RegionRat
Gold $$ Contributor
On several posts lately, we see a question being set up by someone starting out and asking for help with load development.
They might say something like "I have my charge and my seating depth worked out pretty well, should I test neck tension and primers? Which do I test first?"
My answer would be "it depends". If you have not met your goals and suspect the rig can still do better, then why not keep looking? That is one of the main points to reloading or competition, to wring out the best performance from ourselves and the gun.
However, too many times the beginner will be given advice to test a primer swap or try a different neck tension, but no other warnings. They may even be told to change one thing at a time. Nothing wrong with that... or is there...
As often as not, they come back frustrated.
It is true that a different neck tension or primer may do better. It is true that testing carefully might mean changing only one thing at a time. However... as often as not folks get frustrated with their results. Often it is because we gave them half the answer but didn't warn them about the rest of the pitfalls of setting out to change just one thing.
What we often fail to tell them is in a complex system that one change, by itself, may give a false negative result, but instead of giving up you may need to change some other parameters to find the optimum. Sometimes, that negative result is the final answer and spending any more on it is a waste. There is risk in the question of how to know? Then the swirl over the method to find the truth begins. We start to argue over math and statistics, or the costs of a chronograph, and before long there is more swirl about the next steps.
Nothing I will add here today will prevent all of that from happening again. Nothing I am adding here will turn everybody into a role model of efficiency for load development. My purpose is just to shed light into DOE for the folks who never got a chance to ever even hear about it, and to point out some possible pitfalls to what we call OFAT testing, (One Factor At A Time).
Can an answer be found without burning up the barrel? What is the test method? How many samples? Etc.
I won't get into statistics today because that gets enough play on other threads. But I rarely see a discussion on the design of the load development test or warnings about blocking factors and other pitfalls. So while I am stuck waiting here today, I thought I would kick it open without cluttering up someone else's thread.
There is an entire study called design of experiments (DOE) that can become a specialty all by itself. Some of you do this well instinctively, others will never consider it. To each their own.
DOE deals with how to set up tests for everything from simple problems to complex ones. Optimizing a gun and recipe combination is not a simple context. Many of you don't give yourselves credit for years of experience that resulted in a good instinct for what to test and what not to test. Others here waste many shots by not realizing this has all been done before and there are possible pitfalls. Nobody has all the answers. That is okay too as long as you enjoy the shooting.
The main point of brining up DOE study in this discussion is the efficiency of tests when there are complex interactions between the parameters like seating depth, neck tension, primers, etc., that might force changes in each other to find optimum performance.
Too many times, you will see folks who think they simply swap a primer and get a result, when in some cases they should have known better. Sometimes, when you change "one thing" and test, you still don't know anything till you change another to find the optimum.
Many (okay, most) shooters are not trained scientists. That clearly doesn't matter cause many of the best have no formal credentials and set world records. But, the point of this forum is to share and teach each other that sometimes you have to keep an open mind to testing methods and statistics if you are going to become a good load developer. You don't have to become a statistician, but you sure better use them in your favor if you are short of materials or money. I will waste a little time today shedding light into complex testing and DOE.
Rather than just telling someone new to test primers or neck tension, it is also fair to warn them at the same time that you can't count on the answers from a single parameter change like leaving everything else the same and just making one change. If you hit that lottery, great! But, if not, sorry it just doesn't always work out that way.
It may be true that for one test you make one change, but the main point here is that you may be forced to change something else in addition, if your real goal is to improve performance.
There is a statistical chance you get a positive answer from a primer swap or neck tension change, but there is also a chance that when you get a negative answer and that the answer could have been different and positive if combined with another change.
Here is a little paper on the concept while using something other than shooting to illustrate the whole DOE concept. Just ignore the statistical math, names, and terminology and push through it to get an idea. Pretend tomato yield in the example is group size or score and fertilizer is some new change and you will get the idea.
In the link to the web page, you will see an example of One Factor At A Time or OFAT testing and how sometimes you don't find an optimum when you stick to it.
https://www.jmp.com/en_au/statistics-knowledge-portal/what-is-design-of-experiments.html
The whole point to this isn't to give the no-maths a headache, but in the hope that the incomplete advice and swirl gets minimized. The other hope is to get at least some folks interested in the stats and DOE methods that help keep more folks shooting when we are in a drought of supplies and primers are precious.
Keep or pitch. YMMV
They might say something like "I have my charge and my seating depth worked out pretty well, should I test neck tension and primers? Which do I test first?"
My answer would be "it depends". If you have not met your goals and suspect the rig can still do better, then why not keep looking? That is one of the main points to reloading or competition, to wring out the best performance from ourselves and the gun.
However, too many times the beginner will be given advice to test a primer swap or try a different neck tension, but no other warnings. They may even be told to change one thing at a time. Nothing wrong with that... or is there...
As often as not, they come back frustrated.
It is true that a different neck tension or primer may do better. It is true that testing carefully might mean changing only one thing at a time. However... as often as not folks get frustrated with their results. Often it is because we gave them half the answer but didn't warn them about the rest of the pitfalls of setting out to change just one thing.
What we often fail to tell them is in a complex system that one change, by itself, may give a false negative result, but instead of giving up you may need to change some other parameters to find the optimum. Sometimes, that negative result is the final answer and spending any more on it is a waste. There is risk in the question of how to know? Then the swirl over the method to find the truth begins. We start to argue over math and statistics, or the costs of a chronograph, and before long there is more swirl about the next steps.
Nothing I will add here today will prevent all of that from happening again. Nothing I am adding here will turn everybody into a role model of efficiency for load development. My purpose is just to shed light into DOE for the folks who never got a chance to ever even hear about it, and to point out some possible pitfalls to what we call OFAT testing, (One Factor At A Time).
Can an answer be found without burning up the barrel? What is the test method? How many samples? Etc.
I won't get into statistics today because that gets enough play on other threads. But I rarely see a discussion on the design of the load development test or warnings about blocking factors and other pitfalls. So while I am stuck waiting here today, I thought I would kick it open without cluttering up someone else's thread.
There is an entire study called design of experiments (DOE) that can become a specialty all by itself. Some of you do this well instinctively, others will never consider it. To each their own.
DOE deals with how to set up tests for everything from simple problems to complex ones. Optimizing a gun and recipe combination is not a simple context. Many of you don't give yourselves credit for years of experience that resulted in a good instinct for what to test and what not to test. Others here waste many shots by not realizing this has all been done before and there are possible pitfalls. Nobody has all the answers. That is okay too as long as you enjoy the shooting.
The main point of brining up DOE study in this discussion is the efficiency of tests when there are complex interactions between the parameters like seating depth, neck tension, primers, etc., that might force changes in each other to find optimum performance.
Too many times, you will see folks who think they simply swap a primer and get a result, when in some cases they should have known better. Sometimes, when you change "one thing" and test, you still don't know anything till you change another to find the optimum.
Many (okay, most) shooters are not trained scientists. That clearly doesn't matter cause many of the best have no formal credentials and set world records. But, the point of this forum is to share and teach each other that sometimes you have to keep an open mind to testing methods and statistics if you are going to become a good load developer. You don't have to become a statistician, but you sure better use them in your favor if you are short of materials or money. I will waste a little time today shedding light into complex testing and DOE.
Rather than just telling someone new to test primers or neck tension, it is also fair to warn them at the same time that you can't count on the answers from a single parameter change like leaving everything else the same and just making one change. If you hit that lottery, great! But, if not, sorry it just doesn't always work out that way.
It may be true that for one test you make one change, but the main point here is that you may be forced to change something else in addition, if your real goal is to improve performance.
There is a statistical chance you get a positive answer from a primer swap or neck tension change, but there is also a chance that when you get a negative answer and that the answer could have been different and positive if combined with another change.
Here is a little paper on the concept while using something other than shooting to illustrate the whole DOE concept. Just ignore the statistical math, names, and terminology and push through it to get an idea. Pretend tomato yield in the example is group size or score and fertilizer is some new change and you will get the idea.
In the link to the web page, you will see an example of One Factor At A Time or OFAT testing and how sometimes you don't find an optimum when you stick to it.
https://www.jmp.com/en_au/statistics-knowledge-portal/what-is-design-of-experiments.html
The whole point to this isn't to give the no-maths a headache, but in the hope that the incomplete advice and swirl gets minimized. The other hope is to get at least some folks interested in the stats and DOE methods that help keep more folks shooting when we are in a drought of supplies and primers are precious.
Keep or pitch. YMMV
