Cookie settings

Choose whether to allow website analytics. You can change your mind here at any time.

Essential features Always on

Contact-form security, your privacy choice, age verification and navigation. These are separate from analytics.

Read the cookie and privacy policy

Regular Dude Defense

Privacy policy

Effective October 22, 2026

Loading privacy policy…

Skip to content

Measuring accuracy honestly: groups, averages and sample size

Understand group measurements, mean radius, sample size and selection bias so rifle accuracy records support conclusions instead of misleading claims.

By Dr. Dylan J. Johnson, D.Sc.

8 min read

A small group is a result. It is not, by itself, a description of what a rifle will do repeatedly. The difference matters when somebody is comparing ammunition, discussing a service concern or deciding whether a claim about a firearm is believable.

Keep the target you are proud of. Keep the ordinary targets too. A useful account includes the number of shots, the conditions, the measurements and the results that did not make a good photograph. Without that context, “sub-MOA” can be a precise-looking label attached to very little information.

State what the number describes

In ordinary range conversation, accuracy often means group size. More carefully, the size of a group describes the dispersion of the observed shots, while the group's location relative to the intended aim point is a separate part of the result. A tight cluster can be displaced from the intended point.

That distinction prevents a misleading comparison. If two targets have similar spread but different centers, a single group-size number has hidden the location difference. If their centers are similar but the spreads differ, a statement about where the rifle is sighted has hidden the dispersion difference.

Use the terms clearly enough that another person knows what you measured. You do not need to turn a range notebook into a laboratory report. You do need to avoid using one measurement to claim that it answered every question about the firearm, ammunition and shooter.

The NIST discussion of measurement terminology is useful background on repeatability, reproducibility and the need to state conditions. It is general measurement guidance, not a firearm-testing standard.

Extreme spread records the farthest pair

Extreme spread is the distance between the centers of the two holes farthest apart. It is familiar because it can be measured directly and reported as one number. It also has a limitation: those two observations determine the result regardless of where the other shots lie.

Two groups can have the same extreme spread and different arrangements of the remaining shots. One might contain a close central cluster and one distant shot. Another might have the shots distributed across the full group. The shared spread number does not make those patterns identical.

A distant shot can change extreme spread substantially. That is a reason to retain the target and describe the result, not a reason to erase the shot automatically. The observation may be meaningful even when it is inconvenient.

When comparing recorded spread, keep the number of shots clear. A three-shot result and a larger sample have different opportunities to include an observation farther from the center. Presenting both as if the group-size figure were independent of sample size makes the comparison harder to interpret.

Mean radius uses every recorded shot

Mean radius is the average distance of the recorded shots from their group center. It includes each shot rather than being determined only by the farthest pair. That gives it a different view of the same target.

It is not immune to a distant shot. Nor does a small mean radius prove that the group was centered on the intended point. The reference is the group center, so the measure addresses spread around that center rather than answering every question about placement.

Software can help calculate it from a target image, but the input still matters. The scale, the marked hole centers and the identification of shots all need to be correct. A number with several decimal places does not repair a poorly measured target or an uncertain assignment of overlapping holes.

Record which measurement you used. Do not compare one person's extreme spread with somebody else's mean radius as though the smaller number settled which result was better. They describe different properties and should be labeled accordingly.

A single group leaves a wide gap in the evidence

A target contains a sample of a process with variation. Ammunition, conditions, equipment and the shooter all contribute to what is observed. One group cannot establish how much variation will appear across other groups or on another day.

The NIST statistics handbook's sample-size discussion explains the general relationship between variability, sample size and the precision of an estimate. It also makes the purpose of the estimate part of choosing the sample. There is no universal count that settles every measurement problem.

For the range notebook, the important habit is to state the sample rather than hide it. “Best of five groups” conveys something different from “the only group measured.” An average across several groups conveys something different again. None should be abbreviated to the same unexplained claim.

More observations do not compensate for a badly defined comparison. If the conditions or measurement method change without being recorded, a larger collection of numbers can still leave the original question unanswered. Quantity and a clear method belong together.

Selection can make a photograph tell the wrong story

Suppose a shooter records several targets and shares only the smallest group. The photograph may be authentic, and the measurement may be accurate. What has changed is the question it answers: it shows a selected result rather than the usual result in that session.

Selection can happen unintentionally. The owner remembers the good target, discards the damaged paper and leaves the ordinary groups in a range bag. Later, the saved target becomes the basis for a claim about the entire setup.

Decide what you intend to report before reviewing the result. If the purpose is comparison, keep the comparison complete. If the purpose is simply celebrating a personal best, label it that way. A personal best does not need to be disguised as an average to be worth keeping.

The same applies to when a test ends. Stopping after a favorable observation and describing it as representative builds the preference into the result. A record of the planned sample and what was actually recorded makes that decision visible.

A “flyer” needs an explanation

Calling a shot a flyer describes where it landed; it does not explain why it landed there. Deleting it without a reason replaces an observation with an assumption. Keep the original measurement and state any known circumstance separately.

If an identifiable interruption occurred, record it when it happened. A note made at the time is better evidence than a story constructed after the target looks disappointing. The distinction is between a documented event and a guess about what must have caused a distant hole.

There are cases where an observation should be excluded from a particular analysis. When that decision is made, identify the rule and show what was excluded. Retaining the complete target allows somebody else to understand whether the reported subset fits the stated purpose.

Do not let “the shooter pulled it” become the answer to every inconvenient result. It may be correct in a specific case, but a mechanical concern and a measurement concern should not be dismissed solely because the owner would prefer a smaller group.

Compare like records before comparing results

Keep the setup and conditions consistent enough to make the comparison understandable. Record distance, ammunition identity, relevant conditions and any known change in the equipment or barrel condition. Those notes explain what the measurements can reasonably be compared with.

Do not combine results merely because they were produced by the same rifle. A change in ammunition, conditions or measurement method may mean the records answer different questions. Label the groups before reducing them to an average.

Around Pahrump, the difference between one session and another can be substantial. A note saying “hot and windy” is limited, but it is better than pretending the conditions were unchanged. If exact conditions were not measured, do not insert precise numbers later from an unrelated weather report and present them as observations at the target.

Measurements also need consistent units and definitions. Keep center-to-center group measurements distinguishable from a casual outside-edge estimate. If a target image is used, preserve the scale reference and the unedited original so the figure can be checked.

Composite results need their method stated

A composite can combine observations from several groups, but there is more than one way to align them. A result aligned on the original aim points describes something different from one in which each group is shifted to its own center before combination.

Centering individual groups can remove between-group shifts from the display. That may be appropriate for a stated question about dispersion within groups, but it should not quietly stand in for a result that includes placement changes between groups.

Explain the alignment in the record. Keep the original targets and the individual measurements beside the combined result. A composite should make the evidence easier to understand, not conceal the work that produced the final image.

If software performs the combination, learn what it is doing before adopting the output as a conclusion. The program's display is an analysis choice. It does not automatically establish that the choice matches the question the owner intended to answer.

Small differences deserve modest conclusions

If two reported averages are close, the data may not support a strong preference. The practical decision could still depend on cost, availability or consistency of supply. A tiny numerical difference is not automatically a reason to replace a large quantity of ammunition or authorize work on a firearm.

State the result at the precision supported by the measurement. Reporting hundredths of an inch from an uncertain photograph makes the output look more certain than the input. Extra decimal places are a formatting choice, not additional evidence.

Keep a separate place for the conclusion and the observation. “This sample had a smaller average spread” is an observation about the recorded data. “This load is always better” is a much broader claim and needs more support.

A restrained conclusion is useful. It tells the next person what remains uncertain and prevents ordinary variation from becoming an unnecessary repair decision. It also makes a later comparison easier because the earlier record has not promised more than it established.

Bring evidence to a service discussion

If results change unexpectedly, retain the targets and the information needed to identify the firearm and ammunition. Describe when the change appeared and what else changed around that time. The original records are more useful than a recollection that the rifle “used to be a tack driver.”

A target alone cannot diagnose a mechanical problem. It can document why the owner wants the firearm assessed and help the gunsmith understand the concern. Have suspected safety or mechanical faults evaluated rather than treating additional range results as permission to keep using a firearm that may need attention.

For a gunsmithing appointment in Pahrump, contact Regular Dude Defense at (775) 471-9010. Bring a clear account of the concern and the records already available. Honest measurement is valuable because it improves that conversation, including the occasions when the correct conclusion is that the current evidence is not sufficient to identify a cause.

Last updated (first published )

All posts

Discuss firearms instruction in Pahrump

Explain your experience, equipment and learning needs so we can discuss a suitable training appointment. Visits are by appointment, 8:00 AM – 4:30 PM, seven days a week.

7 min read

Low-light shooting fundamentals

How poor visibility changes safe identification, light selection, battery care and preparation for supervised low-light instruction.

  • Training
  • Carry & Home Defense

8 min read

Natural point of aim: letting the rifle settle

What natural point of aim is, how to find and adjust it with your body instead of muscling the rifle, and why it makes your shooting more accurate.

  • Rifles
  • Training

8 min read

Weapon lights for pistols

How to think about pistol weapon lights: lumens vs. candela, holster compatibility, activation switches, and practicing safely so the light helps you.

  • Carry & Home Defense
  • Optics