updated for those who care. Definitely made a big mistake, but maybe the updated post shines a bigger light on small sample sizes and sample STDs.
 
  • Like
Reactions: brianf
So, weird thought I had while back at shot.

If you take a stick and hold it straight out, the tip is going to wobble around, because no one is perfectly steady.

If you add a fixed weight to the far end of the stick, it's going to wobble around, but move slower.

If your wobble zone moves slower, I'd think you could shoot smaller groups.

Unrelated, I used to have a pdf about lot testing for the Olympics with an accelerometer. Trying to dig it up.
 
So here's a fun one I did recently. I sorted through roughly 1500 6mm bullets to finally get 30x of each on the high and low end of weight variation (about a 0.5gr total spread). And if you figure my average velocity in my 6mm ARC is ~2750 and that because the KE equation is 1/2mv^2 the mass term should be pretty linear and you do KE1 = KE2 (assuming the powder charge [chemical potential energy] being the same will result in the same final KE) with the two different masses (108.7 and 109.2, for example), you should see about 7-8fps change in the average of those weight sorted extremes.

30 shot strings:
108.7- 2735.6fps
109.2- 2735.2fps

ES within 3fps, SD well within 1 fps between the two.

So big picture takeaway, the transition of chemical potential energy into KE isn't 100% efficient, and there are many factors included as to what makes the peak pressure happen, the duration of pressure, and the sum of all of it that finally results in the muzzle exit velocity of a particular shot. At least in this test, the effect of weight variation is being grossly out-weighed by some other factor(s). Unfortunately I didn't think to measure base-to-ogive on these before I shot them but I still have a bunch of the rest of them I could go through and measure to see if there's any trend there.

At some point, however, things happen that are out of our control on a scale that's not feasible to control. At some point the scale to which a tuner "must need to be" to be "in tune" slips inside of variation of the forcing function we see from P&V results and that's where my level of skepticism about the whole thing comes from.

On aliasing...
The testing that Jayden did puts a pretty big damper on the whole "shoot 2, if they're not touching adjust the tuner...." until all 3 touch. We kind of suspected that going into it. The 3-shot strings do not correlate to the 20-50 shot strings (shocker). So maybe we aliased and skipped over the best settings, but what are the odds that we did it exactly 3-5 times per barrel to achieve pretty much the same results within the error brackets of what a 20-50 shot test will repeat to with/without/at different settings?

As someone else pointed out, to hit every setting with statistically confident results would take someone weeks of dedicated range use and several thousand rounds. To anyone saying we did it wrong, the world is just lying there waiting for you to do it better and share your results.

Again, I'm not saying that muzzle devices don't play a part in dispersion. I have been conducting a series of tests with this and in bolt guns, gas guns, the accuracy fixture I can document different levels of dispersion with different muzzle devices and that's repeatable. I just don't know that the reason why that happens if because of the weight/vibrations/mode shapes or muzzle exit gas flow, both, neither, something else.... Still digging into it.
Have you ever thought to do a correlation matrix on the different components? Like if you ran an experiment where you have 30 shots with one variable changed, then all the rest the same, then pick a different variable and change that, and keep doing that till you’ve covered all, or most of the variables you want to cover you could create a correlation matrix that shows how much a certain variable contributes to velocity. If you put your data in excel, you can easily load in pandas with .read_xlsx(), then run .corr() on the data frame To generate that matrix
 
So, weird thought I had while back at shot.

If you take a stick and hold it straight out, the tip is going to wobble around, because no one is perfectly steady.

If you add a fixed weight to the far end of the stick, it's going to wobble around, but move slower.

If your wobble zone moves slower, I'd think you could shoot smaller groups.

Unrelated, I used to have a pdf about lot testing for the Olympics with an accelerometer. Trying to dig it up.
Found the article, leaves a lot of questions
 
That was another factor that Hornady was considering that I have not seen addressed elsewhere but I don't know everything, by quite a long shot. Essentially, the center of gravity of a bullet. I call it mass distribution. If the material of the bullet is not uniform throughout, then, the moment arm would seem "lop-sided." That would introduce some wobble, I think.
You don't generally see that addressed elsewhere because it's incredibly difficult to properly measure, at least not without having all the variables and parameters available to you. Jeff Siewert mentioned this in some of his information. Imagine you have an X, Y, Z earth coordinate system. Z is your horizontal (to the right), Y is elevation straight up, and X is looking down range from the base of the bullet in the direction of fire.

Where the moment arm stuff comes into play is with a second coordinate system that is not earth fixed, but instead fixed to the projectile, we'll use i, j, k. Axis "i" runs through the projectile CG (assuming it's has perfect symmetry), from the nose of the projectile to the base. This is the polar moment of inertia. The other two axis, j and k, run perpendicular to the i axis through the CG, and are known as the transverse moments of inertia.

Here's where your point is important - if the CG of the projectile is offset some number from the axis of symmetry, then you now have an arm running through either the j or k transverse moment of inertia, perpendicular to the polar moment of inertia. That arm now wants to cause the projectile to "spin" within its current state of spinning (if that makes sense). This mass asymmetry (you called it mass distribution) is what leads to lateral throw off, and if I recall correctly, is the largest contributor to what Jeff Siewert described as a "cross velocity". With a pile more math and so on, you would be able to visualize the projectile motion (along the i axis)as a sine wave, depicting the frequency at which the projectile is wobbling.

Remember too, the motion of the projectile is "reasonably" constrained within the the bore of the firearm as it travels down to the muzzle. It's still not technically rigidly constrained in the bore, but significantly more-so than when it exits the muzzle. When it exits the muzzle with an offset CG, and it's all of a sudden not constrained to the bore, it can do and I say this broadly, whatever it wants.
 
  • Like
Reactions: Ronws
Status
Not open for further replies.