How many shots to establish a receipe?

[mugsie]

When trying to determine the "perfect" bullet / powder combination, I shoot groups. Usually I load 4 shots 1/2 grain apart from the lowest to the highest load, then shoot for groups. This will narrow it down, then I'll split the two best and redo it using finer increments till I have what I think will work.

For instance - .223 Varget 25 min to 27 max. I'll shoot 4 at 25, 4 at 25.5, 26, and finally 4 at 26.5. I'll hold off on the max at times but may use it later.

If I find two groups fairly small, say 25.6 and 26.5, then I'll redo the test using 25.5, 25.9, 26.3, 26.5, so it takes maybe 40 shots or so.

How many does it take you and what are your methods?

 

[winchester1886]

I load one round say 26.0 max,one at 25.8, one at 25.6,and one at 25.4.I Then I shoot one round at a time from low to high.On every round I check for any signs of pressure.If there is no indication of pressure. I will start loading groups of 3 each for a total of say 15 rounds. I always use a powder that will give me the most velocity for a given load.I start from the lowest group to the highest group for the most accurate group.I'm always hopeful that the most accurate load will be the fastest load.In the end I will go for a accurate load over the speed

 

[armoredman]

I usually load tester groups of 10 rounds, fire two groups of five. If I have favorable results, I explore in increments both directions from that spot, until I find the "sweet spot" for that bullet/powder combo. Getting there is 75% of the fun.

 

[redbullitt]

I like to do a ladder test first, provided I have adequate range to work with. A rough one first, say in .4ish or .5ish increments depending on range. Then narrow down to try to nail down the node. THEN I like to shoot some groups and mess with little things like OAL etc.
I like the 4 shot groups too for this. Helps keep the group workable if I mess one up lol.

I make it a point not to chase the velocity dragon. As long as I am getting adequate velocity to work with, I go for accuracy and consistency. Let the rifle and load tell you what to shoot here.

I always try to make sure I try different primers etc. Sometimes one kind will work better than another, noticeably better.

 

[HEAVY METAL 1]

Sound statistical method calls for shooting 7 shots in a group to establish group size.

For centerfire rifles I load in 1/2 grain increments until I find what groups well,make note of it, then keep going up if no high pressure signs are seen as often there is a second sweet spot higher up (not always). I go a little higher until the groups start to open up again then quit. Then starting slightly above the sweet spot, backing down generally in 3/10 gr increments until THE most accurate load is found. At that point I start fiddling with oal. The manuals all give starting loads, but now I start in the mid range as the low end has never been high pressure in any of my guns, but has always been poor accuracy wise. For your own protection follow the book though. If you still can't get acceptable accuracy then you need to change powders & start the process all over. I usually pick the bullet first and that depends on the "mission"-hunting or target.

 

[dacavasi]

Standard statistical methods almost universally call for a sample size of 30 to establish a reliable standard deviation estimate. That said, for my reloading experiments, I generally load in groups of 10 for each variable. That usually seems to do the trick as I home in on a reliable recipe for any specific combination of bullet/powder/OAL/Primer.

 

[kingmt]

I start with 3. If I can't get 3 to group tighter the 1" then there no need to waste the components. There is times I wish I had a few more when I'm working outside of the box but there is always another day at the range. I normally have to pull a few when I get home to salvage the components but that's better the 50.

I work in .4 increments. I'll start at say 25.2, 25.6, 26, 26.4, & so on.

 

[Funshooter45]

I do it similar to the OP mostly. I will pick a given bullet and powder combo and then load up 4 rounds each at 1.5% powder weight variation, basically covering from max recommended down to about 9% below max. The one or two that look best, I go back and load at half the variation, 5 or 7 rounds each, maybe about +/-0.7% weight variation around my suspected sweet spots.

When I first started doing it, I only used 3 rounds per test weight. I didn't like that much. If I had just one shooter induced flier, it kind of made that group useless. It's hard to judge what is a "flier" if you only have 3 holes. It's much easier to do with 4 holes. Obviously, 5 or 7 holes is better when you think you have a sweet spot identified.

 

[amlevin]

I've started to use my Chronograph to do the judging rather than totally rely on my targets to tell the tale.

I recently went looking for a favorable load in .308 using 2495BR and a 175 gr. Nosler Custom Comp bullet.

I loaded 10 rounds of each powder weight, working up to what my internal ballistics program indicated would be maximum.

Each batch was shot over the chronograph. First, group size was evaluated. Second, I looked at the Standard Deviation and also the Mean Average Deviation. It was clear that one batch had far less "deviation" in velocity than the rest. One batch had a Mean Average Deviation of 3 fps with an average speed of 2519.9 fps. In analyzing the group on the target, it wasn't the smallest but the spread was horizontal. There was less than 1/4" of vertical spread. (wind was @90 degrees, 3-6 mph)

One thing about measuring the Standard Deviation or Mean Average Deviation. I's giving a clear measurement of the consistency that is unaffected by wind, hiccups, or impatience.

 

[918v]

It's not the number of shots, but the conditions under which they are fired. More often than not, a "prerfect" load falls apart when the temperature changes +/- 20 degrees and the group size doubes.

To me, a perfect load remains consistent inside a 50 degree temperature swing.

 

[Slamfire]

I regularly shoot ten shot groups with rifles. Maybe I am a data addict. I will shoot over chronograph screens if I can, sometimes it is too cold, too windy, or the range is too busy to set up the chronograph. That chronograph is critical in my opinion as it tells me real world velocities.

If the ten shot string does not meet my expectations, I put my thumbs on the outer most shots and claim success!

For pistol, I load 50 rounds at a time on my Dillion 550B. I don't weigh powder charges for pistols so whatever I see over the screens is a real world average of throws. My data is only limited by the number of shots a Chrony will average. My chrono only averages a maximum of 32 rounds per string, and I have the Beta version. I called up the factory and some Professor set up the program that way.

I believe the more shots you put over the screen the higher the confidence you can have that your velocities, standard deviations, and extreme spreads are real, not just luck of the draw.

 

[Clark]

In the 90's was at a start up company designing ultrasound and every time I walked past a big water tank where some testing was done, I saw people pointing at pages in statistics text books and arguing.

Normally I would jump into any technical argument, but they did not seem to be having any fun.

 

[28lx]

I usually start out with the starting load for a given powder. I load up 5 round batches with the bullet seated at the rifling and 5 rounds each in increments of - .010 until I reach .030 off the rifling. Almost always I find a sweet spot in the COAL and increase the powder charge from there. It works for me. I should add that if I cant load the bullets out to the rifling I start at the maximum length I can get to feed from the magazine.