Thoughts About the Satterlee Method of Load Development

[denton]

The short version: I have no confidence in it.

The long version:

The basic idea of the Satterlee Method is that you load 10 cartridges with equal increments of powder charge. You then fire them, looking for a flat spot in the MV vs. powder charge curve. (Not position on a target, like the Audette Ladder Method) That flat spot is where you're you will get minimum MV variation, and where you are likely to find your most accurate load.

What is wrong with this? Let me count the ways......

One of my references is this video.



If you plot the data presented in the video, you get this scatterplot:



And sure enough, it looks like we have a flat spot at Point 7.

Now getting low variation in your MVs is a good thing, but there are big problems.....

1. I ran the stats on the data set, and none of the points are detectably statistically different from the trend line. Some points are above the trend line, and some are below. As far as we can tell, that is all explained by random variation. That means that if we repeat the experiment, the flat point will only appear again if we are lucky. It's not repeatable.

2. Mean (central location) and variation (standard deviation, range) are independent of each other. You can't predict one from the other. I can find no reason to believe that the standard deviation of the MV is any different at Point 7. From experience, I would guess that you need two samples of ~100 shots each, taken at Point 7 and one of the adjacent points to demonstrate that the variation of Point 7 is any different from 6 or 8 or any other point. Small changes in variation are notoriously hard to pin down.

3. I have run many tests of MV and pressure as a function of powder charge. I've read the writings of Dr. Brownell (great series of well researched articles in the 60s), and I've analyzed data from Ken Oehler. I have never, ever found a real (not random variation) flat spot, the kind that Satterlee fans seek.

So can you get a real flat spot? I can think of two ways:

1. AA2520 and AA4350 both start to plateau near maximum charges. MV variation increases in that plateau, contrary to what we'd like to have. I think it's from unburned powder.

2. A small change in illumination will have an effect on the readings of an optical chronograph. So if you have a cloudy day, and you get a spot where the clouds are thicker or thinner, you'll get slightly different readings. I noticed this 20 years ago, and contrary to most, did my chronographing only on sunny days until I switched to LabRadar, which doesn't have that problem.

So can you find an optimum target load if you find something like Point 7? Sure. But as far as I can determine it's not likely to better or worse than any other point in the string.

If you want to argue that it does work, let me indicate that to evaluate a claimed difference of 12% in group size, you'll need to compare about a dozen 5-shot groups from both the test and control batches. Otherwise, sadly, you're probably only comparing one random number with another.

 

[jmorris]

I think a lot of people just try and do things people that win gun games are doing. Every now and then, someone comes up with a new and better way to do things and start winning and people shift. Not unlike folks fishing.

Then there are people that don’t win and still think their way is best, even if it’s just best for them.

Life is too short to argue.

 

[thump_rrr]

Here is a ladder I shot on Monday.
Can you determine one way or another with the data presented if it is valid.
I forgot to turn on the Chronograph on the first shot.
This is why there are only 2 velocities at position #1.

 

[JFrank]

Maybe your rifle didn’t read that article. I’ve used one shot per charge as a pressure ladder as well as the perverbial flat spot (overlap) indicator.
I’ve moved to different methods of viewing test rounds, regardless they do get the job done.

 

[JFrank]

Here is a ladder I shot on Monday.
Can you determine one way or another with the data presented if it is valid.
I forgot to turn on the Chronograph on the first shot.
This is why there are only 2 velocities at position #1.
View attachment 1145563

Your’s is interesting to me, there’s not a lot of movement of group centers along the line and that’s what this style is about, I also see inconsistent grouping ( spitting shots) for whatever reason. Why so much annealing ?

 

[barnfrog]

View attachment 1145581 Maybe your rifle didn’t read that article. I’ve used one shot per charge as a pressure ladder as well as the perverbial flat spot (overlap) indicator.
I’ve moved to different methods of viewing test rounds, regardless they do get the job done.

In your example, where would you say the flat spot is? 6-7-8?

 

[AJC1]

I agree that a single data point per charge is not meaningful to me. Additionally I have watched a monster pile of testing videos and the primer test show more impact on flats and curve shape than powder. Three shot data points aren't mathematically supported any better, but if they support a trend I'm much more inclined to believe that data set.

 

[Howa 9700]

I come from a background where my job for past 40 years or so was one research project after another, where one aspect was to make maximum use of limited data. Out of a nearly infinite number of variables, which one's mattered. In time, came to realize very few of them actually did. Find what did and focus on that.

Having tried the satterlee method, have come to realize there is the simple concept of the 10 shot incremental ladder, then there is the rest of the story, which is he is running these thru high end custom guns and already has an idea of what is to come. Mostly is confirming what he already knows. He isn't running an off the rack gun using components he isn't familiar with (like I am). He is also using a high end chrony, not an entry level Caldwell (like I am). Beyond that, if you really wanted to try this, the best way would be to double down on the limited data. Shoot these over a chrony, but also at a long enough distance that similar velocities will print at the same drop level. Minimum 300 yards and 500 would be better. If chrony and target give you the same answer, you might be on to something.

There is another part of this that seldom gets mentioned, but if you dig deep enough you may find the part where he says the accuracy nodes for the rifle/bullet combo centers around a velocity node......and the same accuracy point shows up at the same velocity node, regardless of powder used. Again, these limited data are only to confirm what he already knows.

And modern stuff related to him has moved on to seating depth nodes to fine tune the load ladders.

For load development work for hunting ammunition, I'm now in the camp that is using a series of six incremental steps of 3 shot load ladders with half grain increment jumps, ending at the max load for bullet/powder combination. Goal is best accuracy that is close to factory ammo velocity levels. Anything sub MOA and cooking along at factory level velocity is all that is needed for my purposes. But a bench rest guy might look at that and throw up.

Have also come back to what I remember from work in that you have to eliminate all variables except for the one you are trying to measure. Unless you are able to do that, one is just burning thru powder, primers bullets and wearing out brass and barrels with no hope of finding anything useful other than by stumbling upon it by random chance.

 

[JFrank]

In your example, where would you say the flat spot is? 6-7-8?

There are three, 1-2 …6-7-8.. 9-10
Loading in the middle of those will for the most partView attachment 1145601 repeat on paper.

 

[JFrank]

Now days I use a slightly different ( Blaney)method, it looks like this.

 

[N555]

A flat spot in velocity is good but it isn't everything.
I use the chronograph for handgun load development and with rifle I use it after the load is developed to figure drop.

If he is annealing twice means he isn't getting them hot enough the first time and is looking for consistency.
I anneal them dead soft then size them.
It's the only way to be able to take once fired brass, range brass, multi fired brass with no annealing, multi fired with multiple annealings and bring them all back to close to factory neck tension.
If annealing after sizing, if you actually get the brass hot enough for annealing to occur then the neck is going to shrink then neck tension will really be all over the place.

 

[denton]

Here is a ladder I shot on Monday.
Can you determine one way or another with the data presented if it is valid.
I forgot to turn on the Chronograph on the first shot.
This is why there are only 2 velocities at position #1.
View attachment 1145563

I think we have discussion of two different methods going on.

The Satterlee Method does not use a target. It looks at chronograph data, MV as a function of powder charge. I don't think that works to find an accurate load, though it is very handy for other purposes.

The Audette Method looks at point of impact vs. powder charge on a target. That has a very valid physical basis, since you are trying to map out the nodes on your barrel whip. I think that probably works, and is probably more clever than most people realize.

I've made a couple of changes to the original post so as to be more clear about this.

 

[AJC1]

I come from a background where my job for past 40 years or so was one research project after another, where one aspect was to make maximum use of limited data. Out of a nearly infinite number of variables, which one's mattered. In time, came to realize very few of them actually did. Find what did and focus on that.

Having tried the satterlee method, have come to realize there is the simple concept of the 10 shot incremental ladder, then there is the rest of the story, which is he is running these thru high end custom guns and already has an idea of what is to come. Mostly is confirming what he already knows. He isn't running an off the rack gun using components he isn't familiar with (like I am). He is also using a high end chrony, not an entry level Caldwell (like I am). Beyond that, if you really wanted to try this, the best way would be to double down on the limited data. Shoot these over a chrony, but also at a long enough distance that similar velocities will print at the same drop level. Minimum 300 yards and 500 would be better. If chrony and target give you the same answer, you might be on to something.

There is another part of this that seldom gets mentioned, but if you dig deep enough you may find the part where he says the accuracy nodes for the rifle/bullet combo centers around a velocity node......and the same accuracy point shows up at the same velocity node, regardless of powder used. Again, these limited data are only to confirm what he already knows.

And modern stuff related to him has moved on to seating depth nodes to fine tune the load ladders.

For load development work for hunting ammunition, I'm now in the camp that is using a series of six incremental steps of 3 shot load ladders with half grain increment jumps, ending at the max load for bullet/powder combination. Goal is best accuracy that is close to factory ammo velocity levels. Anything sub MOA and cooking along at factory level velocity is all that is needed for my purposes. But a bench rest guy might look at that and throw up.

Have also come back to what I remember from work in that you have to eliminate all variables except for the one you are trying to measure. Unless you are able to do that, one is just burning thru powder, primers bullets and wearing out brass and barrels with no hope of finding anything useful other than by stumbling upon it by random chance.

Most of the guys doing a lot of that are just doing a barrel verification on a cartridge they have loaded for years. 284 win with the same barrel length same barrel maker same action. I do a lot of these tests to build a level of knowledge and experience but I will never be as refined as a guy competing because of volume and the things learned from fellow competitors....

 

[Krogen]

One data point does not a data set make! Satterlee is trying to glean information from one data point at each powder charge. Statisticians will argue, but probably 30 data points are needed to assess random variation. That would be at each powder charge. Other variables would have to be controlled as well to isolate random variations. I'd be interested to see if the same flat spot materialized with just 5 shots at each powder charge. My guess is that the range of variation for 5 shots at each powder charge completely obscures the flat spot conclusion.

 

[Kaldor]

I dont find the Satterlee to be perfect, but Ive done a few dozen Satterlee tests now and only very rarely does looking at velocity nodes not work.

There has to be some assumptions made to get an accurate test. You have to assume you are getting absolute accuracy in your SD and ES because you are only shooting one shot per charge weight. Whether you are loading one round or five or ten, you need to be assured the data is actually good and viable. I have found with my loading techniques and equipment if I am not getting single digit SDs, then something is really wrong.
Brass needs to be absolutely consistent, so forget about using mixed headstamp or even commercial brass for the most part. I like something like Lapua, Peterson, but Ive had good luck with Starline when buying in bulk boxes because very likely it was all run on a single day on the same machine.
Your powder drops need to be to the kernel accurate. For general loading, this is over kill, but for a test, it needs to happen for the argument of ES/SD. A couple kernels wont make much difference in general loading, but in a test it really does. This is also why we look for nodes so it gives the flexibility of less than dead on powder drops.
You need to have an accurate chronograph. Even an old school chrono will produce good results in the right conditions. I do own a LabRadar.
You have to use a powder charge that is small enough to actually see the nodes. Big jumps dont work. I use .2 for short action cartridges, and .3 for long action stuff. This gives you enough resolution in the test to actually see the data. Its also covers 2 or 3 grains of powder in a given load range which is fine for my usage.
You have to use powder that is stable enough to produce consistent accuracy. Ball powders need not apply unless you are only shooting in temperature controlled environments. You also have to understand that at some points it takes more powder to make velocity to increase that what one would think and this is what creates a node.

The shooting part is the easy part. Just shoot them into the berm and gather the data. I do take it a step further and I used to plot my string but now just use my Shotmarker because it saves time. I can look at the chrono data, and see if anything coincides on the target.
I ran in to this exact situation yesterday when testing 300 PRC with N555 and 220 SMKs. Shots 4, 5, and 6 were my node with a ES of 17, SD of 9.3 (velocities 2693, 2708, 2710 so 3 different charges at .3 jumps, and a wide .6 total node) and on the target 4, 5, and 6 stacked into a 1/2" group. Looks like a good place to start and Ive only expended 10 shots to find this data. I will probably concentrate more on the 2708 and 2710 node and dump the 2693.
Ive also run into situations where I have had the luxury of finding two nodes in a 10 shot Satterlee with .2 jumps in 6 Creed with the 107 SMK and H4350.
Best luck Ive ever had was with a 109 Berger/H4350 in my 6 Creed. Did a 10 shot Satterlee, found a decent node. When I tested with 10 shots, it shot right at a 1/2" on 2 different range days. 20 shots to find a fairly acceptable group that averages 1/2"? Yeah, Ill take that and I still shoot it to this day 4 years later.
I have had it not work either. My 6 Grendel is such a case. I see good nodes, I test them, they suck. So it doesnt work everywhere. I reverted back to 5 5 shot groups running the chrono for 5 shot ES/SD and looking at accuracy on the target and making a judgement call based on those.
I also dont use a Satterlee when testing ball powders with cheap bullets for the reason that they just arent not stable and consistent enough.

Its not a magic bullet but it does work, you just have to work within the confines of the data set and have alot of confidence the consistency of your reloading. And even then, its just a starting point. You still have to test your nodes, tune COAL, etc, which is no different than any other method. Even when it doesnt work, you can still glean useful info from the test like pressure signs which you can in turn use to define a standard 5 shot test for a given load range.

 

[Varminterror]

Here’s where theory and reality diverge.

I run triplicates of Satterlee tests before most matches I shoot, for 7 seasons now, finding the same node in the same cartridge. Pretty simple math - I shoot the same 8 charge weights 3 times per test, ~5-8 times per year, for 7 years… so over 100 shots at each charge weight… and the flat spots repeat…

So I will continue to put more trust in statistical analysis of my ACTUAL data, rather than your speculation about statistical analysis of imaginary data.

 

[Nature Boy]

Having tried the satterlee method, have come to realize there is the simple concept of the 10 shot incremental ladder, then there is the rest of the story, which is he is running these thru high end custom guns and already has an idea of what is to come. Mostly is confirming what he already knows. He isn't running an off the rack gun using components he isn't familiar with (like I am)

Satterlee method aside, this would apply to any load development method. Expectations have to be set based on the equipment being used. My pedestrian rifles will not match the accuracy potential of my custom rifles but they both have an accuracy node for a given bullet/powder/primer combination.

Have also come back to what I remember from work in that you have to eliminate all variables except for the one you are trying to measure. Unless you are able to do that, one is just burning thru powder, primers bullets and wearing out brass and barrels with no hope of finding anything useful other than by stumbling upon it by random chance.

I could not agree more.

Regardless of which load development method you prefer, fundamentally you got to have 3 thing:

1. A rifle that’s stable enough to net repeatable results throughout the load development process.

2. A reloading process and components that are capable of producing consistent ammo.

3. A shooter that’s steady enough to eliminate himself as a cause of variation in the process.

 

[Varminterror]

The Satterlee Method does not use a target. It looks at chronograph data, MV as a function of powder charge. I don't think that works to find an accurate load, though it is very handy for other purposes.

The Audette Method looks at point of impact vs. powder charge on a target. That has a very valid physical basis, since you are trying to map out the nodes on your barrel whip. I think that probably works, and is probably more clever than most people realize.

The irrational bias displayed here is absolutely steeped in irony.

In the context of a thread disavowing one method for supposed statistical invalidity, you’re favoring another method which has even greater variable count, and greater variability. POI based long range load development methods - which also DO work, as has been proven for several decades - are dependent upon velocity errors in the same way as the Satterlee method, including all of the independent variables contributing to velocity error, AND ON TOP OF THOSE velocity error contributors (which are isolated in the Satterlee method), POI based methods ALSO are sensitive to (and dependent upon) additional error contributors, including bullet BC inconsistency, shooter ability and mechanical POA errors, environmental/atmospheric condition variability, raw accuracy potential of the rifle, variable harmonic resonance response (variable "barrel whip," as you called it) to barrel time variation, etc.

So it's exceptionally untoward to make the claim made in the opening of this thread - that a proven method isn't statistically valid - but then claim a method which is dependent upon the same error contributors as the first, but then a dozen other MORE variable and MORE influential error sources. Either both are completely meaningless, or Satterlee Method has more merit than Audette method - but you can't get away with talking out of both sides of your mouth.

 

[JFrank]

Here’s where theory and reality diverge.

I run triplicates of Satterlee tests before most matches I shoot, for 7 seasons now, finding the same node in the same cartridge. Pretty simple math - I shoot the same 8 charge weights 3 times per test, ~5-8 times per year, for 7 years… so over 100 shots at each charge weight… and the flat spots repeat…

So I will continue to put more trust in statistical analysis of my ACTUAL data, rather than your speculation about statistical analysis of imaginary data.

I have no idea if this would produce results I could compete with or not, I’ve tried most others but heck I have trouble keeping track of which guy wrote which article.
Let’s see if I have it right;
Shatterlee, uses chronograph data to find and maintain tune.
Audette, uses the one shot per charge on paper.
New berry, ( ocw) has a horizontal foremat judging group placement.
Blaney, ( the one I use now) overlays groups using one point of aim at longer ranges.

 

[Varminterror]

I come from a background where my job for past 40 years or so was one research project after another, where one aspect was to make maximum use of limited data. Out of a nearly infinite number of variables, which one's mattered. In time, came to realize very few of them actually did. Find what did and focus on that.

What you've described here is known as "Sensitivity Analysis," and you're correct, understanding outcome sensitivity to different input errors is critical to process development and product improvement - which in our context, this means "making better ammo by better understanding of how to make better ammo." You'll see me communicate this principle on forums like this by talking about "big knobs" and "small knobs" - meaning turning one independent variable, one knob, has far greater influence than turning some others.

Having tried the satterlee method, have come to realize there is the simple concept of the 10 shot incremental ladder, then there is the rest of the story, which is he is running these thru high end custom guns and already has an idea of what is to come. Mostly is confirming what he already knows. He isn't running an off the rack gun using components he isn't familiar with (like I am). He is also using a high end chrony, not an entry level Caldwell (like I am).

I'll mention directly, that if you have experienced the Satterlee method does not work for common or less expensive gear, then you've had some anomalous result. There is nothing about the Satterlee Method which is sensitive to what quality of components or gear used, nor any sensitivity to shooter skill, nor any sensitivity to using unfamiliar components. Frankly, if you watch the video you actually posted yourself in your Satterlee 2.0 thread, at ~13:30 of the second video, Scott even mentioned he had used 3 Savage barrels along side 3 Proof Research barrels and got exactly the same results.

I will point out here, however, that POI based methods like Chris Long OBT (foundation of QL and GRT), Creighton Audette Ladder, and Dan Newberry OCW methods absolutely DO depend upon shooter skill and equipment quality. If we can't rely upon our target pictures (POA) to be exactly reliable, and if we can't rely upon our POI's to be exactly at the center of their potential variability, then we have experimental error which obscure the actual results - plainly, if you can't shoot worth a damn, or your rifle can't group worth a damn, then POI methods don't work - because the POI's will not represent statistically valid results for each shot, or even for each group. Say you shoot the Blaney method (which is a multi-shot group of the Audette method) and your groups of each charge weight are 8" at 600yrds, and the total height of your test is 10", you can't pretend the relative position of those group centroids are meaningful. BUT... There's nothing the shooter's skill can do to ruin the result of a velocity dependent test. There's nothing a rifle's raw precision can do to ruin the result of a velocity dependent test. But both of these CAN ruin POI dependent tests - and we know from generations of proven results, POI dependent methods like OCW, OBT, and Audette Ladders DO work for common gear and unfamiliar components - it's just harder to discern nodes from highly instable experimental results.

Beyond that, if you really wanted to try this, the best way would be to double down on the limited data. Shoot these over a chrony, but also at a long enough distance that similar velocities will print at the same drop level. Minimum 300 yards and 500 would be better. If chrony and target give you the same answer, you might be on to something.

I've said on this forum and others, countless times, that I've done exactly what you're describing, and verifiable results found via POI based methods DO coincide with verifiable results found via velocity methods. HOWEVER - it's very, very simple to get false output in POI based methods by shooting the test at distances too short. POI methods at under 300yrds can produce false results, and in these short distances, the velocity based method is actually more reliable. At long distances, where POI becomes more inherently dependent upon trajectory consistency and less dependent upon mechanical consistency (shooter POA and raw rifle precision), then we see alignment between the nodes found in the velocity based method with the nodes found in POI based methods. Congruency of outcome is good. But plainly, again, shooting POI based methods at short range is one way to introduce experimental error which taints the results sufficiently to undermine the outcome. Velocity curves can be shot at any distance, but if you want meaningful results, POI based methods MUST be fired at 300+, preferably 500+.

There is another part of this that seldom gets mentioned, but if you dig deep enough you may find the part where he says the accuracy nodes for the rifle/bullet combo centers around a velocity node......and the same accuracy point shows up at the same velocity node, regardless of powder used. Again, these limited data are only to confirm what he already knows.

I've been watching Scott and others using this method for a long time, and I think you're getting twisted somewhere - velocity nodes, as in "this cartridge shoots small at this velocity regardless of bullet or powder" or even "this bullet shoots small at this regardless of powder," are 100% snake oil. @South Prairie Jim posted a video recently in which world-leading marksmen and gunbuilders Erik Cortina and Alex Wheeler discussed that topic, directly, and dismissed the premise in just as few words.

And modern stuff related to him has moved on to seating depth nodes to fine tune the load ladders.

Be careful, it seems you're reading between the lines into something which isn't on the page. Mark Gordon (SAC) and Scott Satterlee's documentation out there which talk about seating depth nodes "first" are NOT refining to a razor edge, but rather have 1) disproven that we needed to chase exactly ONE seating depth, and 2) proven we can find huge forgiveness windows where good bullets can shoot well despite huge seating depth variabilities. Scott's method, by his own admission, takes HUGE jumps, beyond the BETTER statistically verified and MORE valid process used by Mark Gordon, but he is finding forgiveness in long jumps, largely as a means of minimizing sensitivity to that ONE variable. Moving farther out has less sensitivity than moving closer in. There's really nothing more to it than that. In a manner of speaking, circling back to the Sensitivity Analysis paradigm discussed above - all we're doing with most of these load development methods is trying to turn Big Knobs to a place on the dial where they behave like Small Knobs. For example - if I shoot 107 SMK's, I know I have much greater jump sensitivity than if I shoot 105 Hybrids, so seating depth is a bigger knob for the SMK's than the Hybrids. I also know if I seat Hybrids at 5 thou off of the lands, my loads are more sensitive to a few thou of seating depth error than if I seat them at 50thou. I ALSO know, if I load my 6 creed at 42.2grn H4350 in Hornady brass under 105H's, my velocity variability increases, especially more dependency of velocity error as a result of charge weight error, and my groups get taller, but loading at 41.8, my velocity variability becomes much, much less dependent upon charge weight error, and that vertical error shrinks. So all I have to do is turn that Big Knob to the right RANGE, not the right exact frequency, and the music will come through clearly, and turning it within that range won't change the clarity much at all. But alternatively, in the wrong range, a little turn to the big knob makes a huge change in signal clarity.

 

[Varminterror]

Audette, uses the one shot per charge on paper.
New berry, ( ocw) has a horizontal foremat judging group placement.
[Baney] ( the one I use now) overlays groups using one point of aim at longer ranges.

I don't really differentiate the 3 POI based methods you mentioned here - because they're not really different. The "Blaney Method" - as described in the 2010 article by Jason Baney at AccurateShooter - is simply a triplicate of the Creighton Audette method. Otherwise stated, it's just a Dan Newberry OCW method, but fired at the same POI instead of multiple POI's. Or the Audette method is just one string of a Baney process...

They're all analyzing relative vertical position of POI for adjacent charge weights - each "brand name" is just a different means of measuring the same thing. Repeating 3-5 Audette Ladders is a Baney test. Shooting a Newberry OCW on one point of aim is a Baney test, and vice versa... It's all the same thing. Hell, I was taught the Baney Method in 1998 by one of my mentors which called it the Audette method - and it wasn't until well into the 2000's which I finally ran across an online copy of one of Creighton's old articles, was it Handloader Magazine?, which described only 1 shot per charge weight. But I accepted that "doing it right, multiple times, wasn't the same as doing it wrong," so I've continued to do multiple shots and continued to call it the Audette Ladder. Would not a rose, by any other name...

***ETA - not so different, most guys I shoot with don't actually shoot ONE shot per charge weight on Satterlee tests either. I shoot 3 shots each, AND for that matter, I shoot either a short range OCW or a long range Baney/triple Audette with the same shots. Whether we shoot 1 shot or 3 shots, if I'm looking for flat spots in the velocity curve, I'm doing the same analysis whether I call it the Satterlee Method or the 6.5 Guys Method, or the Varminterror Method.

 

[JFrank]

I don't really differentiate the 3 POI based methods you mentioned here - because they're not really different. The "Blaney Method" - as described in the 2010 article by Jason Baney at AccurateShooter - is simply a triplicate of the Creighton Audette method. Otherwise stated, it's just a Dan Newberry OCW method, but fired at the same POI instead of multiple POI's. Or the Audette method is just one string of a Baney process...

They're all analyzing relative vertical position of POI for adjacent charge weights - each "brand name" is just a different means of measuring the same thing. Repeating 3-5 Audette Ladders is a Baney test. Shooting a Newberry OCW on one point of aim is a Baney test, and vice versa... It's all the same thing. Hell, I was taught the Baney Method in 1998 by one of my mentors which called it the Audette method - and it wasn't until well into the 2000's which I finally ran across an online copy of one of Creighton's old articles, was it Handloader Magazine?, which described only 1 shot per charge weight. But I accepted that "doing it right, multiple times, wasn't the same as doing it wrong," so I've continued to do multiple shots and continued to call it the Audette Ladder. Would not a rose, by any other name...

***ETA - not so different, most guys I shoot with don't actually shoot ONE shot per charge weight on Satterlee tests either. I shoot 3 shots each, AND for that matter, I shoot either a short range OCW or a long range Baney/triple Audette with the same shots. Whether we shoot 1 shot or 3 shots, if I'm looking for flat spots in the velocity curve, I'm doing the same analysis whether I call it the Satterlee Method or the 6.5 Guys Method, or the Varminterror Method.

Nicely said my friend.
I would agree, they are essentially the same. I think the Blaney is very easy to read and quite effective for mid to long range tuning, the last five years or so I’ve witnessed some of smallest groups the world has seen at 1000 yard Br. I’ve shot a couple myself and one thing guys need to remember is we shoot different games where in Br small ES doesn’t ensure small groups and small groups don’t mean much to prs shooters that need to calculate shooting solutions.
The Audette method IMHO is best used as a course powder/ pressure ladder. It gets to the meat and potatoes quickly and saves barrel life.

 

[Kaldor]

@Varminterror
Well said sir.

 

[denton]

The irrational bias displayed here is absolutely steeped in irony.

In the context of a thread disavowing one method for supposed statistical invalidity, you’re favoring another method which has even greater variable count, and greater variability. POI based long range load development methods - which also DO work, as has been proven for several decades - are dependent upon velocity errors in the same way as the Satterlee method, including all of the independent variables contributing to velocity error, AND ON TOP OF THOSE velocity error contributors (which are isolated in the Satterlee method), POI based methods ALSO are sensitive to (and dependent upon) additional error contributors, including bullet BC inconsistency, shooter ability and mechanical POA errors, environmental/atmospheric condition variability, raw accuracy potential of the rifle, variable harmonic resonance response (variable "barrel whip," as you called it) to barrel time variation, etc.

So it's exceptionally untoward to make the claim made in the opening of this thread - that a proven method isn't statistically valid - but then claim a method which is dependent upon the same error contributors as the first, but then a dozen other MORE variable and MORE influential error sources. Either both are completely meaningless, or Satterlee Method has more merit than Audette method - but you can't get away with talking out of both sides of your mouth.

Oh come on VT, tell us how you REALLY feel!

I'm happy to have my statements disproved, if they can be. That allows me to upgrade my understanding. But there are some important facts that you have missed:

Change in POI is much easier to reproducibly test that change in group size. For POI, you get to do a T Test, which converges quickly. For changes in variation you have to do F, Bartlett, or LeVene's Test, which do not. If the data fail the test, then the result is usually not reproducible. If it won't reproduce, it's not of much use.

Barrel whip is well known to be a first order factor in accuracy. Because of the peculiar way that variation adds, second order factors contribute only very little to total variation. You have to focus on the first order factors. MV is almost always a second order effect and the reasons why require a long discussion.

Given all that, chasing non-reproducible variations in MV, and comparing group sizes to test the outcome is a real uphill battle. That's Satterlee.

Chasing POI shifts to characterize barrel whip uses much more powerful tools, and focuses on a first order effect. That's Audette. And Audette was wise enough to require randomization, which washes out the effects of barrel temperature, neck tension, bullet mass variation, changes in illumination, etc.

Now the remaining question is, how many shots do you need to do for each Audette load? I don't know. I haven't run that math yet. But the Audette Method is on much firmer physical and statistical ground than Satterlee.

 

[Slamfire]

Making decisions on one shot or three shot groups, give me a break.


There is an excellent article at the end of the Oct 2014 Shooting Sports USA on group size and accuracy: Page 38 of http://www.nxtbook.com/nxtbooks/nra/ssusa_201410/ This foundational article was written by small bore prone competitors who wanted to shoot perfect scores. In small bore prone a Match is a 40 shot event of two twenty shot targets. The typical 1600 round Smallbore bore prone tournament is 160 rounds fired for record, divided up into four 40 round Matches. Therefore the referenced article assumes that a 40 round group is the baseline.

As anyone can see in table six, at least at 100 yards, a five shot group is 59% of the size of a 40 shot group, a 10 shot 74%, and a twenty shot 88%. A three shot group is below contempt, but three shot groups are the current standard for the shooting community because the leaders of the shooting community, that is in print Gunwriters, have convinced the shooting community that three shot groups are an exact measure of accuracy and consistency.

This is another good article on the limitations of five shot groups

Accuracy Testing: Shortcomings Of The Five-Shot Group

by Brad Miller, Ph.D. - Wednesday, September 25, 2019

https://www.ssusa.org/articles/2019/9/25/accuracy-testing-shortcomings-of-the-five-shot-group


one that addresses statistics for accuracy.

Shot Group Statistics for Small Arms Applications
https://apps.dtic.mil/sti/pdfs/AD1034885.pdf

People see patterns all the time where they don't exist. Which is why low shot groups are so popular. Three shot groups are the Rorschach test of the shooting community.You see what you want to see, and what you see is all in your head.