Head angle, fork offset and trail: three quantities, two trail conventions
Read the steering axis, perpendicular axle offset and ground trail separately. Work through a fictional construction without treating a trail number as a handling score or fork recommendation.
Published 10 October 2026 · OpenBikeFit
Front geometry field guide · not to scale
The direction changes the answer.
Fictional R = 350 mm; α = 73° above horizontal; perpendicular f = 45 mm.
Normal n = R cos(α) − f ≈ 57.3 mm; ground t = n / sin(α) ≈ 59.9 mm.
Ground trail is horizontal; normal trail is perpendicular to the axis. Not a handling score or target.
On this page
- 01Name the steering axis and head angle
- 02Fork offset is measured across the axis
- 03Ground trail is not normal trail
- 04Follow one complete fictional construction
- 05Change one model input without changing its meaning
- 06Wheel radius is not a tyre-width label
- 07Keep the complete chart condition together
- 08A cockpit match does not match the whole bicycle
- 09Retain the references before any decision
Read before comparing
Keep the unresolved references on paper
An unfilled free worksheet creates no bike, selects no fork and certifies no clearance or installation.
Open toolA geometry chart might list a head angle and a fork offset while leaving trail out. Another might publish trail but omit the wheel and tyre condition used to calculate it. Those rows are related, but they do not measure the same thing. Understanding the relation starts with three named lines, not with deciding whether one number is good or bad.
The construction here is a rigid, upright side view with a straight-ahead wheel on level ground. Its dimensions are fictional teaching inputs, not recommended bicycle specifications. It explains how a declared model produces a distance. It does not test self-steering, stability, suitability, tyre behaviour or whether a fork can be fitted to a particular frame.
Name the steering axis and head angle
The steering axis is the line about which the fork turns. In this guide the head angle α is the acute angle between that line and horizontal ground, with the axis rising rearward from its ground intersection. A greater angle in this convention means closer to vertical. An angle measured from vertical would use a different number for the same unchanged line.
Write the reference explicitly: 73° above horizontal is not 73° away from vertical. A photograph of a fork blade is not a measurement of the bearing-defined steering axis, and a curved blade does not reveal that axis by itself. Use the exact chart or drawing, keep the frame size and configuration, and retain an unknown angle instead of substituting a nearby model.
Fork offset is measured across the axis
Fork offset f is the perpendicular separation from the steering axis to the front axle centre in the side view. Bike Insights calls this quantity fork offset and notes that rake is another name used for it. The axle can sit forward of the axis even when the visible fork blades look straight; silhouette and offset are separate information.
The direction matters to the arithmetic. A perpendicular offset of 45 mm is not a horizontal 45 mm shift at ground level. Do not substitute a head-tube angle, a fork blade bend or a front-centre row for this length. Rake can carry other meanings in other vehicle contexts, so preserve the diagram and unit rather than relying on the word alone.
Ground trail is not normal trail
Extend the steering axis to its intersection with the ground. Mark the ideal wheel contact point directly below the axle. Positive ground trail t is the horizontal distance from that contact point forward to the steering-axis intersection. It is a ground-line quantity; the fork offset is not measured along that line.
Mechanical trail, also called front normal trail in the cited definition, instead uses the shortest perpendicular distance from the contact point to the steering axis. Call it n here. For this specific upright construction n = t × sin(α). Both can describe the same geometry while having different numeric values. A table that says only trail needs its convention resolved before comparison.
Follow one complete fictional construction
Declare wheel radius R = 350 mm, head angle α = 73° above horizontal and perpendicular fork offset f = 45 mm. The wheel centre is R above level ground. The normal distance associated with the radius is R × cos(α); subtract the axle offset to obtain n = R × cos(α) − f. Convert that perpendicular distance to horizontal ground trail: t = n / sin(α).
Using raw trigonometry gives n ≈ 57.3 mm and t ≈ 59.9 mm. Equivalently t = (R × cos(α) − f) / sin(α). Calculate before rounding; the displayed tenths are a reading aid, not measured precision. These two answers are different conventions for one unchanged fictional front end. Neither is a target, a pass band or a claim that the bicycle handles well.
Change one model input without changing its meaning
Hold R and α fixed in that example and declare f = 50 mm instead. Ground trail becomes approximately 54.7 mm. The five-millimetre offset difference changes trail by −5 / sin(73°), approximately −5.2 mm, not simply −5 mm. The denominator records the change of direction between the two dimensions.
That comparison holds the head angle fixed as a model assumption. A real fork replacement may also change axle-to-crown dimensions, frame orientation, support height and other interfaces. Do not use the isolated offset calculation to predict every consequence of a swap. If the assumed dimensions are incompatible or unknown, the example cannot resolve the real assembly by choosing a convenient value.
Wheel radius is not a tyre-width label
R names the radius of the ideal wheel circle in this construction. A rim bead-seat diameter, a marketed wheel name and a tyre width are different inputs. A mounted tyre has its own profile and loading condition. You cannot read an exact 350 mm support radius from a printed tyre-width number, then label the resulting trail physically measured.
Separate a manufacturer-reported trail from a model-derived trail and from observations of an installed bicycle. Preserve any wheel/tyre and static or loaded assumptions given with the chart. More decimal places in a calculator do not compensate for a missing definition. A contact patch and tyre deformation also exceed the ideal point-contact drawing used here.
Keep the complete chart condition together
A frame with adjustable geometry or suspension can have several stated conditions. Keep head angle, fork identity, length, offset, wheel configuration and state from the same row or specification. Combining one size’s angle with another size’s offset produces a newly invented model, not a faithful reading of the published bicycle.
If a chart does not state whether a suspension fork is extended or loaded, ask for that information before treating differences as physical changes. This guide gives no procedure for setting sag, changing a headset or modifying a fork. Limits and permissions remain with the exact frame and component documentation, independently of whether the arithmetic can be evaluated.
A cockpit match does not match the whole bicycle
Two frames can place a named bar-clamp point at the same BB-relative X/Y using different stems, yet retain different steering axes, wheelbases and front-centre values. The existing OpenBikeFit frame example intentionally models the clamp endpoint. Its small residual does not become evidence that the front wheel geometry is the same.
Likewise, equal ground trail alone does not make every other property equal. This page does not translate category-level steering descriptions into a handling rating for a reader. A dimensional record answers what the chart describes; actual ride behaviour, component condition and suitability require other evidence rather than a green label next to a calculated distance.
Retain the references before any decision
Write frame model/year/size, configuration, angle reference, offset definition, wheel condition, reported or derived trail, its direction and the source date. Keep unknowns in the record. The diagram and the arithmetic should be reproducible from those declarations without requiring a saved rider profile or camera estimate.
Continue to the fork-length guide when a candidate has a different axle-to-crown number; use the wheelbase guide when the question is where the axles sit relative to the BB. The blank worksheet can hold technical references and questions for an exact manufacturer or mechanic. Reading, calculating this example and opening a link neither authorise installation nor certify handling.
Practical questions
Frequently asked questions
Is fork offset the same as trail?
No. Offset is perpendicular between steering axis and axle; ground trail lies along the ground between the contact point and steering-axis intersection.
Why do two trail calculations give different answers?
Check the direction first: horizontal ground trail and perpendicular normal trail differ. Also compare angle convention, wheel condition, offset definition and chart state.
Does increasing offset reduce trail by exactly the same millimetres?
Not in the declared fixed-angle ground-trail model: Δt = −Δf / sin(α). A real component change can alter other inputs and is not approved by that calculation.
Does matching trail guarantee the same handling?
No. A single geometric distance does not test the whole bicycle, loading, tyres, rider or component condition. This guide supplies no handling score or ideal trail.