Thoughts About the Satterlee Method of Load Development

I think we just had this conversation’ three shot per increment tuning is quite effective and quite popular in the long range Benchrest community. LRBR shoot arguably the smallest groups in the world and I can assure you we don’t waste barrels with extra needless shots.
 
@denton - your claim that "variations in velocity" are non-reproducible is false, so you're building subsequent hypotheses on a broken foundation.

I'm teaching a new shooter how to shoot long range next month. I'll let him shoot an Audette Ladder with one of my rifles, with the LabRadar behind it - I'll let you do the data analysis to determine whether his 20" groups at 1000yrds have produced meaningful data - since, as you claim, "barrel whip is well known to be a first order factor in accuracy." Pretty easy to hide barrel whip in piss-poor shooting. On the other hand, his virgin hands won't be able to lie to the chronograph.

You'll also note - if you'd read the original articles by Creighton Audette himself - that what we call Satterlee method today was actually an aspect of the Audette method. Yes, Audette was wise.

Patently false to say barrel whip "washes out" (whatever you believe that could possibly mean) the influence of temperature and neck tension. It's well demonstrated that neck tension is a "big knob" and directly influences harmonic response - aka, barrel whip. And not for nothing - POI based method, again, rely on perfect mechanical POA, which is ALSO inherently dependent upon "changes in illumination," just as is an optical chronograph, while electromagnetic and doppler chronographs are not.

You're intentionally card stacking here, and ignoring proven ballistic science.

Both methods - really all of these methods SPJ and I have mentioned here - are proven to work by folks which are ACTUALLY shooting long distance. But it's bad science to try to convince people that an experimental method which is inherently sensitive to MORE sources of error, including the primary source of error in the other method, will have less error sensitivity than the other method. Carrying a cat, a dog, and a duck is not easier than carrying just a dog.
 
Last edited:
@denton - your claim that "variations in velocity" are non-reproducible is false, so you're building subsequent hypotheses on a broken foundation.

I'm teaching a new shooter how to shoot long range next month. I'll let him shoot an Audette Ladder with one of my rifles, with the LabRadar behind it - I'll let you do the data analysis to determine whether his 20" groups at 1000yrds have produced meaningful data - since, as you claim, "barrel whip is well known to be a first order factor in accuracy." Pretty easy to hide barrel whip in piss-poor shooting. On the other hand, his virgin hands won't be able to lie to the chronograph.

You'll also note - if you'd read the original articles by Creighton Audette himself - that what we call Satterlee method today was actually an aspect of the Audette method. Yes, Audette was wise.

Patently false to say barrel whip "washes out" (whatever you believe that could possibly mean) the influence of temperature and neck tension. It's well demonstrated that neck tension is a "big knob" and directly influences harmonic response - aka, barrel whip. And not for nothing - POI based method, again, rely on perfect mechanical POA, which is ALSO inherently dependent upon "changes in illumination," just as is an optical chronograph, while electromagnetic and doppler chronographs are not.

You're intentionally card stacking here, and ignoring proven ballistic science.

Both methods - really all of these methods SPJ and I have mentioned here - are proven to work by folks which are ACTUALLY shooting long distance. But it's bad science to try to convince people that an experimental method which is inherently sensitive to MORE sources of error, including the primary source of error in the other method, will have less error sensitivity than the other method. Carrying a cat, a dog, and a duck is not easier than carrying just a dog.

You have profoundly misunderstood and misstated what I wrote.

I have not said that variations in velocity are non-reproducible. I have said that non-reproducible variations in velocity are not useful in optimizing accuracy.

I have not said that barrel whip washes out the influences of temperature and neck tension. I have said that if you are conducting a test for barrel whip, it's wise to randomize the test so that the effects of temperature, neck tension, and other lurking variables wash out.
 
I have said that if you are conducting a test for barrel whip, it's wise to randomize the test so that the effects of temperature, neck tension, and other lurking variables wash out.

But you can't, because barrel whip is dependent upon neck tension... Pretty simple.
 
When I test prior to a long range match, I do so with flags I’m subject to poor weather skewing the data I get on the target, if you know what velocity your ride likes it makes it a lot simpler for sure.
I agree on neck tension, get that right and data much looks better.
 
I was watching this thread because I thought I might learn something, and I have. I learned that watching two statisticians debate a point is like watching two economists debate a point. Because my level of knowledge on either subject is so small, both (all) sides sound like they're right, leaving me more confused than before.

I'm out.
 
I was watching this thread because I thought I might learn something, and I have. I learned that watching two statisticians debate a point is like watching two economists debate a point. Because my level of knowledge on either subject is so small, both (all) sides sound like they're right, leaving me more confused than before.

I'm out.

What Denton has presented as favorable argument supporting one methodology is reliant upon the very argument he claims undermines the other method.

Moreover, his bias is illustrating his unfamiliarity with the tests in question, and their inherent error sensitivities. So his claim the Satterlee method - which again, is actually truly derived from Creighton Audette’s article and incorporates the same “randomization” which has been brought forth as a false strawman to distract the unwitting - is inherently statistically invalid while promoting the Audette method as valid is simply false.

Either both are valid or neither are, for the same weaknesses - but the LEAST error sensitive method certainly has more opportunity for validity than the one with the same sensitivity PLUS greater error sources, which in this thread, Denton has claimed backwards.

It sure sounded good to someone not familiar with experimental design, but what he’s doing is insisting error sensitivity is non-existent in group sizes - which is easily disproven by the fact we have groups at all, rather than holes.

Bottom line - error sensitivity being what it is: the typical Hunter, gun owner, shooter which is shooting common rifles and bullets which together can’t shoot sub-MOA groups at 100yrds, let alone the 600-1000yrds necessary for these tests will struggle to define nodes via POI based methods, much more than defining the node via velocity based methods. ESPECIALLY if they follow the original and most popularly communicated “Audette ladder” which only fires one shot per charge weight - but even discerning any trend within multi-shot versions of Audette ladders such as the Baney version will still only be about as reliable as reading tea leaves.
 
Last edited:
I hesitate to chime in here at all because I haven't done enough precision-oriented load development to have a justified opinion one way or another. I'm also hesitant to disagree with @Varminterror or certain others who appear to have done such testing and who shoot and reload at a level I do not. But . . .

The claim in the OP was that @denton has no confidence in the Satterlee 10-shot velocity ladder because the "flat spots" aren't repeatable. The guys over at Hornady made basically the same (if not broader) claim recently in their podcast series (you can find it on YouTube) stating in effect that in their testing of how-ever-many-stupid-number of rounds they shot in their controlled conditions (with admittedly limited bullet types and cartridges), they did not see material differences in velocity ES and SD over different charge weights. Quality components and reloading technique yielded low velocity ES and SD pretty well regardless of powder charge if you shoot enough at each different powder charge, or so they say.

My own limited testing with my rifles and reloads is inconclusive but suggests a velocity ladder with only one round per charge weight is of limited value when considered by itself. If you made me vote today on whether I think a one-shot-per-charge-weight velocity ladder reliably identifies a charge weight that produces less velocity variability than any other charge weight with that bullet and powder combination, I would vote no. I don't always see a "flat spot" repeat from one string to the next. In fact, in my experience, "flat spots" found with only 1 shot per charge weight fail to repeat so often that I won't conclude I've found a "flat spot" unless I've shot at least 2 shots per charge weight, and 3 is preferable (I'm not claiming 3 shots is statistically valid - just that 3 shots per charge weight gives me more data than 1). Sometimes, however, flat spots do appear to repeat (particularly those identified with more than 1 shot per charge weight), but I've never gone back and retested the "antinodes" or charge weights that "didn't work" to see if the difference I observed during load development actually holds up over time--I just go with what I see working because I have limited time and components. I still use velocity ladders for various purposes, but I don't do all of my load workup based on one shot per charge weight over a chronograph. FWIW, I only reload for production rifles and do not shoot competitions seriously, so if you're one of those guys and know better, I have no argument with you.

A final thing to keep in mind: Just because someone critiques your preferred load development method does not mean it doesn't work. The point is often that the method will not find a load that is objectively/repeatably better than all other loads for that rifle. But that doesn't mean the method won't repeatably find a "good load" that works for your rifle and your purpose---too many shooters have used each of these methods for a long time to develop stellar loads for many different purposes. They work. But just because they work doesn't necessarily mean they develop "the best" load for your rifle. Without shooting out the barrel doing nothing but testing, I don't know how you'd figure that out and prove it. The flip side of that coin is that methods which are not "statistically valid" do in fact produce "good loads"---the identified load may not be scientifically proven to be better than all other loads not selected, but the selected load is demonstrated to work. Use whatever method gives you confidence in your reloads and your rifles for your applications.

Addendum: my current working theory is that load development of most any method is better at telling you what doesn’t work than it is at identifying “the best” load.
 
Last edited:
The guys over at Hornady made basically the same (if not broader) claim recently in their podcast series (you can find it on YouTube) stating in effect that in their testing of how-ever-many-stupid-number of rounds they shot in their controlled conditions

I talked to Miles about that podcast (the dude with the mustache in the Hornady video you mention) at a PRS match two weeks ago and mentioned my ~100rnd per charge weight data sets, and he confirmed the meaning of his statement in the video - from barrel to barrel, we see node shifts and velocity shifts and Satterlee tests, when stacked over and over with multiple barrels, end up a broad swaths of linearity between velocity and charge weight - so when HORNADY LOADS FACTORY AMMO, there’s no reliable nodes and picking any charge weight will not be “in a node” for all barrels, nor result in the same velocity. The same can be said about Audette ladders - when we pile Audette ladders from multiple barrels, we’d just get one massive group. But we have irrefutable evidence that, within a given barrel (rather pertinent, since most of our rifles only have one barrel at a time), we can find repeating flat spots and predictable velocity shifts around sustained nodes.

If it were truly random error which is presenting flat spots as red herrings as Denton claimed, then no triplicate tests should offer the same node result, but when 30+ tests show repeating nodes, never failing to show the same flat spot with the same width and position, it’s pretty obviously not random error.

When I pile ALL of my different 6 creed barrels together, between different manufacturers, different reamers, and different lengths, I have a broad swath of apparent error around a linear trend, which is what Miles described in the Hornady podcast video. When I review each barrel alone, the linearity reliably, repeatably, and repeatedly breaks, even for high volume data sets which defeat the ignorant claim that “it’s too small of data set to be statistically valid.” Nobody argues that 30+ replicates is insufficient, because it’s not.

Because this simply isn’t true:

The claim in the OP was that @denton has no confidence in the Satterlee 10-shot velocity ladder because the "flat spots" aren't repeatable.

The flat spots, for an individual barrel, ARE repeatable. Not for 11 barrels of varying manufacturer, reamer, and length, but for each barrel, they are.

The same claim has been made about Audette ladders as well, which yielded the multi-shot Baney version and the Newberry OCW method. I’ve mentioned frequently here, I always shoot triplicates of my node confirmations to identify errant data, which is how I’ve established such a large database - I do the same test during load development AGAIN before almost every match all season, so it’s easy for me to catalog 20-30+ data points for a given barrel of each charge weight. As I mentioned, most guys I know do multiple Satterlee strings, randomizing in the same way as the Audette ladder, for the same reason Jason Baney and others suggest shooting multiple Audette Ladders instead of just 10 single shots per charge. The “6.5 Guys” can be (probably rightfully) attributed for making the Satterlee method famous several years ago, and they published their free spreadsheet to include multi-shot strings as “standard” within the application of the Satterlee method… sure, if someone has huge 70fps ES’s on a 3 shot replicant, the data is likely no more meaningful than the guy shooting 24” groups at 1000yrds trying to read an Audette/Baney ladder, but it’s just false to pretend velocity data isn’t repeatable and just as false to say velocity flat spots only occur due to random error.
 
Last edited:
I talked to Miles about that podcast (the dude with the mustache in the Hornady video you mention) at a PRS match two weeks ago and mentioned my ~100rnd per charge weight data sets, and he confirmed the meaning of his statement in the video - from barrel to barrel, we see node shifts and velocity shifts and Satterlee tests, when stacked over and over with multiple barrels, end up a broad swaths of linearity between velocity and charge weight - so when HORNADY LOADS FACTORY AMMO, there’s no reliable nodes and picking any charge weight will not be “in a node” for all barrels, nor result in the same velocity.

When I pile ALL of my different 6 creed barrels together, between different manufacturers, different reamers, and different lengths, I have a broad swath of apparent error around a linear trend, which is what Miles described in the Hornady podcast video.

That’s interesting. I’m not doubting your conversation, but that’s definitely not how I understood his comments in the videos. They called the Satterlee method out by name in their podcasts and were critical of basically all conventional load development methods. Which is why I thought their comments were interesting—as I said, I too have seen flat spots appear to repeat (just not every perceived flat spot in any ladder) and so their comments were contrary to my understanding and received wisdom.

I don’t dismiss your data where you get flat spots to repeat in your rifles. But I also don’t discount the OP’s claim that he hasn’t seen them. I don’t know the full explanation; just what I’ve seen in my own reloading experience.

I think I see repeatable flat spots on certain charges in my 6.5 creedmoor. But try as I might, I can’t find them in my 6.5 Grendel with CFE 223 and any 123 grain bullet. I’m still deciding on my 30-06s—since they are purely hunting rifles I care less, shoot them less, and generally just load to a predetermined velocity so long as accuracy is “good enough.”

While I don’t have your extensive database, I take my chronograph with me about every time I go to the range and record velocity data for every shot in my reloading notebooks when I take it. My load work up usually uses both on target and velocity data for each shot. Maybe my Chargemaster lite or Caldwell chronograph aren’t precise enough to get repeatable results, or maybe I’ve found a way to introduce shooter error into velocity data.

In any event, I stand by the position that a single velocity ladder, considered by itself, is not a reliable way of finding “the best” or least velocity variable load. But if you are shooting multiple of them, or have experience with a barrel and bullet, I don’t discount its usefulness in helping you find where you want to be.
 
What I find somewhat baffling, frustrating and borderline amusing is that 50 years after I was first exposed to reloading for rifles, there has yet to be any standardized consensus on how to go about load development. Most of the reloading manuals will tell you to do load development, but aside from the 5 or 6 basic steps required to put something together that goes BANG, specific how to do effective load development steps dry up in a hurry. So guys are left wandering in the wilderness trying to cut thru all the noise to find something that works.

So if it were up to me to put one together, it would go something like this:

Step 1: Know the limits of your rifle. In general, budget rifles off the rack do not have the same accuracy potential as high end custom jobs, so temper expectations. Use factory ammo as your benchmark. If it won't shoot factory ammo worth beans, hand loads alone won't do much better. Load development should be performed on barrels free of any fouling, copper or carbon, as fouling is a variable that has potential to skew the results.

Step 2: Form follows function. When selecting a bullet to test, match it to the end use. Hunting or target, long range or short. Hunting bullets, in particular, need to be matched up to expected hunting conditions. Hunting bullets have an optimal impact velocity, so pick a bullet to match expected range and velocity where it will be used. At this point you also need to determine your end game goal. Is it optimum accuracy at any velocity that is reasonable? Or to make ammunition that is equal to or better than factory ammo of same use. Same or better accuracy and velocity.

Step 3: During any load development work, best results will come if you eliminate as many variables as you can. Pick the components you want to use and stick with it. Same lot of brass, primers, powder and bullets. The bullet is the driver. Once you pick a bullet, you can chose any components you want to go with it, but once you select the set, you stay with it. You should be able to get to an end conclusion with one bag (50 pieces) of brass, one 100 count sleeve of primers, one 100 count box of bullets and one pound of powder. Be aware that new, unfired brass, even from the same lot, creates a variable that may differ from end result after brass has been fire formed to fit your chamber. Consistency should improve after brass has been shot once or twice.

Step 4: Using fire formed brass as a guide, setup full length sizing dies to bump shoulder from 0.001 to no more than 0.002. Size all brass, fired and new. For any bullet being used, determine COAL that results in bullet jump of 0.020 to lands (assumes rifle barrel is not shot out and has remaining life left in it) and setup seating die to seat bullet of choice to this level (20 thousands off the lands). Seating depth may be tested and altered later, but for now, eliminate the seating depth variable so as to be able to determine optimal powder amount.

Step 5: Experience has shown that best consistency, accuracy and velocity often are found with powder charges near full case capacity, which means reloaders are often walking the edge of what is safe to use. As such, it is critical that reloaders learn and understand signs of pressure. You need to know when you have reached the safe limits of the loads you are working on. Accept they may be more or less than what published data suggests. Stop when you reach obvious signs of pressure, even if there is room in the load data to go higher. Conversely, if you are seeing no signs of pressure, you may be able to go higher than maximum until you do. Depending on load data sources, there are differences in what one source says is maximum vs. others. Your actual results will mean as much as numbers on paper. And remember, this is driven by your end goals. If you have reached good accuracy at factory level velocity or better, and that is your goal, then you do not need to go any higher into the danger zone.

When searching for load data, try to find at least 2 or more independent sources for the load combination you are working on, as these do vary, and sometimes by a lot. From these different sources, conclude what you feel is a reasonable maximum load for your components.

Step 6: Once you have selected a load combination to try, load 3 rounds each at 6 different load levels. Load increment for this 6 step ladder is up to loader, but a good place to start is 1% of max load in grains, rounding down to next whole number.

Example, if max load is 42.5 grains, then 1% is .425 grains, rounded down to .4 grains. If one began at the maximum and worked their way down or backwards, load increments would then be:

6: 3 x 42.5
5: 3 x 42.1
4: 3 x 41.7
3: 3 x 41.3
2: 3 x 40.9
1: 5 x 40.5 (3 shots plus 2 foulers)

End result is one 20 round box of ammo to test, consisting of 2 fouling shots and 6 incremental loads of 3 shots per load.

This differs from the standard recommendation where one starts 10% (i.e., 4.3 grains) below maximum to get to minimum or safe starting place. Instead it is only a 5% reduction from maximum, which in most cases remains safe. It also cuts to the chase much faster. If the two fouling shots show no signs of pressure, one can safely continue on up thru the ladder. Continue until you shoot all 6 loads (reach maximum) or if signs of pressure are found, stop at that point and go no further.

Step 7: Shoot loads at 100 yard targets, shooting for best accuracy and if possible, use a chronograph to measure and record velocity. Shoot the rounds round robin....1 each from loads 1, 2, 3, etc then if you make it all the way to #6, start over. Some rifles (perhaps light weight sporters with pressure point bedding) will demand you allow barrel to cool down between shots so as to not distort the results.

Step 8: Examine completed targets for best accuracy (smallest group) along with correlation of velocity (extreme spreads). Assuming velocity, both amount and consistency) is acceptable, most accurate group wins.

Step 9: If best group is acceptable as is, then one can stop load development there and you are done. If it looks good, but you want to pursue this further, you can then enhance the load development process by going above and below the selected load in 0.2 grain increments (2 loads of 3 shots each, shot round robin). You would then be left with 5 loads shot in 0.2 grain increments, clustered around a level that gave you the best velocity. Pick one you like and you are done.

Step 10: If you want to continue fine tuning the development process, you can do the same thing as before with the single variable now being seating depth. Load 3 rounds each of your selected best load from powder charges, but now increase the seating depth of each by 0.003 (three thousands). That would have you starting out at 20 thousands off the lands (you already have that group), then load and fire 3 rounds each seated at 23, 26, 29, 32, 35, etc. off the lands for as far as you want to go. Expect groups to open and close as you pass thru seating depth nodes. Final seating depth should be at entry point of a node, which gives the most life to a load, as it will continue to work as lands erode away until such time as node is lost.

End result is a load specific to each component combination. Change any one component (and especially bullet or powder) and you start over from the beginning.

Something like that.
 
Last edited:
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.

Here’s the bottom line.
 
Know the limits of your rifle. […] If it won't shoot factory ammo worth beans, hand loads alone won't do much better.

I’ll tell you directly, this is bad advice and false perspective. Many, many rifles shoot relatively terribly with many factory loads, but will pull down MOA or sub-MOA with tailored loads. One of my 300wm’s is an example - I tried a dozen factory loads in it before flying out of state for a hunt, just in case I would end up separated from my ammo in transit, and none of them shot better than 1.5moa, with several 2-3moa. Absolutely abysmal. But it hangs onto ~3/4moa with hand rolls.
 
I'd say for the vast majority of hunting rifles out there, 1.5 to even 2 MOA is not considered terrible but even so, can be improved upon to the level you mentioned. That could almost be considered normal or expected level of improvement for a lot of hunting rifles.

I tried loading for a Mossberg Patriot youth gun with package scope and it wouldn't shoot 5 MOA with anything. Was stunned at how bad it was. No amount of load development is going to bail that gun out. That was the gun I was thinking of. The off the shelf guns with flimsy plastic stocks, rough barrels and entry level package optics. The "crap shoot" guns.
 
Many, many rifles shoot relatively terribly with many factory loads, but will pull down MOA or sub-MOA with tailored loads.

Some of my rifles there is no “factory ammo”, others factory ammunition in said caliber, wouldn’t even fit the chamber reamed into the barrels I have. That said, there are lots of people out there loading and shooting ammunition that, while much less expensive, won’t shoot on par with premium factory ammunition.
 
It’s funny, as a young man shooting rifles, we worked loads up looking for spots where it shot good groups, which could be more than one spot. Since chronos were expensive back then, we didn’t have any idea what the speeds were, just where it shot tight, accuracy nodes.
 
Back
Top