I'm tending to believe I'm only locating a "trend" based on 7 shots. Then retesting the most promising combinations with at leased 3 more groups of 7 before I'm willing to say I found the happy load for a given firearm, bullet, powder, primer and case combination.

Any others care to weigh in (so to speak) on this?

------------------
Abe

If everyone thought like me, I'd be a damn fool to think any differently!

On competition rifles I then do ten shot groups.

On 30 cal hunting rifles I do five shot groups.

On my 416 I do 3 shot groups, as it is expensive and I don't want to develop a flinch.

On all the rifles I keep statistics on all groups shot. Do not fool yourself with one good group. When I get 3/4 MOA or better long term average with a hunting rifle I feel like I am in business.

Don

With a sample of 5 (a 5 shot group) the range (distance between the two farthest apart shots) is 90% as effective as standard deviation at representing spread. That is, there is practically no reason to use SD at that sample size.

If your rifle shoots a 1" group with 5 shots, you can be 95% sure that the real long term group size is somewhere between .5" and 1.5".

5 shots in varmint weight barrels.

10 shots out of Service rifle barrels.

Mike

------------------
Victory through superior firepower!

I've gone from 7-shot, to 5-shot, to 3-shot. I will work my way back up depending on the purpose.

If you shoot 5-shot groups (let alone 7) you can quite easily empty your box of bullets before you even come to a load that works. If you do it while working up from starting loads you will also waste a lot of components, not to mention barrel-cooling time and cleaning time.

What I am doing now is firing 3-shot groups over the chrono to get an idea of velocity and pressure. If I get a 3-shot group that achieves a velocity I'm looking for (whether that be high or low) and shows promising accuracy, I'll load some more of those and try again with a larger number of rounds in the group.

This will allow me to try more loads and more quickly winnow out those that serve no purpose for me. It also means I'll spend less time pulling bullets from loads I don't shoot because they are over max based on results with lesser charges.

Three shots tells you all you need to know about a game rifle, because if you make more than three shots in rapid succession at game, then MOA accuracy is unlikely to be a factor with subsequent shots (offhand at a fleeing target).

With target rifles, the number of shots in competition obviously should correlate with the number of shots in a test group.

As for varmint rifles, you really need to know how they're going to perform in extended shooting from a heated and fouled barrel. I wouldn't suggest 40 shot test groups, but 3 five-shot groups fired before the barrels cools back to ambient temperature makes a good indicator of field performance.

For velocity measurements, I find 3 shots very adequate, provided your chronograph is performing properly.

BTW, In varmint calibers I shoot 3 shot groups to get an idea of velocity, then shoot 5 shot groups with a "narrowed down" selection of loads on my next testing session. For deer rifles, I shoot 3 shot groups.

------------------

[This message has been edited by phurley (edited 08-11-2001).]

I always check my zero with a cold barrel for 3 shots prior to a hunt if I can.

------------------
Ray Atkinson

ray@atkinsonhunting.com
atkinsonhunting.com

If you choose a perfectly calm day, so that wind is not a variable, and shoot a single 1" 5 shot group, you do not really know that much about the rifle's performance. Statistically, you can be only 95% sure that the true, long term characteristic is somewhere between .5" and 1.5". To narrow that uncertainty, you need to take many 5 shot groups, and average them.

Many people go to the range, shoot a .5" group, and say that their rifle is performing wonderfully. Others go shoot a 1.5" group, and start tearing the rifle apart to fix whatever is wrong. Neither of these actions is statistically correct. If the rifle is truly a 1" machine, it is entirely normal for the groups to vary unpredictably between those limits.

The NRA's old "Handloading" book had a long article applying statistical formulas this--pretty heavy going. Somewhere around 4 5-shot groups the curve gets a good bit steeper. That's how NRA has always tested rifles.

I have less time to run tests and fewer rifles to rotate between than they do, so I shoot 3-shot groups to test velocity & pressure, then a 5-7 shot group or two to confirm accuracy of loads that look promising.

John

Once I get it narrowed down, I usually shoot four 5 shot groups per load as a fine tune verification.

5 shot groups tend to fit my style best. In my 50 round boxes, I dedicate one row of five to each load while working up a load. It is much easier to keep track that way.

Throw in randomizing factors such as wind, mirage, shifting sandbags, sling swivels getting caught on the bag, scope failures, recoil induced flinch and it is a miracle we ever find a load! Dutch.

Rifles are prone to all sorts of reactions due to various influences. These cover the gamut: wood stock swelling, altitude changes, temperature effects, loose stock mounting screws, loose scope screws or rings, stock wear from recoil, forearm pressure on the barrel, burnt powder fouling, caking and copper buildup in the barrel, barrel effects from machining, dings and disformation in the muzzle crown, headspace problems, bolt problems, loose and worn parts in the bolt face and assembly, chambering abnormalities, cartridge case anomalies, triggers and more. Some guns will "throw" erratic fliers and change point of impact due to being fired cold, hot or oiled and/or dry.

The minimum recommended groups to shoot, for the purpose of a complete and thorough evaluation of a rifles performance ( particularly where there are unusual problems ) is four (4) five (5) shot groups shot on different days under different climatic conditions groups. The procedure entails using 5 targets, a master and 4 others. Targer #1 is overlaid/superimposed precisely on the master and the group is shot. Then later, a different day, target #2 is overlaid on the same master, Then #3 same and #4 same on different days. You will have a master with 20 holes representing each of the 4 slave targets and 4 targets with 5 holes each.

By measuring the group on the master you will then have a true and reliable representation of your rifles performance in the field and it may come as something of a surprise that the so-called 3 shot wonder group of 1/2 inch isn't that at all but rather something quite larger than expected.

When you have a couple of fliers that cause the group size to balloon out to 5" or 6" at 100 yards, then the head scratching begins.

If you notice a different point of center from 5 shot target to 5 shot target by more than 3/4" it is a sure sign you rifle needs a tune up.

If the 20 rounds show excessive vertical stringing 3 times higher than wide it is a sign of upward stock pressure.

If there is excessive horizontal stringing it is a sign of barrel channel pressures.

From there is is necessary to work through all of the other afflictions ranging from ammunition on through to all of the previously mentioned defects.

Some shooters are perplexed as to why one rifle will shoot consistently while another is erratic. This procedure will help make finding the answers faster and easier.

Kendall Dace

The author did slip up by not checking the Bonferroni confidence intervals for his sample size, or is happy with a fairly imprecise answer. With n=5, your 95% limits will have a ratio of about 3:1. In other words, a 1" machine will print 5 shot groups between about .5" and 1.5" 95% of the time. With n=20 (four groups of five) the limits do come in, but not much. You can expect about a 2:1 ratio between the upper and lower limits. If you go to n=40, you'll get about a 1.5:1 ratio between the limits. Even this isn't really super, and the bad news is that as the sample size goes up from there, precision improves very slowly.

With n=20, 1" groups and 2" groups are statistically indistinguishable.

Statistics for means (where the center of a group is) are robust. Statistics for comparing variation (how big the groups are) are notoriously weak.

quote:

---

Originally posted by William E. Tibbe:

The minimum recommended groups to shoot, for the purpose of a complete and thorough evaluation of a rifles performance ( particularly where there are unusual problems ) is four (4) five (5) shot groups shot on different days under different climatic conditions groups.

---

I agree that this is the correct procedure *if* you are looking for a complete and thorough evaluation.

If you are merely working up loads, starting, say 4 grains down from your expected max to check for velocity and pressure, it's just a good way to waste 51 rounds of ammunition and multiple days at the range, shooting ammo that's probably about 200 fps below what you're looking for (in a typical .270 or .30-06).

I will in fact be following this procedure to evaluate a problematic rifle. But, at the same time, I am working up loads for that rifle and will be shooting 3-shot groups to check velocity and pressure, larger groups to double-check accuracy potential of those with promising velocities.

Quite right. I had a rifle that was very old in years but not worn much. At the time it was new it was a super shooter. But over the years its accuracy declined. So I gave it the 20 shot test and found two things.

1. To my surprise the glass bedded forearm barrel channel was actually touching the barrel, whereas years back there was free float. I remedied that with sandpaper in a matter of a few minutes.

2. I still hadn't achieved the desired accuracy. A few eyeball looks down the barrel didn't reveal anything dramatic. But I bought some of the leading copper fouling snake oils and began using them including plugging the chamber and filling the barrel to let is soak. I kept getting blue staining. I took the rifle to the range and shot groups. After EACH shot I swabbed the barrel with copper cleaner while it was hot then ran an oil soaked swab through between each shot. That was to duplicate inner barrel conditions from shot to shot. It removed burnt powder and copper from the previous bullet and "slicked up" the barrel so each new bullet would have the same lubricity.

After several of those swabbing and cleaning sessions the blue tint diminished. Finally the accuracy improved a lot. But it still isn't all the way back to "like new".

Over time the gun had changed craracteristic all by itself. ( stock/forearm swelling or warping ). I wasn't, obviously, doing a thorough cleaning job over the years and had a buildup of fouling that took a great amount of elbow grease and time to dislodge.

Now, after I run a string of 20 shot diagnostics and know how the rifles shoot, I too will only shoot a short series of shots to quickly check it out. If it shoots where it is supposed to without throwing fliers or tracking, I call it good and go hunting.

Kendall Dace.

If you are going to target shoot or even varmint shoot then you must go a different route, I guess....I still think we get to wound up in technical and theoretical ga ga, but if that is ones pleasure then have at it.
Most of my guns shoot a 1/2" to 1" very consistently and that's good enough for me. I'm primarily a hunter..

------------------
Ray Atkinson

ray@atkinsonhunting.com
atkinsonhunting.com