Load Development — Method, Not Recipes

Load development is the part of handloading with the widest gap between what is confidently taught and what the evidence supports. The safety half of it is settled, well documented and not controversial: start low, increase in increments, watch for specific signs, stop when they appear. The optimisation half — the methods that claim to locate a charge weight at which a particular rifle shoots best — is on much weaker ground than its popularity suggests, and the reason is not that its practitioners are careless. It is that the effects being hunted are smaller than the instruments used to hunt them.

This volume covers the method and the critique. The instruments themselves are Volume 22, and the record that carries results forward into the next batch is Volume 23.

21.1 Start From Published Data, Every Time

The starting point is a current published source: a manufacturer’s loading manual, or the powder maker’s own published data. Those sources exist because the load was assembled and fired in a pressure gun by someone with instrumentation, which is a thing no private bench can do.

Forum data is not a source. Not because the people posting it are dishonest, but because a charge weight is meaningless without the rest of the specification — the brass, the primer, the bullet, the seating depth, the barrel — and forum posts almost never carry it. A number without its context is not data.

The house rule for this dive follows from the same reasoning. Where a published table is reproduced anywhere in this collection, it is attributed to the publisher by name and closed with an instruction to work up from the starting load in the current published source. No charge weight in this dive is ever offered in its own voice as a recommendation.

21.2 Working Up

The procedure is unglamorous and it is the part that actually matters.

Begin at the published starting load, not somewhere between start and maximum. Load a small number at that charge. Increase in small increments, firing and inspecting at each step, and stop at the first sign of excessive pressure or at the published maximum, whichever comes first.

What to watch for, and how much to trust it, is where care is needed — because the popular pressure signs are far less reliable than their reputation.

Case-head expansion is not a usable pressure gauge. Instrumented work by Denton Bramwell found identical case-head expansion from peak pressures differing by a factor of two, using the same lot of brass with the same loading history. Measuring case heads to a ten-thousandth and inferring pressure from the result is measuring precisely the wrong thing.

A flattened primer is among the least reliable indicators available. Primer appearance is affected by the primer’s own cup hardness, by how the pocket has loosened with use, and by headspace.

A cratered primer is confounded by the rifle, not the load. Firing-pin diameter and bolt-face clearance produce cratering on some rifles at every pressure, and the sources genuinely conflict on how much diagnostic weight it carries. That conflict is reported here rather than resolved.

The signs that do carry weight are stiff bolt lift, ejector-mark swipes on the case head, difficult extraction, and a pierced primer.

The correct framing for all of them, and the one sentence worth carrying out of this volume, is this: a pressure sign is a stop signal, never a green light. Its absence proves nothing. Working up is bounded by published maxima, and the signs are there to catch the case where something has gone wrong before the maximum — not to license going past it.

The full diagnostic vocabulary for reading a fired case — the bright ring, the paper-clip test, extraction feel, what a separated case looks like, and the order to read them in — belongs to the Headspace dive, which covers it thoroughly and is not duplicated here.

21.3 Any Component Change Restarts the Work-Up

This is stated everywhere and believed selectively, so it is worth a concrete case rather than an exhortation. The case is the best-documented thing in the research for this dive.

In a study of powder-lot and component variation, a group of handloaders with seventeen years of experience substituted one brand of brass for another in an otherwise unchanged load — Lapua cases in place of Remington-Peters, with roughly 0.6 grain less case capacity. Same powder, same charge, same bullet, same primer. The result was a stuck case and a sheared bolt handle on a Remington 700.

A brass change alone did that. Not a powder change, not a charge increase, not an error in weighing.

The reason is straightforward: a charge weight is safe only relative to the volume it burns in. Less case capacity at the same charge is a higher loading density and a higher pressure, and the relationship between the two is steep near the top of the range.

The practical rule that follows applies to brass brand, powder lot, primer, bullet and seating depth, and it is not a formality. A change to any of them means backing off and working up again.

21.4 Simulation Is Not a Pressure Gun

Pressure-prediction software — QuickLOAD, and the Gordons Reloading Tool — is genuinely useful for understanding relationships: what happens to pressure as case capacity falls, how burn rate interacts with barrel length, why a given powder suits a given case. Used that way it teaches things that are hard to learn otherwise.

It cannot certify that a load is safe, and the same documented case above shows why with unusual clarity. QuickLOAD evaluated the load that sheared the bolt handle as still safe, at 57,585 psi — while predicting its velocity nearly correctly.

That combination is the instructive part. The model got the velocity close, which is the output a user can check, and got the pressure wrong in the dangerous direction, which is the output a user cannot check. A simulation that agrees with the chronograph feels validated, and it has not been validated on the quantity that matters.

Software is for insight into relationships. It is not a substitute for working up from published data.

21.5 The Optimisation Methods, and What They Claim

Two families of method dominate.

The ladder test loads a series of single rounds in ascending charge increments, fires them at a distant target, and looks for a group of consecutive charges whose impacts or velocities cluster — interpreted as a region where the rifle is insensitive to small charge changes.

The optimal charge weight approach loads small groups at several charges, looking for a charge where adjacent charges shoot to the same place, on the reasoning that such a region will tolerate variation in temperature and in metering.

Both rest on the same premise: that there exist velocity or dispersion “flat spots” — nodes — that can be located with a modest number of rounds.

21.6 Why the Premise Is Below the Noise Floor

The criticism that follows is not that these methods are fraudulent or that their practitioners imagine their results. It is narrower and harder to argue with: the effects are smaller than the measurement error of the instruments used to find them.

The instrument argument is the cleanest one. Two good chronographs, measuring the same shots, agree with each other only to about 7 to 10 feet per second. Meanwhile a tenth of a grain of powder is worth single digits of feet per second. So a velocity flat spot narrower than roughly 10 fps is entirely inside the uncertainty of the device used to observe it. A three-charge plateau of a few fps is not evidence of a node; it is evidence that the chronograph cannot tell those charges apart.

The statistical argument is well established and points the same way. A five-shot group has roughly plus or minus fifty percent natural variation in its size — two five-shot groups from the identical load routinely differ by that much. Extreme spread, the statistic most commonly reported, uses only the two most extreme shots, discarding most of the information in the sample. Distinguishing two genuinely different loads to a reasonable confidence takes on the order of fifty rounds or more, not three or five. And Applied Ballistics’ own testing found that ladder testing does not reliably locate a charge window — repeat the ladder and it frequently points somewhere else.

The honest conclusion is not that load development is pointless. It is narrower and more useful: a small-sample test cannot resolve a small difference, so most of what a three-shot or five-shot ladder appears to reveal is sampling noise being read as signal.

21.7 What To Do Instead

Several things survive the critique intact, and they are the ones worth spending components on.

Spend the rounds on fewer questions. A comparison between two candidate loads, fired in enough quantity to mean something, produces a defensible answer. Ten comparisons at five rounds each produce ten unreliable ones.

Use standard deviation rather than extreme spread when reporting velocity, because it uses every shot in the sample rather than two of them. Report the sample size alongside it, since a standard deviation from five shots is itself a noisy estimate.

Test in a way that controls for the things that are not the load. Firing all of load A and then all of load B confounds the comparison with barrel heating, changing light, changing wind and the shooter’s own warm-up. Alternating between them distributes those effects across both.

Accept indistinguishability as an answer. If two loads cannot be separated by the available sample, the correct conclusion is that either will do — and the choice can then be made on grounds that are actually measurable, such as which meters more consistently, which is less temperature-sensitive, or which uses a component that is easier to obtain.

Seating depth is worth testing, and for the same reasons deserves the same discipline. It is measured to the ogive rather than to the tip, for the reason Volume 2 gives, and its effect on dispersion in many rifles is larger and more repeatable than the effect of small charge changes.

21.8 The Record

None of the above is worth anything unretained, which is the subject of Volume 23. The minimum is that every batch fired for development is recorded with its full specification — brass brand and lot, powder and lot, charge, primer, bullet, seating depth measured to the ogive, the date and the conditions — and that the results are recorded against it in a form that can be compared later.

The lot numbers are not clerical detail. Volume 4 makes the storage case for recording them; the stuck case and the sheared bolt handle earlier in this volume are the reason.