How BikeSize calculates bike size

Inseam does the work, not height

Height is a poor predictor of frame size on its own. Two riders at 5'9" can differ by 6 cm of inseam, and that difference is the whole ballgame for standover and saddle height. So the calculators are built around inseam: stand against a wall in socks, pull a book up firmly into your crotch, and measure from the top of the book to the floor.

If you only give us height, the road and mountain calculators estimate inseam as height x 0.468 and say so. That estimate is a fallback, not a feature. It carries the average leg-to-torso ratio of the population, which is exactly the thing that made height unreliable in the first place. Measure the inseam if you can.

The formulas, by calculator

These multipliers are the long-standing conventions of bike shops and fit literature, applied consistently rather than reinvented. Each row matches the code that runs when you press the button.

What the adjustments are, and are not

The riding-style and terrain nudges are deliberately small, usually a centimetre. They exist because a racer on a slammed stem and a weekend gravel rider carrying camping gear want different amounts of reach from the same legs. They are not a hidden model. If you see plus 1 cm for endurance riding with low flexibility, that is the entire rule, and it is the only thing separating the two answers.

Where a calculator snaps to a stocked size, it rounds to what manufacturers actually build. Frames do not come in 54.7 cm.

The triathlon calculator, in full

The triathlon tool is the one that does more than multiply an inseam, so it is worth spelling out. Frame size is inseam x 0.65, and choosing a full Ironman adds 1 cm on top, because a longer day in the aerobars is easier on a fractionally larger frame.

Everything else on that result card is a target position rather than a frame size, and it comes from a different set of inputs:

• Stack and reach targets. With height alone, stack is height x 3.1 plus 20 mm and reach is height x 2.15 plus 5 mm. Enter a torso measurement and both are recalculated from it instead: stack becomes torso x 8.2 plus 50, reach becomes torso x 5.8 plus 30. Torso length predicts how far you can stretch out better than total height, so it takes over when you provide it.
• Arm length. Pad width is arm length x 0.36. Arms longer than 65 cm add (arm minus 65) x 0.8 mm of reach, and arms shorter than 60 cm subtract (60 minus arm) x 0.8. With no arm measurement, pad width falls back to height x 0.13.
• Race distance. Sprint lowers the stack target by 10 mm, Olympic by 5 mm, half Ironman leaves it alone, and full Ironman raises it by 10 mm.
• Seat tube angle. This is a published range for the distance, not a computed number: 78 to 80 degrees for sprint, 77 to 79 for Olympic, and 76 to 78 for half and full Ironman.

Two limits are worth stating plainly. These coefficients are fit-shop conventions scaled to body measurements, and they rest on thinner ground than the frame-size multipliers above, so they get you to a plausible aerobar position to test rather than a fitted one. And the frame size and the position targets are computed independently, not read off one frame model, so a bike that matches your stack and reach may still carry a different size label. Use the position numbers to filter bikes whose published geometry you then check yourself.

Nothing in this calculator sees hip angle, shoulder mobility, or how a position affects your run off the bike. Those are the reasons a triathlon fit exists, and this is the calculator on the site with the strongest case for booking one.

How accurate this is

A frame-size estimate from inseam lands you in the right size for most adult riders, and within one size for nearly everyone else. That is genuinely useful when you are narrowing a shortlist or checking whether a bike on a classifieds listing is worth a drive.

It is not the last word, because the math never sees:

• Your torso and arm length, which decide reach far more than inseam does
• Hip mobility, back history, and how long you actually ride at a stretch
• A brand's own geometry, where two frames labelled M can differ by more than a size step
• Toe overlap, cleat position, saddle setback, and everything else that only shows up on the bike

The last one is worth repeating. A number here is a starting point for a test ride, not a substitute for one. Where a brand publishes its own chart, the brand wins. That is why every brand size chart on the site sits next to the calculator rather than under it.

When to see a professional fitter instead

Go see a fitter, and treat anything on this site as background reading, if any of these apply:

• You have knee, back, neck, or hand pain that shows up on the bike
• You are between sizes on the brand chart and the two candidates feel equally plausible
• You are buying a triathlon or time-trial bike, where position is the point and the adjustment range is narrow
• You are spending enough that a few hundred on a fit is a rounding error against the frame
• You are returning to riding after injury, surgery, or a long time off

A fitter watches you pedal. A web form cannot. We say this on the calculator pages too, and it is in the editorial policy: BikeSize is not a fit studio.

What we do not do

• We do not lab-test or ride-test bikes. Nothing here comes from hands-on testing, and no page should imply that it does.
• We do not invent geometry. If a manufacturer has not published a measurement, the field stays empty.
• We do not tune formulas to steer you toward a product. Affiliate links pay for the site, and the affiliate disclosure explains how they are marked. They do not touch the arithmetic on this page.

Corrections

If a formula described here does not match what a calculator returns, that is a bug worth reporting, and it is the kind of thing this page exists to make visible. Open an issue on the public issue tracker. Reports are handled in the open. We aim to respond within five working days, and a confirmed error in a published number gets fixed on the page rather than quietly patched in code.

This page is checked against the calculator source whenever a formula changes, and at minimum once every twelve months, on the same review cycle described in the editorial policy.