If it is true that mighty oaks from little acorns grow, then it is equally true that monumental research projects spring from tidbits of information. Sometimes they take on a life of their own. This one began when I learned that one of the factories used a powder that was much slower than the Bullseye or W-231 we usually associate with loading for the .45 ACP.

Most of my working life was spent not at a loading bench, but at a laboratory bench. I earned my living doing chemical analyses. So what, you ask? Well, before I was a chemist, I was a gunsmith and bullseye pistol shooter. The two have much in common. Both chemistry and shooting demand that you do the same thing in exactly the same way over and over, and both will punish you smartly if you don’t. Science requires us to challenge and see if we can get the same answer another lab does. The boss sends us a bottle of stuff, and we have to find out about it. Of course, the boss already knows, and things get pretty grim if we don’t get the same answer he does. It all boils down to testing, which, not surprisingly, is exactly what I do today, but this work is a lot more fun. My gunsmith training involved Mr. Browning’s wonderful 1911 and so did two-thirds of my bullseye shooting.

A gun writer’s life is ruled by things that may not be quite as easy as bench analysis. They are mechanics, power and accuracy. Most of the time you could distill it down to power or accuracy. Power is easy. Buy a chronograph and a simple calculator and you too can be an expert. I guess some could argue that buying a Ransom Rest is about the same. I don’t think so. I can’t begin to count the number of times I’ve read that “this” load or “that” gun is simply wonderful based on a 3-, 5- or 10-shot group. Science taught me that anyone who makes statements like that is - how can I put this politely - totally full of . . .

Shooting, just like analytical chemistry, is governed by something called the bell curve - random distribution if you like. Basically this states that any set of experimental data is going to have a few results at each end of the spectrum, but the majority will fall somewhere in the middle. Specifically it says that 68 percent of results will fall within plus or minus one standard deviation of the mean, 95 percent will be within two SDs and 99 percent within three. Perhaps this is the most compelling reason to believe that standard deviation is not a predictor of accuracy.

Accuracy testing is one of the best ways I know to prove the bell curve is real. If you shoot five groups, most of the time you’ll find one that is much larger and one smaller while the other three come somewhere in the middle - often pretty close to what we call “average.” You can even use that analogy when looking at the shots within a single group because very often one bullet hole will double the group size. More than a few shooters ignore those “flyers” because they don’t want their data to look bad.

Since the bell curve rules, conclusions need to have some degree of statistical reliability. This really doesn’t have anything to do with the phases of the moon, your sign or the presence or absence of gremlins (although sometimes I wonder about those). It means that we must have a meaningful number of data points before we can say this or that is true. The more data you have the happier you can be about your conclusions. How do you decide what’s enough? How many shots in a group and how many groups? Sometimes it helps to be arbitrary.

Since most of my testing began with pistols used in bullseye competition, the first one is easy. Everything in a bullseye match is based on 10-shot groups. I know five-shot groups are popular, but   if you’ve watched the shots stack up, it’s very common for the second half of the string to have a marked effect on group size, and if we made conclusions based on the smaller number, we could well be led astray.

The number of groups should - according to statisticians - be an odd number. Almost everyone uses either three or five groups, and since ammunition comes in 50-round boxes, it seems reasonable to shoot them all. What good are 20 loose rounds? So, with ground rules established, let me tell you how I got in this mess.

Over the years I have maintained that headspace and brass are not significant issues in the accuracy of the 1911 pistol. I’ve said that headspace is determined by the extractor and doesn’t matter. Well, it doesn’t, but the cartridge really does headspace on the case mouth unless the case is so short that it can’t get there. I also believe that sorting and trimming .45 brass is an exercise in time-wasting. Discussions with Editor Scovill prompted all the following and the humble acknowledgement that it does headspace on the case mouth.

Materials and Methods

Any proof must begin with some givens. Lots of shooting taught me that Alliant Power Pistol and Remington’s 185-grain .45 wadcutter (WC) bullet gave wonderful accuracy in target loads. So it was logical to wonder if the same might be true with full-charge loads using Remington’s 230-grain full metal jacket (FMJ).

All ammunition was loaded on an RCBS Pro 2000 with standard RCBS carbide .45 ACP dies. The extra station of the press let me use a Redding taper crimp die that was adjusted for a firm crimp. Seating depth was set to duplicate most factory ball ammunition, which measures about 1.260 inches. Care was taken to be sure all components were from the same lot and, with the exception of brass, that was done. I have a fairly good supply of once-fired brass saved from other tests but no track of the lots. The new Federal brass was some I’ve had for years left over from a project long forgotten. New Winchester and Remington brass was purchased from MidwayUSA as were the bulk bullets.

All test shooting was done from a Ransom Rest at a range of 25 yards divided into five, 10-shot groups. Each round was chronographed so the velocity figures represent the average of all shots. Everything was shot in my Rock River Arms basic pistol, which has proven itself time and again to be among the most accurate pistols I own. You can’t believe how it bugs me when somebody says “accuracy was poor” but doesn’t furnish any information about the gun’s qualifications. “Poor” in my book would be the 5- to 6-inch groups we used to get from GI guns. “Good” isn’t any better without something to use for reference. There is no magical good = “this” or poor = “that.” Accuracy just is.

We simply must have a good gun to be able to even see small differences.

Some preliminary work suggested that the combination would work, so I set up a truly worst-case scenario for the test. I happened to find a small bag of primed .45 ACP brass at a gun show. It was clean and included a total of eight different headstamps: four each of commercial and GI origin. I’ve no idea at all what primers were there. You can hardly get worse than that. For comparison I prepared 50 brand-new Federal cases.

The load was 6.8 grains of Alliant Power Pistol with CCI primers and 230-grain Remington FMJ bulk packed bullets. In the Ransom Rest it was obvious my experiment wasn’t going to turn out the way I wanted it to. The load using new brass averaged 1.08 inches for five, 10-shot groups. The mixed brass averaged 1.38 inches with a standard deviation almost twice as large. Interestingly there was only 2 fps difference in velocity. The    results are listed in Table I.

In practical terms either of those accuracy measurements would have met the needs of all but the pickiest shooter and would have even been the envy of more than a couple factories, but the fact is the load using new brass is significantly more accurate.

Given the rather surprising outcome of the first effort, I fired a couple other loads. The first load was a bit hot compared to factory ball ammunition, so the charge was reduced to 6.6 grains. Then 50 rounds were loaded in once-fired Federal brass and another 50 in mixed once-fired cases from Remington, Winchester and Federal divided as equally as possible. This time some factory ammunition was included for reference. And, as an afterthought, 50 rounds were loaded in once-fired Federal brass using CCI, Remington, Winchester and Federal primers. Once more the mixed brass groups were a little larger at 1.81 inches compared to 1.54 inches for the group with Federal brass.

The next phase used once-fired Federal cases with the primer as the only variable. There was really no difference in velocity and only a little in accuracy with 6.6 grains of Power Pistol and a 230-grain Remington FMJ.

Statistically the load with Remington primers was not as accurate, but there wasn’t 2 cents worth of difference in the rest.

The real shocker though was the results with the factory loads, which are summarized in the table below. Again, all are 230-grain FMJ factory loads, and the results made me retest both the Remington and Federal loads.

There are some conclusions here, and all of them run counter to what I would have expected: (1) There did appear to be a difference between segregated and mixed brass, (2) primers made no difference in velocity and only a little in accuracy, and (3) all the handloads shot better than or equal to the factory ammunition. Had I been a gambling man, I’d have lost my shirt.

The tests so far had been done over two days, and it was clear there was more to learn. I wanted to know if the apparent differences between new and once-fired brass would repeat. So the plan for phase three became to load 50 rounds in new and once-fired cases from Federal, Remington and Winchester. While CCI primers had been most common in the earlier tests, an additional 50 would be loaded with either Winchester or Remington primers. All the results are summarized in Table II.

Most of these differences are not significant but some are. It’s a good thing science doesn’t always have to explain “why” as long as the data is properly gathered and reported.

Guns

Let’s talk a minute about Government Model pistols. I wish there were some objective way to rate them. Wine experts use a scale - often from 70 to 100 points. Anything below 70 is not worth mention, or maybe if it is made from grapes and has a little alcohol it gets 70 points. When we buy wine we also look for the name of the winery or chateau, Silver Oak or Chateau Lafitte. Not unlike wine, just because you pay a bunch for a gun doesn’t mean it will shoot better than anything on the planet. I’ve tested accurized pistols with $4,000 price tags and found some that cost 75 percent less that shot a lot better.

My choice of the Rock River pistol for this project wasn’t based on cost. In fact you could pay as much or more for a semi-custom pistol from one of the factories. My choice was based on knowledge of this particular gun and a lengthy record of outstanding accuracy.

Interpretation of Results

Very often the conclusions we reach are swayed by how the data is presented. That happened to me. I shared my data with a good friend who has a lot more science education than I, and when he sent it back, he had separated the data into two groups that included only the handload with 6.6 grains of Power Pistol and then sorted it in ascending order. That way you can see there really aren’t any definable trends related to brass make or degree of use or primer make. You can also see the “cussed randomness” of accuracy testing where it isn’t all that unusual to have a group in a series that is twice as large as the one before.

I know this data will have a different value for different people. For some it will be of only academic interest, but for others it might be enough a make them change their loading practices. But we have to be careful to not give numbers more than their just due. A difference in the second or third decimal place is irrelevant - even .1 or .2 may not matter much.

Statisticians can do wonders with numbers, but I don’t have those skills. Instead I’ve got a simple little computer program called Baltec1 (Bill Davis, Tioga Engineering, Box 913, Wellsboro PA 16901; e-mail: tiogaeng@epix.net) that does lots of handy things. One of those is to calculate the probability that A is more accurate than B. One of the things I set out to find was whether or not new brass made a difference. Well, it turns out that it does - sometimes. With Federal and Remington the probabilities that loads in new brass are more accurate are 99 percent and 95 percent, respectively. But with the Winchester cases there is a 91 percent probability that the once-fired cases are more accurate.

So here we are after shooting 30 different loads - five, 10-shot groups of each for a total of 1,500 rounds - and the earth still wobbles around on its axis, and Merlin’s pointy little black hat is only slightly askew. And here am I, eating some crow for having said that the days when a .45 ACP handload could outshoot factory ammunition were past. That was not the case here. Perhaps the most interesting information involves Remington. Big Green makes good primers, brass and bullets, but when they put them together, the whole doesn’t seem to do justice to the parts. I have no idea what powder Remington uses in its factory loads, but I sure could make a suggestion.

It is my normal practice to share data like this with a few friends and colleagues to help look for glaring errors or things I missed. Not all are helpful. I usually get a few “Why don’t you?” questions, but every now and then somebody will tell me all the things I “shoulda” done. Patient I am, but everyone has their limits, and there are really times when I’d like to say, “I wanted to leave something for you to do for your article.”