Ask five shops what size bike you need, and you will get five answers, most of which are based on your height. Height is the worst single predictor of fit in cycling, and everyone in the industry knows it. Two riders at 5'10" can differ by 60mm in inseam and 80mm in arm span, which puts them on genuinely different bikes.
Here is our honest take on the sizing methods currently in use, including ours, with the strengths and the holes in each.
The default: manufacturer height charts
Every brand publishes a size chart mapping height to frame size. It exists because it has to exist, and it is the method behind the overwhelming majority of bikes sold.
One virtue is that it is usually not catastrophically wrong. Its problem is that it collapses every dimension of a human body into a single number and says nothing about how the bike is set up once it leaves the shop. A rider on the correct frame with a 35mm stem and 20mm-rise bars and a rider on the same frame with a 60mm stem and flat bars are in materially different positions. The chart cannot see that.
Charts are also increasingly at odds with modern geometry. Top tube length has become nearly meaningless as a comparison figure now that seat angles have steepened and top tubes have lengthened across the board. Seat tube length no longer tracks frame size the way it did when most brands still used it as the naming convention.
Stack and reach
The improvement on height charts is reading raw geometry: reach and stack, the horizontal and vertical distance from the bottom bracket to the top of the head tube.
This is real information, and it is comparable across brands, which is why enthusiasts live in geometry charts. The limitation is that reach and stack describe the frame, not the rider's position. Your hands do not sit on the head tube. They sit out past it, determined by stem length, stem angle, bar rise, bar sweep, and headset spacers, and none of that appears in the number you are comparing. Two bikes with identical published reach can put your hands 40mm apart.
Reach is the right starting variable. It is just not the finishing one.
Road-derived fit systems
Motion capture fit studios bring trained fitters and good tooling to the problem, and they are genuinely effective at what they were built for.
The catch is what they were built for. These systems are seated-position tools with road DNA, and much of their logic descends from KOPS, the practice of dropping a plumb line from the knee to the pedal spindle. KOPS has no biomechanical justification and never did. It survived because it happened to correlate with sensible positions for average femur lengths on 73-degree road frames. On a modern trail bike with a 77-degree effective seat angle, a strict KOPS application would have you slam the saddle rearward, undoing the geometry the frame was designed around and your climbing traction.
These systems also have very little to say about the standing position, which is where a mountain biker spends the part of the ride that matters most.
Worth having for pedaling comfort and injury work. Not sufficient for sizing a trail bike.
Rider-first frame platforms
Several brands now decouple seat tube length from frame size, building short seat tubes with long dropper travel across the range so a rider can choose a size based on how the bike handles rather than on standover and saddle height. Specialized S-Sizing is the best-known implementation and represents a real improvement in the hardware. Many riders can now legitimately consider two adjacent sizes.
Better hardware does not tell you which one to pick.
Where our thinking comes from
Credit where it belongs. The most important contribution to mountain bike fit in the last decade came from Lee McCormack, the coach and author behind Lee Likes Bikes and the book Dialed. His RideLogic method argued that sizing should be based on the two points where a rider actually contacts the bike, the bottom bracket and the grips, rather than on a frame dimension that only partly determines where those points end up.
He gave the industry two ideas worth keeping. The first is that the functional cockpit is the straight line from the bottom bracket to the grips, which already accounts for the stem, spacers, bar rise, and sweep rather than ignoring them. The second is that the angle of that line matters separately from its length, and that the angle is what tunes a bike to a riding style rather than to a body. As a rough map, he puts cross-country bikes around 55 to 56 degrees, all-around trail bikes near 58, and downhill bikes around 62 to 63.
That framework is sound, and we generally like the concept. Our disagreement is with the numbers it produces.
McCormack derives a target cockpit length from body proportions, and in our experience that calculation falls short. Riders we fit consistently prefer more cockpit than the formula predicts, often by a meaningful margin, and this is not a fringe observation. It is one of the most common reports from experienced riders who work through the method, many of whom land 40 to 50mm longer than their calculated number and stay there.
We think there are reasons for that. A short cockpit loads the front wheel well and rewards an aggressive attack position, which is exactly what the method optimizes for. But it sacrifices seated climbing comfort, closes the hip angle, and on a long ride it shifts more of the rider's weight forward than most people want to carry for four hours. It also predates the current generation of steep seat-angle trail bikes and the arrival of eMTBs, where climbing is no longer the part of the ride you endure to get to the good part.
None of this is a knock on the method. RideLogic works extremely well for many riders, and Lee McCormack has done more than almost anyone working today to shape how mountain bikers think about fit and skills. We simply hold a slightly different philosophy about where the ideal cockpit sits.
So we developed our own sizing recommendations on a similar framework, biased longer.
The TGI method
We work in two numbers.
TGI Reach is the straight-line distance from the bottom bracket to the middle of the grips. It is the functional cockpit length. Unlike frame reach it already accounts for stem, spacers, bar rise and sweep, because those parts sit inside the measurement instead of outside it. Our recommendations are derived from the rider's foot-to-hand geometry, modified by bar width, crank length, and pedal and shoe stack, and they run longer than the RideLogic figures for the same rider.
TGI Rise Angle is the angle of that same line relative to level. Same distance, different angle, and the bike behaves differently. Raising the angle brings the bars up and back, which favors technical handling and steep descending. Lowering it sends them out and down, which favors seated climbing and sustained efforts. Reach tells us whether the bike fits the body. Rise angle tells us whether the setup fits the riding.
Both are calculable from a candidate bike's geometry plus its cockpit parts, which means we can evaluate a bike a customer cannot throw a leg over. That matters for a shop that builds rather than stocks.
The number is still a hypothesis
We should be as skeptical of our own recommendation as we are of anyone else's. Any method that produces a target from body measurements is making a prediction about a specific person, and predictions get tested. A calculated cockpit length that a rider hates is not a rider who is wrong. It is a prediction that missed.
So, we generate the number, then run it against four tests.
Balance before geometry. The first test is unweighted hands. At a moderate effort, lift your hands gently off the bars and hold for several pedal strokes. A balanced position holds itself because the pedaling forces support the torso. Tip forward onto the bars, and your hands are carrying you. Feel like you are pushing the pedals forward rather than down, and you are too far back. This costs nothing and it discriminates better than any single measurement. We then put a number on it by calculating weight distribution at the front and rear contact patches, seated and in the attack position, so we are working from measured load rather than an impression of it. The target is enough weight on the front tire that it tracks and bites on climbs and in flat corners, and not so much that the rider ends up carrying the bike through their hands.
Hip angle at the top of the stroke. Thirty seconds of side-on video, scrubbed to twelve o'clock, with torso-to-femur measurement at the hip. Absolute numbers vary by protocol, so we use the delta between candidate positions rather than a target. A closed hip angle and lateral knee symptoms travel together, so this is the one we watch when a rider arrives with an IT band complaint.
Stance width, deliberately. Q-factor gets almost no attention at the point of sale, and it should. Drive units push stance width well past analog norms on most full-power eMTBs, and the spread across current motors runs from 135mm up past 175mm. A rider changing platforms can face a 40mm step change, which is large enough to be symptomatic on its own. To see what a given stance width is actually doing, we film the pedal stroke head-on and track the angle between the hip-to-knee segment and the knee-to-ankle segment through the rotation. A near-straight line through those three points is what we are after. Deviation in either direction, with the knee falling inside the foot or tracking outside it, is the signature of a stance width the rider is compensating for, and it shows up on video well before it shows up as pain. We establish a rider's current Q-factor before discussing bikes, not after.
Task validation. Nothing in a fit studio tests what a trail bike actually has to do. We pick a steep, sustained seated climb and score two things at each candidate position: does the front wheel wander, and does the rider migrate to the nose of the saddle without deciding to? On an eMTB, where weight bias sits aft, front-end loading is usually the binding constraint, and this test outranks the other two.
One variable at a time, three rides minimum. Fit changes have a fatigue lag. Day one impressions lie in both directions.
The honest summary
Research on mountain bike fit is limited. There is no validated optimum for saddle fore-aft, no consensus formula for cockpit length, and no peer-reviewed basis for most of what is sold as a fit system. That applies to our recommendations, too. We think longer is right for most of the riders we see, though our numbers still land shorter than most manufacturer suggestions and longer than McCormack's. We will tell you when the tests say otherwise on your bike.
What works is a measurement that describes the rider's actual position rather than the frame's dimensions, a set of field tests that discriminate among candidate positions, and a rider willing to run a structured experiment on themselves over a few weeks. That combination is the current best practice, not a fallback from something better.
Q-factor remains a critically overlooked variable in the point-of-sale experience, despite its direct impact on biomechanical efficiency. Modern eMTB drive units have significantly widened stance widths, with some platforms forcing a 40mm shift that can lead to subtle but persistent compensatory patterns. By filming the pedal stroke head-on to monitor the alignment between the hip, knee, and ankle, we can identify these compensations long before they manifest as chronic pain. Establishing a rider's preferred Q-factor early in the process prevents costly mismatches when switching between motor platforms. Taking this proactive approach ensures the bike's hardware respects the rider's natural anatomical tracking.