Your Knees Don't Know Where the Bottom Bracket Is
For decades, bicycle fitters have used a remarkably simple method to determine saddle fore-aft position. Place the crank horizontally, drop a plumb line from a point near the rider's knee, and adjust the saddle until the line intersects the pedal spindle. Knee Over Pedal Spindle. KOPS.
It's easy to measure, easy to teach, and produces a reassuringly definitive result. The knee is either in the correct place or it isn't, and the fitter has an apparently objective basis for deciding whether to move the saddle. There's just one problem: the measurement doesn't establish the biomechanical relationship it claims to represent.
A rider may function exceptionally well with the knee positioned directly above the pedal spindle. In fact, I wouldn't be surprised if many thoughtfully developed positions happened to produce something close to that alignment. But the fact that a position satisfies KOPS doesn't establish that KOPS is responsible for its success. The alignment may simply be a consequence of other positional decisions that were made for entirely different reasons.
And yet, I'd be reluctant to dismiss KOPS as completely meaningless. There is a legitimate mechanical argument worth considering, particularly when we examine the contribution of vertically directed forces to crank torque. The problem is that a plausible mechanical advantage, however small, is not the same thing as a universal fitting prescription. That's the distinction we've largely failed to make.
How a Measurement Became a Rule
KOPS became popular for understandable reasons. It requires little equipment, can be performed quickly, and provides an easily communicated reference point for saddle fore-aft. A plumb line, a horizontally positioned crank, and a recognizable anatomical landmark are all that's needed. Different versions of the method use slightly different knee landmarks, including the patella, tibial tuberosity, lateral malleolus, and so on. That variation alone should raise questions about how precisely the resulting alignment can represent an ideal biomechanical relationship.
Nevertheless, the method became deeply embedded in bicycle-fitting education, particularly when fitting was approached as a sequence of standardized measurements. Saddle height could be determined through one procedure, saddle fore-aft through another, and handlebar position through a third. Once the rider satisfied the prescribed relationships, the bicycle was considered appropriately configured.
There was an appealing logic to this approach. It offered a repeatable process that could be taught to bicycle shop employees and applied to a wide range of customers without requiring an extensive understanding of anatomy or biomechanics. But there's an important distinction between a repeatable measurement and a valid prescription.
A plumb line can tell us where a particular point on the knee is positioned relative to the pedal spindle. It cannot tell us whether that alignment optimizes muscular recruitment, reduces joint loading, improves power production, or creates a sustainable relationship between the rider and bicycle. Those are separate questions, and they require separate evidence. The fact that a measurement can be performed accurately doesn't establish that the result has the meaning we've assigned to it.
The Knee Is Not an Independent Reference Point
One of the fundamental problems with KOPS is that the knee's position is the consequence of several interacting variables.
Femur length, tibia length, crank length, saddle height, saddle fore-aft, ankle orientation, and the rider's pelvic position all influence where the knee appears relative to the pedal spindle. Change any of those variables and the relationship may change, even when the rider's underlying functional capabilities remain the same.
Consider two cyclists with different femur-to-tibia proportions, but similar total leg length. Both may be positioned with their knees vertically above the pedal spindle when the crank is horizontal, yet the configuration of their hips, knees, and ankles may differ considerably. The muscular demands and joint moments associated with those positions may also differ.
The same issue arises when we change crank length. Shortening the crank changes the location of the pedal throughout its rotation and alters the joint configurations required to reach it. Depending on the other dimensions of the position, the knee may appear farther forward or backward relative to the pedal spindle, even though the rider's relationship with the saddle hasn't changed.
Ankle orientation introduces another variable. A rider who pedals with a relatively plantarflexed ankle may produce a different knee-spindle relationship from someone who maintains a more dorsiflexed position, despite having otherwise similar bicycle dimensions.
This is why your position cannot be evaluated in isolation. A measurement taken at one point in the pedal stroke describes a particular configuration of the system, but it doesn't explain how that system functions.
The knee doesn't know where the pedal spindle is. It responds to muscular forces, joint moments, limb geometry, and the movement demands imposed by the rider's interaction with the bicycle. And the bottom bracket, despite its importance to the bicycle's geometry, is not an anatomical reference point.
But There Is a Legitimate Argument From Physics
Here's where I think the discussion becomes more interesting. Although I don't believe KOPS provides a valid universal prescription for saddle fore-aft, I also don't think we should dismiss the possibility that certain limb configurations offer mechanical advantages.
Consider the crank positioned horizontally forward, approximately at three o'clock. At this point, a vertically downward force applied to the pedal acts perpendicular to the crank arm. For a given downward force, this produces the greatest possible positive crank torque associated with that force direction, considering a gravitational component, which some researchers have hypothesized as signficant.
When the force acts perpendicular to the crank, its entire magnitude contributes to torque. As the direction of force approaches alignment with the crank arm, the effective tangential component decreases. This is elementary mechanics, and it would be difficult to discuss bicycle biomechanics seriously without acknowledging it.
Now consider a rider whose knee is positioned vertically above the pedal spindle when the crank is forward and horizontal. The configuration creates a potentially interesting relationship between the lower extremity and a vertically directed pedal force. One might reasonably hypothesize that this alignment offers some advantage in transmitting a downward-directed load through the leg at a point where that force has particularly favorable leverage on the crank.
It's a plausible question, and I think it deserves consideration. But there's an important qualification. The mechanical advantage of a vertically directed pedal force at three o'clock is determined by the orientation of the crank relative to the force, not by the location of the knee. Moving the knee directly above the spindle doesn't change the crank's moment arm for that force.
What it may change is the configuration of the lower extremity and the muscular and joint moments required to produce or transmit the force. That's a much narrower claim than suggesting that knee-over-pedal-spindle alignment inherently maximizes power production.
Gravity Doesn't Act Through a Plumb Line From Your Knee
There's another distinction that becomes important when we talk about the contribution of gravity to pedaling. Gravity acts on the mass of the rider's body segments. It doesn't act on the knee as though the joint were a weight suspended directly above the pedal spindle.
The amount of gravitationally related loading transmitted to the pedal depends on the distribution of body mass, the acceleration of individual segments, the forces supported by the saddle and handlebars, and the muscular actions required to maintain the rider's position. In other words, placing the knee directly above the spindle doesn't automatically place more of the rider's body weight onto the pedal. Nor does it establish that any additional load would be translated into useful crank torque.
A rider seated on a bicycle is supported through multiple contact points, and those relationships change throughout the pedal stroke. The saddle supports some portion of the rider's mass, the handlebars may support another portion, and the pedals experience forces generated through a combination of muscular action, segment motion, and external loading.
This is one reason I've always been cautious about explanations that describe pedaling primarily as pushing body weight onto the pedals. Gravity certainly matters, and body mass can contribute to the forces experienced at the pedal. But gravity isn't an independent source of continuous propulsive energy during seated cycling. The rider must repeatedly reposition and accelerate body segments, while muscular work ultimately supports the sustained mechanical output.
The fact that a downward force has favorable leverage at three o'clock is indisputable. Whether a particular knee position meaningfully improves the rider's ability to generate that force is a different question. And even if it does, the magnitude of that advantage still needs to be established.
If this feels contradictory to what I wrote above, no, you’re not crazy…it is. A little. But that’s the point, none of this is straightforward. Bike fitting isn’t straightforward. After 22 years, sometimes I’m stumped, and I’m throwing the proverbial spaghetti at the wall to see if I can introduce an improvement.
A Small Advantage Isn't Necessarily an Important One
Let's assume, for the sake of argument, that positioning the knee directly above the pedal spindle provides a small mechanical advantage during the power phase. What would that tell us? It would tell us that a particular limb configuration may improve one aspect of force production at one portion of the crank's rotation. That's useful information, but it doesn't establish that the configuration produces the greatest net power over an entire pedal stroke, much less that it represents the most efficient or sustainable position for the rider.
The lower extremity is not a rigid structure that simply transfers downward force to the pedal. The hip, knee, and ankle are connected through muscles with changing lengths, moment arms, contraction velocities, and activation patterns. A configuration that improves the mechanical conditions at one joint may increase demands at another. A position that facilitates force production at three o'clock may create less favorable conditions elsewhere in the pedal stroke.
The rider also has to maintain pelvic stability, support the torso, manage fatigue, and distribute body mass appropriately between the bicycle's contact points.
These considerations don't disappear because we've identified a potentially favorable mechanical relationship.
If satisfying KOPS requires moving the saddle to a position that compromises pelvic control, increases upper-body loading, or creates an undesirable relationship between the rider and the pedals, it's difficult to justify that change on the basis of a hypothetical advantage at one crank angle.
The size of the benefit matters, as does the cost of obtaining it. And I would be particularly cautious about assuming that a theoretically favorable alignment produces a meaningful improvement in performance without demonstrating that improvement in the rider.
Saddle Fore-Aft Is About More Than the Knee
This brings us to what I believe is the more useful way to think about saddle fore-aft. Moving the saddle changes the relationship between the pelvis and the bottom bracket. That affects the configuration of the lower extremities, but it also influences how the rider organizes the trunk, maintains pelvic control, and distributes support between the saddle, handlebars, and pedals. Those relationships are interconnected.
A rider positioned farther forward may experience a different distribution of muscular demand than someone positioned farther back. Depending on anatomy and movement strategy, the adjustment may improve or compromise the rider's ability to maintain a stable lumbopelvic system. It may also change the demands placed on the upper extremities and influence the rider's relationship with the bicycle's steering and handling characteristics.
These are not secondary considerations. They are part of determining whether a position works.
In The Saddle-Forward Power Myth, I explored the increasingly popular assumption that moving the saddle forward inherently improves power production. The problem with that argument is much the same as the problem with KOPS: it elevates one positional characteristic into a presumed explanation for performance. But the appropriate position isn't determined by how far forward the saddle sits, any more than it's determined by whether the knee aligns with the pedal spindle. It's determined by how the individual rider functions in that configuration.
For some riders, a relatively forward position may improve the interaction between the pelvis and lower extremities. For others, greater setback may provide a more stable and sustainable relationship with the pedals. Neither outcome can be established by a plumb line.
A Good Position May Satisfy KOPS Entirely by Coincidence
This is perhaps the most important distinction in the entire discussion.
Imagine working with a rider whose position has been developed through careful evaluation. We've considered their anatomy, mobility, injury history, pelvic control, muscular demands, riding objectives, and the mechanical characteristics of the bicycle. Through observation and experimentation, we've arrived at a configuration in which the rider functions comfortably and effectively.
Now imagine placing the crank horizontally and dropping a plumb line from the knee. It falls directly through the pedal spindle. Does that mean KOPS was correct? Not necessarily. It means the position happens to satisfy the measurement.
The rider's position was established through an evaluation of functional relationships, not by attempting to align the knee with an arbitrary reference point. The fact that the final configuration happens to produce the traditional KOPS relationship doesn't demonstrate that the relationship was responsible for the outcome.
It may be nothing more than coincidence.
This distinction is fundamental to scientific reasoning. An observed relationship doesn't establish a causal mechanism, and a measurement that agrees with a successful outcome doesn't necessarily explain why that outcome occurred.
If we reverse the process and begin by forcing the knee into the prescribed position, we're assuming the measurement has predictive value that it may not possess. That's a very different proposition.
Why Rules Like KOPS Survive
I don't think KOPS survived for so long because bicycle fitters were foolish or because the method was entirely without practical value. It survived because it was simple, repeatable, and easy to communicate. It provided a convenient starting point at a time when bicycle fitting was largely concerned with establishing standardized positions across a broad population of riders. And perhaps it frequently produced positions that were reasonably close to what riders could tolerate or use effectively. And maybe it sustains simply because we don’t have a better solution??? Or, rather, the masses or “at scale” don’t have a better solution.
That possibility shouldn't be dismissed. A fitting convention doesn't need to be biomechanically optimal to have some practical utility, particularly when applied to riders with relatively ordinary positional requirements. But a useful starting point is not the same thing as a universal endpoint. The danger emerges when we stop treating a measurement as a reference and begin treating it as an explanation.
A rider's knee falls behind the spindle, so we move the saddle forward. Another rider's knee falls ahead of it, so we move the saddle backward. In both cases, the adjustment is justified by the measurement rather than by evidence that the rider would function better in the resulting position. We've replaced evaluation with compliance.
This is a recurring problem in bicycle fitting, and it extends well beyond KOPS. Joint-angle ranges, symmetry scores, pressure distributions, and other measurable characteristics can become prescriptive targets when the person interpreting them loses sight of what those measurements actually represent.
I've written previously about when data becomes a shortcut for judgment, because the ability to quantify something can create an illusion of certainty that isn't necessarily supported by the underlying biomechanics. KOPS is simply an older, remarkably elegant example of the same problem.
What Should a Bike Fitter Actually Be Evaluating?
When I evaluate saddle fore-aft, I'm interested in the rider's relationship with the pedals and handlebar, but I'm not attempting to place the knee over a particular point on the bicycle.
I'm looking at how the rider organizes their body around the bottom bracket, how they maintain pelvic control, and how muscular demands appear to change as we experiment with position. I'm interested in whether they can produce force effectively without excessive compensatory movement, how they distribute support between the bicycle's contact points, and whether the position remains sustainable as workload and fatigue increase.
I also want to understand the rider's objectives. Someone seeking a sustainable position for long gravel rides may have different priorities from a competitive cyclist optimizing a relatively short, high-intensity effort. A position that represents an acceptable compromise for one rider may be entirely inappropriate for another. That evaluation requires more than measuring a joint at a single crank angle.
It requires observation, conversation, experimentation, and an understanding of how the rider responds to changes in the bicycle. This is where the balance of technology, observation, and intuition becomes important. Measurement tools can help document positional relationships, but they cannot independently establish whether those relationships are appropriate. The information becomes useful only when interpreted in the context of the rider.
Said more simply, when the rider can support themselves with minimal support from the arms/hands, fore/aft is good. The problem lately has been bike companies deciding super steep seat tube angles is the future of cycling. Sorry, but I’ve got a line out the door of riders with your new geometries and hand numbness.
I don't object to measuring KOPS. If someone wants to document the relationship between the knee and pedal spindle, that's perfectly reasonable. I object to assuming the measurement tells us where the saddle belongs.
The Knee Doesn't Know Where the Bottom Bracket Is
There is nothing inherently wrong with using bicycle-based reference points to describe a rider's position. Saddle height is commonly measured relative to the bottom bracket, and handlebar coordinates can be documented in the same way. Those measurements are useful for recording configurations, transferring positions between bicycles, and evaluating changes over time. But the coordinates of the bicycle are not the governing principles of human movement.
The knee doesn't organize its function around the pedal spindle. The hip doesn't select a muscular recruitment strategy because the saddle happens to occupy a particular horizontal coordinate. And the nervous system doesn't attempt to satisfy a plumb-line measurement while producing power.
The rider's body responds to the mechanical and physiological demands imposed by the task.
There may be circumstances in which a knee-over-pedal-spindle relationship coincides with favorable limb mechanics. There may even be a small mechanical advantage associated with a particular configuration during the power phase. I wouldn't dismiss that possibility, because the physics deserves to be considered. But a plausible advantage at one point in the pedal stroke is not sufficient to determine the position of an entire human being on a bicycle.
The objective of bike fitting should be to understand how the rider functions, identify meaningful limitations, and develop a position that supports effective, sustainable performance. If that process happens to place the knee directly above the pedal spindle, wonderful. If it doesn't, that's perfectly acceptable too.
The knee doesn't know where the bottom bracket is. And perhaps it's time we stopped pretending that it should.