Group tuning works like this: take the bow, take the arrows, go to the range. Shoot a group. Adjust something — rest position, nocking point, cam lean. Shoot another group. If it’s tighter, keep the adjustment. Repeat until the cluster closes. Whatever setting produced the best group is the correct setup.
This sounds reasonable. For a well-trained archer with excellent form consistency shooting in calm conditions, it can actually produce a useful result. For most archers, most of the time, it produces the appearance of a result — which is a different thing, and worth understanding clearly.
Why it looks like evidence
Group tuning borrows the structure of a controlled experiment. Variable changed, outcome measured, conclusion drawn. The feedback is immediate and visible: a rest moved an eighth of a turn, the group tightened. That looks like causation. It feels like progress.
The problem is that the experiment is not controlled. The archer is not a constant between groups. The wind is not a constant. Fatigue accumulates through a session in ways that don’t register as fatigue — the hold subtly degrades, the timing drifts, and nothing about how the shots feel indicates this is happening. The grip contacts the riser with slight variation on each shot that is entirely below conscious detection. All of these produce group scatter, and that scatter is indistinguishable from scatter caused by bow configuration.
When the archer moves the rest a sixteenth of an inch and the third group is tighter, the rest gets the credit. The wind does not file a rebuttal. The momentary improvement in hold consistency doesn’t raise its hand.
The signal-to-noise problem
Before examining the noise, examine the constraint. A compound bow in a shooting machine — no archer involved — will hold groups the human archer cannot approach. The mechanical precision of modern compound equipment outstrips the archer’s execution consistency by a significant margin. The ratio is roughly ten to one in terms of group area under controlled conditions: the equipment can out-shoot the archer shooting it.
This matters because it reframes the question group tuning is asking. Equipment fine-tuning assumes the equipment is the limiting variable. For most archers, most of the time, it is not. When the archer is the constraint — which is the default condition — adjusting the bow is adjusting the wrong variable entirely. Group tuning is a fine-tuning technique being applied before the coarse variable has been addressed.
A 1/32-inch rest adjustment shifts the arrow’s departure point by 1/32 inch at the rest. Projected over 20 yards, the expected group centroid change is a quarter inch or less. The shot-to-shot standard deviation of a recreational or intermediate archer’s groups — driven by hold variation, timing inconsistency, and grip contact changes — is typically several times that magnitude. The configuration signal is buried under the archer noise. The two are not separable by eye.
This is not a failure of effort or attention. It is a property of signal detection in a noisy system. When the signal-to-noise ratio is low, the human pattern-recognition system finds patterns that are not there — not because the archer is careless, but because the system is specifically optimized to find patterns, and it does not turn that off when the patterns are spurious.
The session plays out predictably: two mediocre groups, a third one tighter. Something was changed just before group three. The conclusion is immediate. In reality: a brief wind lull. A slightly calmer hold. Three shots in a row where the timing was consistent for reasons that have nothing to do with rest position. These produce a tighter group with no meaningful configuration change. The adjustment is invisible. The conditions are not.
The hunt continues because the evidence is never definitive. There is always another session where a different adjustment looks like it worked. Archers spend entire seasons moving the rest back and forth in small increments, cataloging which configurations produced better sessions, and arriving nowhere stable. This is not tuning. It is pattern-matching noise with a hex key.
At elite execution levels — where form variation is small enough relative to the configuration signal — group analysis can detect real mechanical changes. That threshold is high. Below it, the thing being hunted is not there to be found.
Line tuning — a better measurement, same wrong direction
Line tuning is group tuning done one dimension at a time. Instead of reading a full group, it shoots at a horizontal or vertical line and only reads deviation in one axis — a vertical pass, then a separate horizontal pass. The measurement is cleaner. The signal is easier to read.
The direction is unchanged. The adjustment is still: read the result, move the rest, repeat. The bow is being adjusted to accommodate the arrows. A cleaner measurement of the wrong thing is still the wrong thing.
There is also a subtler problem with how the bare shaft fits in. In line tuning, the bare shaft is shot at the end — wherever it lands becomes the reference for that tune. The reference is discovered. The Axial approach sets it deliberately: intentional bias is established from mechanical reasoning before any arrows are shot. The bare shaft lands where it was designed to land. Intentional beats discovered.
The direction problem
Both group tuning and line tuning share the same underlying premise: when arrows don’t group consistently, the bow needs to change. The bow is adjusted toward whatever configuration makes the arrows more agreeable.
This is the direction Axial is against — not because adjusting the bow never produces a tighter group in the short term, but because the tighter group is a compromise, not a solution.
Where arrow inconsistency stems from internal variation — insert concentricity, spine orientation, nock seating — rest position does not resolve it. Those are arrow properties. Moving the rest redistributes where the inconsistency appears on the target; it does not eliminate the source. What a rest change can do is shift the departure geometry until the arrow’s inconsistency partially cancels against the bow’s new configuration at the test distance. That cancellation is fragile. Change the distance, add a wind component, swap a different arrow into the same slot, and the cancellation breaks. The inconsistency resurfaces.
More directly: a bow cannot be tuned to twelve arrows simultaneously. Each arrow has its own variation. The rest position that reduces arrow 3’s horizontal bias exaggerates arrow 7’s. Group tuning finds a compromise position where the average looks reasonable. The variation across the set is not eliminated — it is redistributed around a different equilibrium. The arrows are still inconsistent. That inconsistency is now partially masked by a bow setup that was configured to mask it.
This is the wrong order. The bow should not be configured in response to the arrows. It should be configured first, from physics, and the arrows should be selected and characterized against that known reference.
Set the bow once — from intent, not feedback
The Axial configuration is set by design. The nock height is set slightly high — this element is consistent across setups. The horizontal bias direction is setup-dependent: it follows from the archer’s natural torque signature and bow geometry. The correct horizontal direction for a given setup is determined through the process in the torque-tuning article. What does not vary is the principle: a direction is chosen deliberately, from measured input, and then fixed.
The nock-high component serves the vertical bias role: arrows depart with a consistent upward tendency, removing the ambiguity at true neutral where form variation pushes arrows up on some shots and down on others. The bias establishes a direction. Variation in execution changes the magnitude of that departure, not its sign.
Set this. Lock it. Do not revisit it because a subsequent group session suggested a different rest position might close things up. The bow’s configuration is not a group-session variable. It is the fixed reference everything else is measured against.
Tune the arrows to the bow
Before fletching, every arrow in the batch is shot bare shaft at the appropriate distance for the shaft diameter — 40 yards works for shafts 6.5mm and smaller; the bare shaft selection article covers the diameter-specific reasoning. The bow’s established bias is the reference. The bow’s established bias is the reference — arrows that belong in the set land at the reference location. At 40 yards, aerodynamic forces have had enough distance to express any rotational misalignment in the shaft as a clear lateral offset. An arrow that lands off the reference group gets a nock rotation before anything else — most apparent fliers resolve with a single index change, because the spine node simply needed a different orientation relative to the string. Only arrows that won’t group at any nock orientation are pulled. Everything else gets its orientation confirmed and proceeds to fletching.
After fletching, DIAG tuning — Dimensional Isolation Arrow Group tuning — confirms each arrow in two independent dimensions. The first pass uses a vertical line and tests horizontal deviation. Only shots the archer calls center count — this filtering step removes the known archer contribution from the horizontal measurement, isolating what is left as the arrow’s behavior. Arrows with a consistent directional bias get nock adjustment in this pass. The second pass rotates the line horizontal and tests vertical deviation. In the second pass, nock adjustment is not available: rotating the nock to correct a vertical issue would undo the horizontal orientation established in the first pass. Arrows that fail the second pass are pulled from the tournament set.
The full procedure is covered in the DIAG tuning article. The relevant point here is what doesn’t happen during it: the bow is not adjusted. An arrow that fails doesn’t prompt a rest change — it prompts a nock rotation, and then a pull if rotation doesn’t resolve it. The bow’s configuration is not the variable. The arrows are being characterized to the bow, not the other way around.
What this ends
Group tuning has no defined stopping point. There is always another session, another adjustment, another configuration that looks like it might be better. The search is chronic because the signal is unreliable enough that no result ever feels definitive. Archers arrive at one rest position, then three months later arrive at a different one, then return to the first.
The Axial workflow has clear stopping points. The bow is configured once, from mechanical reasoning. Arrow selection has pass/fail criteria defined before the first shot is fired. Arrows that complete both DIAG passes are tournament arrows. Arrows that fail are practice arrows. The confirmation work is done once per batch, not restarted every session.
After this process is complete, the bow configuration and the arrow set are known quantities. When groups are inconsistent, the equipment has been characterized and is not the first place to look. The honest diagnostic target is the archer’s execution — which is where it should have been all along.
Settle the equipment. Then look honestly at what remains.
Published September 9, 2026 · Axial Bowstrings
Spotted something? Think we’re wrong? Suggest a correction →