Myths and Mysticism in Load Development

[DingoDeerHunter]

A mob called "Blackburn Defense" has started a YT channel delving into the issue of what testing can show results between performance of variables in shooting.

It has verified what I first thought about when I got into this sport and saw all the videos and blogs suggesting "0.2gn increments, 5 shot groups, finding a node, testing 3thou intervals from the Lands, Barrel Tuners, harmonics etc". I thought "how can such small samples with so many variables at play (not the least being wind, and the execution of each shot by the shooter) produce any meaninfully informative results"? Turns out they actually can't.

The statistical models by these guys show you need thousands of rounds of data (by which time your barrel is worn out) before you get statistically significan results from these small variables to show small improvements. If you test a cheap rifle using cheap ammo against a Fclass rifle with consistent ammo, sure you only need a small sample to see the difference, but that is because the difference is large. However, if you want to see if a small powder charge increase/decrease produces better results then you need thousands of rounds of data. Barrel tuners, same thing - you can't get enough data to know.

I know heaps of guys will swear black and blue that their dozens of rounds of load development testing shows real results (as the youtube commenters have screamed in response to the Blackburn Defense video) but Science doesn't lie, people lie to themselves. It's the placebo effect. If you doubt it, then do 3 x exact replicas of your last "load development test" and compare the results, you will find out the awful truth I suspect.

Stop wasting compenents seems to be the answer science delivers and work on technique and wind reading. All I am going to do in load development is find the safest highest velocity I can (because the transonic zone is a real thing that effects bullet stability) and practice, practice, practice.

T

[Weairy]

Whilst I appreciate the thoughts, I can tell you I’ve got sheets and sheets of A3 card there which will illustrate the effect 0.1gn of powder increments and 5thou depths, as well as barrel tuner adjustments, will make to groups. And it’s significant in F-Class terms.

My 223 goes from a clover-leaf to tea cup over 0.5gn of powders, and even further with 10thou jump increments. And taking the developed load down to zero scopes etc, I can tell you, it replicates those results at 300yd.

[Barossa_222]

A mob called "Blackburn Defense" has started a YT channel delving into the issue of what testing can show results between performance of variables in shooting.

It has verified what I first thought about when I got into this sport and saw all the videos and blogs suggesting "0.2gn increments, 5 shot groups, finding a node, testing 3thou intervals from the Lands, Barrel Tuners, harmonics etc". I thought "how can such small samples with so many variables at play (not the least being wind, and the execution of each shot by the shooter) produce any meaninfully informative results"? Turns out they actually can't.

The statistical models by these guys show you need thousands of rounds of data (by which time your barrel is worn out) before you get statistically significan results from these small variables to show small improvements. If you test a cheap rifle using cheap ammo against a Fclass rifle with consistent ammo, sure you only need a small sample to see the difference, but that is because the difference is large. However, if you want to see if a small powder charge increase/decrease produces better results then you need thousands of rounds of data. Barrel tuners, same thing - you can't get enough data to know.

I know heaps of guys will swear black and blue that their dozens of rounds of load development testing shows real results (as the youtube commenters have screamed in response to the Blackburn Defense video) but Science doesn't lie, people lie to themselves. It's the placebo effect. If you doubt it, then do 3 x exact replicas of your last "load development test" and compare the results, you will find out the awful truth I suspect.

Stop wasting compenents seems to be the answer science delivers and work on technique and wind reading. All I am going to do in load development is find the safest highest velocity I can (because the transonic zone is a real thing that effects bullet stability) and practice, practice, practice.

T

You need to know how measure the data. Need to know what to look for. You're saying that 0.2gr will do nothing. This can be true and false. If you're in the middle of your node everything else considered equal, 0.2 will probably do nothing. If, however you're at either end, 0.2 might push you out and coupled with temperature and fouling issues will make your load uncompetitive. I'm not an expert on anything, but I know what it takes to shoot and load for accuracy out to 1200 yards. A good solid reloading regime paying attention to the right things will get you further along when coupled with that practice and wind reading you mentioned. You can practice as much as you like with a gun that shoots 5 ring, but at the end of the day it's still a 5 ring gun. Wish you the best.

[tachyon]

About the only useful take home from this Blackburn Defence video https://www.youtube.com/watch?v=JSxr9AHER_s is that bullet dispersion looks more like an M pattern (Chi distribution) than a normal distribution.

Screen Shot 2023-06-15 at 10.17.45 pm.png

Let's ignore the fact this guy appears to be an abysmal marksman with 100 yard groups in the 4-8" range and his inability to hold centre greatly amplifies the M pattern. That would be playing the man, not the ball.

Let's look at the task required for precision shooting because it seems neither of you understand it.

The task of load development is not to find the smallest group, rather, it is to find (and AVOID) the large groups.

Whilst it is correct to say you need quite a large number of shots to statistically prove a group is small, you DO NOT require a large number of shots to prove a group is large.

Consider a 5 shot group. Let's say the first shot falls on the left side of our distribution

The probability of shot 2 falling on the left is 50:50 or 1/2
The probability of shot 3 ALSO falling on the left is 1/2 * 1/2 = 1/4
The probability of shot 4 ALSO falling on the left is 1/2 * 1/2 * 1/2 = 1/8
The probability of shot 5 ALSO falling on the left is 1/2 * 1/2 * 1/2 = 1/16

So, 1 time out of 16 you will get all the bullets clustered in 1/2 the total distribution width ie an abnormally small group.

The other 15 times at least one bullet will land on the right side and open up the group. Furthermore, on average bullets will land around both the left and right distribution peaks leading to consistently large groups when a rifle is out of tune....

So, it is actually quite easy to see if a group is large, and have a high degree of statistical certainty about that showing up within 5 shots.

You may find this of interest, because it shows you what a node looks like, and how it was found using 35 rounds.

viewtopic.php?f=5&t=15348&p=110889

69994263094__902AC7C2-8F55-4C32-AA85-E103A3091233 2.jpeg

That powder charge and seating depth held 1/2 MOA (X ring) vertical for the full 130 rounds of the Kings at ranges from 500-1000 yards allowing me to take home the win. The seating depth (12 thou jam) was chosen because it was as far away as possible from the large groups....

[Brad]

Wish you all the best with your approach irt to load development but I’ll say it and if I’m wrong I’m wrong but you’re going to struggle to be competitive….

[DingoDeerHunter]

About the only useful take home from this Blackburn Defence video https://www.youtube.com/watch?v=JSxr9AHER_s is that bullet dispersion looks more like an M pattern (Chi distribution) than a normal distribution.

Screen Shot 2023-06-15 at 10.17.45 pm.png

Let's ignore the fact this guy appears to be an abysmal marksman with 100 yard groups in the 4-8" range and his inability to hold centre greatly amplifies the M pattern. That would be playing the man, not the ball.

Let's look at the task required for precision shooting because it seems neither of you understand it.

The task of load development is not to find the smallest group, rather, it is to find (and AVOID) the large groups.

Whilst it is correct to say you need quite a large number of shots to statistically prove a group is small, you DO NOT require a large number of shots to prove a group is large.

Consider a 5 shot group. Let's say the first shot falls on the left side of our distribution

The probability of shot 2 falling on the left is 50:50 or 1/2
The probability of shot 3 ALSO falling on the left is 1/2 * 1/2 = 1/4
The probability of shot 4 ALSO falling on the left is 1/2 * 1/2 * 1/2 = 1/8
The probability of shot 5 ALSO falling on the left is 1/2 * 1/2 * 1/2 = 1/16

So, 1 time out of 16 you will get all the bullets clustered in 1/2 the total distribution width ie an abnormally small group.

The other 15 times at least one bullet will land on the right side and open up the group. Furthermore, on average bullets will land around both the left and right distribution peaks leading to consistently large groups when a rifle is out of tune....

So, it is actually quite easy to see if a group is large, and have a high degree of statistical certainty about that showing up within 5 shots.

You may find this of interest, because it shows you what a node looks like, and how it was found using 35 rounds.

viewtopic.php?f=5&t=15348&p=110889

69994263094__902AC7C2-8F55-4C32-AA85-E103A3091233 2.jpeg

That powder charge and seating depth held 1/2 MOA (X ring) vertical for the full 130 rounds of the Kings at ranges from 500-1000 yards allowing me to take home the win. The seating depth (12 thou jam) was chosen because it was as far away as possible from the large groups....

I’m sorry you don’t understand the significance of the fact that the model is not a normal distribution and that you believe anecdotal evidence is significant.

This is like a kind of religion, and no amount of evidence will persuade y’all.

Be safe and enjoy

[DingoDeerHunter]

Wish you all the best with your approach irt to load development but I’ll say it and if I’m wrong I’m wrong but you’re going to struggle to be competitive….

I won’t be competitive because I can’t read the wind! Maybe in a couple of years I will have some clue and start to think about subtle load issues.

The point made by the video guys wasn’t that these things don’t make a difference but rather that outside of large effects small samples can’t tell you.

Has anyone actually done a double blind verification of their own testing? Something like the following:

You load up your various samples in triplicate. You give them to another person who hands them to you in no particular order and you don’t know what sample you are shooting. You do this 3 times throughout the day, then results are plotted with no identifier and you pick the best one in each 3 lots without knowing what they are - then pull back the curtain to see if you got the same answer each time?

What chance do you really think there is that the same result will show itself in the 3 runs with 5 shot groups? Pretty much zero is what the stats guys have shown. Unless you’ve done this type of unbiased verification it’s all just Polly waffle ain’t it?

[DingoDeerHunter]

Whilst I appreciate the thoughts, I can tell you I’ve got sheets and sheets of A3 card there which will illustrate the effect 0.1gn of powder increments and 5thou depths, as well as barrel tuner adjustments, will make to groups. And it’s significant in F-Class terms.

My 223 goes from a clover-leaf to tea cup over 0.5gn of powders, and even further with 10thou jump increments. And taking the developed load down to zero scopes etc, I can tell you, it replicates those results at 300yd.

A man was told by a scientist that astrology isn’t real, he stubbornly answered “but I checked my horoscope yesterday and it was so accurate!”

[PeteFox]

The task of load development is not to find the smallest group, rather, it is to find (and AVOID) the large groups.

Whilst it is correct to say you need quite a large number of shots to statistically prove a group is small, you DO NOT require a large number of shots to prove a group is large.
...

James,
this is only meaningful if the shot(s) that make the group large are representative of other groups shot under the same conditions. That is, there is a high confidence that the other groups will be similar.
But if you are discounting a load because of one large group, is it not subject to the same errors as cherry picking a small group without further testing.
What is the rationale for discounting the possibility that the large group is as representative of the abilities of the rile as the small one?

The whole problem with this sort of discussion is that no one seems to accept the existence of statistical outliers. I wiould have thought that reducing the frequency of outliers was the aim here.
Pete

[Weairy]

Whilst I appreciate the thoughts, I can tell you I’ve got sheets and sheets of A3 card there which will illustrate the effect 0.1gn of powder increments and 5thou depths, as well as barrel tuner adjustments, will make to groups. And it’s significant in F-Class terms.

My 223 goes from a clover-leaf to tea cup over 0.5gn of powders, and even further with 10thou jump increments. And taking the developed load down to zero scopes etc, I can tell you, it replicates those results at 300yd.

A man was told by a scientist that astrology isn’t real, he stubbornly answered “but I checked my horoscope yesterday and it was so accurate!”

A man watched a YouTube video that backed up his own thoughts and treated it as gospel too. You do it your way, I’ll do it mine, at least I know mine works.

[MarkS]

I find it deeply concerning that you suggest the advise we get from You Tube isn't always totally awesome Weairy.

Also, I have found a statistical significance in the fact when I'm doing load development the single round groups are always the tightest.
Just sayin

[Weairy]

I find it deeply concerning that you suggest the advise we get from You Tube isn't always totally awesome Weairy.

I know, absolute blasphemy right?

Would it also blow your mind to know that just because it’s on Facebook, it doesn’t make it true?

[DingoDeerHunter]

A man watched a YouTube video that backed up his own thoughts and treated it as gospel too. You do it your way, I’ll do it mine, at least I know mine works.

touché!!!

PS. lt’s a great area of the sport to interrogate, there’s a billion methods promulgated, each with devotees who swear by it. I’m definitely attracted to the framework that brings logic and maths to such a complex system as ballistics. There was another rigorous investigation that tore into the issue of tuning barrel harmonics with science and found the evidence wanting. Reminded me of Peter Brock and the Polarizer.

Happy target hunting all! Good to see some enthusiasm on the forum.

T

[johno]

Wow, and just like that he was gone. Fell clean off the edge of the flat earth.

[tachyon]

There was another rigorous investigation that tore into the issue of tuning barrel harmonics with science and found the evidence wanting.

Apparently you are not familiar with the finite element analysis work of Al Harral, aka Varmint Al, for example:

https://www.varmintal.com/aeste.htm

Who is Al? The retired Group leader of the Advanced Engineering Analysis Group at Laurence Livermore National Laboratory in the USA.

https://www.llnl.gov/news/retiree%E2%80 ... -notoriety

If you're not familiar with LLNL they are most famous for designing Americas nuclear bombs...