Background: I teach experimentation and statistics as part of my work. I have a tendency to apply the methods I teach to reloading, especially for my precision rifles. My personal favorite to shoot is my .30-06, which also lends itself to playing around with different variables. I am in the process of answering two major questions, which will be worked through here as I finish them, but I'm also interested in what questions we can answer for the hive. Post questions you think can be turned into an experiment, and let's explore!

The two questions I'm working on are these:

1.) Does neck sizing vs. full-length sizing significantly affect group size?

2.) Which variable has the strongest effect on group size of these: jump to the lands, powder charge over a small range, primer type, case length?

I'll be walking through the basic statistical methods used to make these determinations - hopefully it will serve as a guide to help you answer your own questions in the future.

Note: If we want to explore some of this with 223, that's fine too - this isn't just limited to my '06

Is this a DOE?

If so, I'd like to see how neck turning affects accuracy (if at all).

I'll be doing a DOE on the multiple variables. If  you just want to look at neck turning on its own, we use a hypothesis test. An example hypothesis would be like so: does neck turning significantly affect group size at a 5% level of significance? Now, the question is whether that effect is changed based on other factors, i.e.: does the effect of neck turning on group size change with different powder charges, different primers, different cases, etc.

We have to be smart about picking variables that we want to control, and we also have to be clear about what the test actually determined - if we simply choose a load and test turned necks vs. non-turned necks, then we have only made a determination on the effect in that situation. If we introduce some othe variables we think may affect the outcome, then we can do a DOE of some sort to account for those variable changes. Are you interested in this for a specific load, or in general? A specific load is easier to test, but the results are limited in scope.

Tag for an interesting thread.
 









 

What effect does crimp have on accuracy when using bullets without a cannelure?  

If crimping is beneficial, how much crimp is optimal?  

If it is detrimental, how much crimp is acceptable without degrading group size by more than 10%?



There are lots of unspecified questions buried in these questions.  How does one measure the amount of crimp?  How many rounds to be used when shooting and measuring groups?  What crimper type is best (taper, roll, LFC)?

How much does annealing affect brass life?

How does "shoulder push-back when resizing" affect case life?

These two are actually answerable with science - we don't really need the stats. I'll be happy to write up a metallurgical analysis, but that's content for another thread. Essentially, we can pretty easily measure the plastic strain induced by resizing, and relate that to cold work on the metal. A certain amount of cold work will eventually embrittle the brass enough that cracks will rapidly propagate when exposed to the pressure of a shot. The details of how to anneal, however (how hot, how long, how often), are statistically interesting. That will require a lot of work on a microhardness tester, however, and while I have access to one, I don't have the time to do the samples at the moment.

Shoulder pushback shortens case life, as does any plastic deformation. No bones about it.

Rocketman,

I think what you're doing is great, will watch with interest but caution. I would stick to what you proposed in your original post to test. The burden of taking every variable a forum throws at you is a bit much. For one focus is expanded and fun becomes work. For two, folks get upset when results vary from others experiences and/or perceptions. I know your sharper than the average knife, just keep it simple is all I'm saying. Don't let your thread be spun into confusion. I can see it starting already.

Good luck,



dc.

Don't worry about the neck turning variable. I concur with 1911smith as to keep it simple.

I can do a test of neck turning v. no neck turning with the same load.

Although I would like to see the crimp v no crimp.  That discussion gets a lot of play here.

Oh, that's what I was going to do. I already have a DOE worked out for the  multiple variables I want to test in this particular test, and the setup for the other hypothesis test. My idea was to put together tests that anyone could run, and walk the members of the forum through doing it on their own, using my own as an example. In the immortal words I learned in Africa: Small Small Catch Monkey.

For your test on neck turning, here's what I'd recommend: I have several targets that are a grid of bullseyes. I'll add a pdf link to this post later. I have written a program that automatically finds each shot's relationship to its respective bullseye, then concatenates them into a group. This prevents overlapping shots from getting lost. Have one target labeled turned, and another labeled non-turned. Put your turned necks in one labeled bag and your non-turned in another. I will generate a random sequence to fire the shots, to eliminate bias as much as possible.  Do whatever you can to eliminate yourself from the shot, e.g.: shoot from a sled, copious sandbags, whatever. Send me scans or pictures of the targets, and I'll walk through the math (which is easier than it might seem).

I can help out some regarding crimp/no crimp relating to pistol cartridges. When you load and test on range results are instant and repeatable. Some sets of components like a crimp, some don't. There isn't a right or wrong regarding pistol crimps. Every variation is different and finding what works best is a handloaders challenge. As to rifle crimp vs sizing neck tension. I think the issue boils down to what the components prefer in your rifle. Once again, this is the burden a handloader takes for himself.

A couple comment.

The powder charge variable should be eliminated as part of a ladder test.. you should see a range of powder charges that all shoot pretty consistently.

brass annealing will have an effect on neck tension and brass life.

neck vs full length resizing... this effects how much the brass is worked.

The question isn't whether powder will have an effect, but rather, how strong of an effect it has, especially when interacting with other variables.



Indeed, that's well established, and rooted in pretty well-understood science



Again, that's true, but that's not the question being asked. The question is whether one gives more precise results than another, and what that behavior looks like. This is not a place to hem and haw over your experience. This is a place to conduct precise experiments, and let the data do the talking.

To be clear, I do not want anecdotal answers to questions in this thread. The purpose of this is to explore what the data actually says, which very well may fly in the face of common thought. We will test some of these hypotheses using standard experimental procedure, which will allow us to draw some limited, but well-supported conclusions. Hopefully, along the way, we will demonstrate how people can do this for themselves in the future. Whether we allow our thinking to be changed by what we see is up to us individually, as one would expect.