Shorter cranks and saddle height: understand the conditional geometry
Derive the vertical compensation and top-of-stroke numbers, then see why setback, seatpost direction and contact layers stop them being a universal saddle instruction.
Published 9 October 2026 · OpenBikeFit
Crank field guide · not to scale
One assumption. Two contributions.
Fictional vertical model: crank 172.5→165 mm; saddle Y 740→747.5 mm.
Bottom gap 912.5 mm unchanged; top gap 567.5→582.5 mm (+15 mm).
With X −90 mm held fixed, direct top-distance change is ≈14.8 mm instead. No saddle instruction.
On this page
- 01Declare the vertical construction first
- 02Keep the bottom gap constant in this model
- 03Explain the doubled top difference
- 04A longer candidate reverses the signs
- 05Setback supplies a counterexample
- 06A seatpost move is not automatically a Y move
- 07Contact layers and separate records still matter
- 08Use the result as a declared model
Understand the model first
Inspect the existing crank comparison
The free tool retains its access, hardware and measured re-check boundaries. Reading creates no bike record or installed claim.
Open toolA crank comparison can report a shorter pedal radius, positive saddle compensation and more top-of-stroke space without modelling a rider. Those numbers follow a deliberately simple geometry. They are useful only when their endpoints and assumptions remain beside them. This guide derives the numbers used by the existing tool and then breaks one assumption to show why the same arithmetic is not a universal instruction to move a real saddle.
All examples are fictional. No part is loosened or moved by reading them, and no amount is prescribed for the reader’s bicycle. The model does not include limb lengths, foot orientation, joint angles, force, comfort or symptoms. An inseam-based product range is another convention, not an observation of the installed bicycle. Keep model output separate from a plan, an installed measurement and a rider outcome.
Declare the vertical construction first
Put the fixed BB at vertical coordinate zero. Let positive Y point upward and let the fictional saddle reference lie directly above the BB at S = 740 mm. For a rigid crank of length L, the pedal-axis top is at +L and its bottom is at −L. The vertical saddle-to-bottom gap is therefore S + L, while the saddle-to-top gap is S − L. These are distances between named model points.
This construction has no setback, no angled post movement and no added pedal/shoe stack. The saddle point is not a pelvis landmark; the pedal axis is not a foot joint. Keeping those exclusions visible prevents a formula for two mechanical points being relabelled as preserved leg extension or a verified change in a body angle. Real movement evidence belongs to a separately defined observation.
Keep the bottom gap constant in this model
Start with L = 172.5 mm and S = 740 mm. The bottom gap is 740 + 172.5 = 912.5 mm. A candidate L = 165 mm has crank delta 165 − 172.5 = −7.5 mm. If the hypothetical saddle coordinate remains 740 mm, its bottom gap becomes 905 mm, a decrease of 7.5 mm. The shorter radius has brought the bottom pedal axis toward the fixed BB.
To preserve 912.5 mm in this vertical construction, set the hypothetical saddle coordinate to 747.5 mm: 747.5 + 165 = 912.5. The model’s saddle compensation is consequently −crank delta = +7.5 mm. Positive means increasing this vertical saddle coordinate. It does not tell how far along an angled seatpost to move, which saddle surface to measure or whether the resulting position is suitable.
Explain the doubled top difference
Initially the top gap is 740 − 172.5 = 567.5 mm. With the compensated model it is 747.5 − 165 = 582.5 mm, an increase of 15 mm. Two contributions add: the top pedal axis moves down 7.5 mm and the hypothetical saddle point moves up 7.5 mm. Thus top-gap change after compensation is −2 × crank delta, not the crank delta alone.
Without that hypothetical compensation, the top gap would be 740 − 165 = 575 mm, increasing by only 7.5 mm. The two scenarios must not share one result label. The tool’s coupled result assumes the compensation; a guide reader cannot use its 15 mm as proof that the installed unchanged saddle gained that amount. Decimal arithmetic does not establish physical measurement accuracy.
A longer candidate reverses the signs
For a fictional 165→170 mm change, crank delta is +5 mm. The model’s saddle compensation is −5 mm and the compensated top-gap change is −10 mm. Zero length difference produces zero in all three model deltas. Those signs describe the declared coordinate system. They do not say that shorter is better, longer is worse or the user should select a length to make the top-gap number positive.
Write the current and candidate identities beside the subtraction. Reversing them reverses every signed delta; a number without an order is incomplete. Preserve unrounded values until the chosen display step. A component swap may simultaneously change lateral position, pedal stack or saddle reference, which this single-length comparison does not represent. Do not silently combine separate changes into one apparently isolated experiment.
Setback supplies a counterexample
Keep a fictional saddle point at X = −90 mm and Y = 740 mm. Its straight distance to the bottom pedal axis with L = 172.5 mm is sqrt(90² + 912.5²), approximately 916.9 mm. After applying the vertical-model +7.5 mm to Y and using L = 165 mm, that bottom distance remains the same because Y + L is preserved and X was held fixed.
But the top distances are sqrt(90² + 567.5²), approximately 574.6 mm, and sqrt(90² + 582.5²), approximately 589.4 mm. Their difference is approximately 14.8 mm, not exactly 15 mm. The vertical gap increased by 15 while the oblique straight distance did not. Even this simple counterexample still holds X fixed; a real seatpost movement can change X as well.
A seatpost move is not automatically a Y move
In another fictional straight-axis construction, a seatpost points upward and rearward at 73° above the rearward horizontal, equivalent to 107° counterclockwise from the forward horizontal. Moving a point 7.5 mm along that axis contributes approximately 7.2 mm upward and 2.2 mm rearward. It does not implement a pure +7.5 mm vertical change. Obtaining that vertical contribution alone would require approximately 7.8 mm along this ideal axis and still introduce a rearward contribution.
These are vector examples, not instructions for post adjustment. Rail movement, curved parts, effective seat angle and the selected surface point can further alter the record. The existing saddle-reference guide distinguishes direct BB distance from X/Y coordinates. Retain the measurement definition actually used by the saved profile instead of relabelling a model coordinate as the same installed quantity.
Contact layers and separate records still matter
The pedal axis is not the shoe sole or foot. A different pedal, cleat, adapter or shoe can change the declared contact layers independently of crank radius. The pedal-stack guide owns that question. Nor does the BB model identify spindle width, chainline or permitted frame clearance. Park Tool’s positioning chart keeps the crank length, saddle dimensions and pedal/shoe record separate, a useful evidence boundary for this comparison.
Keep each current definition and source before considering any separately justified physical work. Re-measure the same named saddle point and installed crank after a change; do not treat a model output as its own verification. An exact manufacturer compatibility chart resolves only its listed assembly questions, not rider suitability. A browser cannot inspect parts, confirm a torque procedure or certify that the bicycle was changed as intended.
Use the result as a declared model
The unchanged free tool reports crank delta, −delta compensation and −2delta compensated top-gap change. It can then offer its existing bike-linked planning and re-check workflow, with separate gates and local records. Reading this guide passes none of them. Opening the blank worksheet creates no installed claim, and reading an external manufacturer source sends an ordinary request to that site, not an offline inspection.
The complete hand-off should keep the candidate identity, model assumptions, original baseline and intended measurement definitions together. An independently observed before/after result is new evidence, not an automatic validation of every model assumption. This guide offers no watts, health outcome, joint-angle prediction or clearance approval. Resolve the real references before deciding whether any mechanical trial is justified at all.
Practical questions
Frequently asked questions
Does a shorter crank always mean raising my saddle by the same amount?
No universal instruction follows. The tool’s −delta result preserves a vertical saddle-to-bottom-pedal-axis gap in its simplified construction, not every real saddle dimension or leg motion.
Why does a 7.5 mm shorter crank give 15 mm top-gap change?
Only after the model’s +7.5 mm vertical compensation: the top axis moves down and the saddle point moves up. Without compensation, the vertical top gap changes by 7.5 mm.
Is vertical saddle height the same as BB-to-saddle tape distance?
No. Setback makes the direct distance oblique, and an angled seatpost move can change both X and Y. Preserve the exact named reference.
Does more modelled space prove a better joint angle or comfort?
No. The model has no rider kinematics or outcome evidence. Independently recorded movement and riding observations remain separate questions.