You have been told to pedal in circles. To scrape mud off your shoe at the bottom. To pull up on the backstroke. To make the force profile on your head unit look like a perfect, smooth ring of even pressure all the way round.
The problem is that almost none of that is backed by the research. Force profiling studies — the ones where scientists bolt instrumented pedals to bikes and measure exactly where power comes from in the pedal stroke — tell a completely different story. One that's less romantic but far more useful.
Here's the good news: the things that actually matter about your pedal stroke are simpler, more fixable, and less dependent on conscious technique work than most coaching advice suggests. And the biggest efficiency gains for most amateur riders have nothing to do with how you think about pushing the pedals.
Where Power Actually Comes From in the Pedal Stroke
Every pedal revolution can be divided into two phases. The downstroke, from roughly 12 o'clock to 6 o'clock, and the upstroke, from 6 o'clock back to 12. If you strap instrumented pedals to any cyclist — amateur, experienced, or professional — and measure the tangential force (the component that actually turns the crank), you find the same basic pattern every time.
The downstroke produces virtually all the propulsive power. Over 95% of the net positive work happens between about 1 o'clock and 5 o'clock, with the peak force landing somewhere between 2 and 3 o'clock. This is true of recreational riders. It is true of national-level racers. It is true of Tour de France professionals.
Prof Jim Martin at the University of Utah has spent decades studying pedalling biomechanics, and his work is unambiguous on this point: the downstroke is where the work gets done, and the differences between good and less-good pedallers are much smaller than most people assume. Trained cyclists produce a force profile that is already remarkably close to the biomechanical optimum. The body self-organises toward efficient patterns with training — no cue cards required.
This does not mean the rest of the pedal stroke is irrelevant. But it means the returns from trying to consciously change your force application during the power phase are marginal at best, and counterproductive at worst.
The dead spot problem
The transition zones — around 12 o'clock (top dead centre) and 6 o'clock (bottom dead centre) — are where the crank arm is vertical and the leg cannot apply tangential force effectively. These are the "dead spots" that coaches have been trying to eliminate for decades.
But Jim Martin's modelling work suggests that even perfectly smooth pedalling — eliminating the dead spots entirely — would only improve power output by about 2-3% at most. And attempts to eliminate them through conscious effort typically increase the metabolic cost of pedalling, because you are recruiting extra muscles to produce force at mechanically disadvantaged positions.
The dead spots are a feature, not a bug. Your body uses them for brief muscular relaxation. Trying to fill them in costs more than it saves.
The Upstroke Myth
This is where most pedalling advice goes wrong.
The idea is intuitive: if you are clipped into your pedals, you can pull up on the backstroke with your hamstrings and hip flexors, adding power that would otherwise be wasted. It sounds like free watts. It is possibly the most persistent myth in cycling biomechanics.
Kautz and Neptune's work in the early 2000s was among the first to systematically measure what actually happens during the upstroke using instrumented pedals and EMG (electromyography, which measures muscle activation). What they found was clear: during normal cycling, the recovering leg exerts a downward force on the rising pedal. Not an upward force. The rising leg is dead weight that the opposite, working leg has to push against.
When they asked riders to deliberately pull up, the riders could do it — they could generate some upward force on the backstroke. But the metabolic cost was disproportionate to the power gained. The hip flexors and hamstrings recruited during a deliberate pull-up fatigue rapidly, and the extra oxygen consumption wiped out any mechanical advantage within minutes.
Jim Martin's lab confirmed this with a different approach. They compared the pedalling patterns of elite track sprinters — athletes producing over 2,000 watts in a maximal effort — with those of recreational riders. Even at maximum sprint intensity, the elite riders did not produce meaningful upward force during the backstroke. The upstroke was less negative (less dead weight), but it was not positive.
Unweighting vs. pulling up
Here's where it gets really interesting. The upstroke does matter — just not in the way most cyclists are taught.
The distinction is between pulling up (actively generating propulsive force on the backstroke) and unweighting (reducing the downward drag of the recovering leg so the working leg has less resistance to push against). The research is clear that unweighting is beneficial and achievable, while pulling up is metabolically expensive and unsustainable.
Think about it this way. If your right leg pushes down with 400 newtons and your left leg is sitting on the rising pedal like a sandbag weighing 50 newtons, your net tangential force is reduced. If you can unweight that left leg — lift your knee, relax the ankle, let the foot come up without pressing down — you keep more of that 400 newtons of pushing force.
You are not adding power. You are subtracting resistance. The difference is about 3-5% of total energy expenditure at higher intensities, according to force profiling data from multiple labs. That's not nothing. But it's not the 20-30% improvement that the "pull up" coaching cue implies.
The practical cue is simpler than scraping mud or making circles. Think about lifting your knee, not pulling your foot. Your body knows how to lift a knee — it does it thousands of times a day when you walk. It is a natural motor pattern. Pulling your foot up against resistance with a clipped shoe is not.
Single-Leg Drills: The Evidence
Single-leg pedalling drills — unclipping one foot and pedalling with just one leg on a trainer — are a staple of indoor cycling sessions. The logic seems sound: isolate each leg, expose weaknesses in the pedal stroke, and teach the leg to produce force through the dead spots.
The evidence is less enthusiastic.
No controlled, published study has demonstrated that single-leg drills transfer to improved power output or reduced metabolic cost during normal two-legged pedalling. That's a strong statement, and it is worth sitting with. Decades of coaches prescribing these drills, and the measured outcomes in laboratory conditions show no performance transfer.
What they do show is some improvement in pedal stroke smoothness, which is a measure of how evenly force is distributed around the revolution. But smoothness and efficiency are not the same thing. You can pedal very smoothly and very expensively if you are recruiting extra muscles to fill in the dead spots.
When single-leg drills are actually useful
That said, there are legitimate uses. Single-leg drills are good at exposing gross left-right asymmetries. If one leg has a dramatically worse stroke than the other, single-leg work makes it obvious — you will feel the clunking and dead spots immediately. That's diagnostic information worth having, especially if you are coming back from an injury that affected one side.
They are also a reasonable warm-up drill. Short sets of 20-30 seconds per leg at low resistance get blood flowing and wake up the neuromuscular pathways before harder work. Used this way — as warm-up and left-right diagnostic, not as a primary training method — they earn their place in a session.
But if you are doing 10-minute single-leg sets three times a week because you believe they will transform your pedal stroke and add watts, the evidence says you would be better off spending that time on structured intervals that actually stress the energy systems you race with.
Cadence and Efficiency
If technique cues produce marginal gains at best, cadence selection has a much larger measurable effect on pedalling efficiency. This is well-trodden ground — we cover the full cadence picture in our cadence guide — but the relationship to pedal stroke mechanics deserves specific attention here.
The metabolic vs. muscular trade-off
At lower cadences (60-75 rpm), each pedal stroke requires more force. More force means more recruitment of Type II muscle fibres, which fatigue faster and consume more glycogen per unit of work. But the cardiovascular demand is lower — your heart rate is lower at 60 rpm than at 90 rpm for the same power output.
At higher cadences (90-100 rpm), the force per stroke drops. Less force means more reliance on fatigue-resistant Type I fibres and more demand on the cardiovascular system. Heart rate rises, but muscular fatigue accumulates more slowly.
The "most efficient" cadence in laboratory terms — the one that minimises oxygen consumption for a given power output — tends to fall between 60 and 75 rpm for most riders. But this is metabolic efficiency over short durations. Over the four or five hours of a road race, the muscular fatigue cost of low-cadence riding overwhelms the metabolic savings. Your heart doesn't fatigue the way skeletal muscle does. That's why trained cyclists naturally gravitate toward 85-95 rpm.
Your body already knows
Here's what matters for the pedal stroke discussion: Jim Martin and others have found that trained cyclists naturally self-select a cadence that is within about 5% of their individually optimal cadence. Your body has an internal governor that converges on the right rpm for your physiology, your fibre type composition, and the intensity you're riding at.
Forcing a cadence 15-20 rpm away from your natural selection — because a podcast or a Zwift workout told you to — almost always costs more energy than it saves. The exception is deliberate cadence training done for specific race demands, like learning to spin at 100+ rpm for the final kilometres of a crit, or practising low-cadence torque work for steep climbs.
You can check what your natural cadence distribution looks like using a cadence tool and your ride data. If you're consistently between 85 and 95 on flat to rolling terrain, you're in the range where the research says you should be.
The interaction with pedal stroke
At very low cadences, your pedal stroke necessarily becomes more "choppy" — the dead spots at top and bottom are more pronounced because there is more time for force to drop to zero and then ramp back up. At higher cadences, the stroke smooths out naturally because inertia carries the crank through the dead spots faster.
This means that much of what people perceive as "improving their pedal stroke" when they increase cadence is simply physics. The faster the crank turns, the smaller the percentage of the revolution where force drops to zero. You did not improve your technique. You changed the speed at which the same technique operates, and the math works better.
Crank Length: The Overlooked Variable
If you want to change how your legs interact with the pedal circle, changing crank length has a far larger and more measurable effect than any technique cue.
Most adult road bikes come with 170mm or 172.5mm cranks. The choice is usually based on rough rules about leg length, and many riders never question it. But crank length directly determines two critical things: the range of motion your hip and knee joints move through on every revolution, and the mechanical advantage ratio between your leg muscles and the pedal.
What the research shows
Longer cranks increase the range of hip flexion at the top of the stroke. At 12 o'clock, the knee is at its highest point, and the hip is at its most closed angle. For riders with limited hip flexibility — which includes a significant proportion of masters cyclists — this can create a mechanical pinch point that limits force production exactly where the downstroke needs to begin.
Shorter cranks (5-10mm shorter than current setup) reduce that hip angle at the top, opening up more space and allowing a smoother transition into the power phase. The research from bike fitting specialists like Phil Burt, who worked with British Cycling's Olympic squad, shows that many riders produce equivalent or better power on shorter cranks with measurably reduced joint stress.
Jim Martin's lab tested crank lengths from 150mm to 190mm and found that power output was remarkably stable across a wide range — changes of 10-15mm in either direction had minimal impact on maximum power in most subjects. What did change was comfort, sustainability, and the aerodynamic position a rider could hold. Shorter cranks allow a lower front end because the knees don't come up as high, which matters enormously for time trialists and triathletes.
Who should consider shorter cranks
This is not a universal prescription. But if you match two or more of these criteria, it is worth exploring:
- You are over 40 and have noticed decreasing hip flexibility
- You ride in an aggressive or aero position
- You experience discomfort at the top of the pedal stroke, particularly hip impingement
- Your current cranks are 172.5mm or longer and you are under 180cm tall
- You want to run a lower front end for aerodynamics
A 5mm reduction — from 172.5mm to 167.5mm, or from 170mm to 165mm — is enough to meaningfully change kinematics without requiring a bike fit overhaul. Most riders adapt within two to three rides. Some notice the difference immediately; others just stop noticing the hip tightness they had accepted as normal.
The practical point for pedal stroke efficiency is this: changing crank length alters the geometry of every single revolution by a fixed, permanent amount. No technique cue does that. No amount of concentration on your pedal stroke will change the angle your hip has to close to at the top of each revolution. But a different crank will.
Bike Fit: The Real Efficiency Lever
If there is one theme running through the pedalling efficiency research, it is this: the variables that have the largest impact on how effectively you apply force to the pedals are not about technique. They are about position.
Saddle height affects knee extension at the bottom of the stroke. Too low, and you cannot fully extend, leaving force on the table. Too high, and you reach for the pedal, rocking your hips and wasting energy on lateral movement. The research suggests a knee angle of 25-35 degrees at the bottom of the stroke (measured from full extension) is the effective range, with most riders performing best around 28-32 degrees.
Fore-aft saddle position determines where your knee sits relative to the pedal spindle at 3 o'clock — the point of peak force. Sitting too far forward shifts the load to the quadriceps at the expense of the glutes. Too far back does the opposite. The "right" position depends on your anatomy and the event demands, but the point is that a 10mm shift forward or back has a measurably larger effect on force distribution than any conscious thought about how to push the pedal.
Cleat alignment affects how force transfers through the foot, ankle, and knee. Rotational misalignment of even a few degrees can create a lateral shearing force on every stroke — thousands of repetitions per hour, compounding into discomfort and inefficiency. Getting a professional bike fit that addresses cleat position is worth more than a year of single-leg drills.
The numbers
A comprehensive bike fit that addresses saddle height, fore-aft, and cleat position can improve a rider's effective power output by 3-8% — not by making the rider fitter, but by removing mechanical losses. That is the equivalent of adding 10-25 watts at threshold for a rider with an FTP of 250-300 watts.
Compare that to the 2-3% theoretical maximum from perfecting the pedal stroke through technique work, and the priority becomes clear. Fix the position first. The pedal stroke will follow.
What to Actually Practise
None of this means you should never think about how you pedal. It means you should think about it differently, and spend your mental energy where the returns are real.
Top-of-stroke smoothness
If you are going to focus on one aspect of your pedal stroke, make it the transition at 12 o'clock — the moment the knee reaches its highest point and the leg shifts from rising to pushing. This is the dead spot that matters most, because this is where the next downstroke begins.
The cue is simple: as your knee comes over the top, think about pushing your foot forward across the top of the stroke, toward the handlebars. Not stamping down. Not mashing. A forward push that begins the downstroke with horizontal force before gravity and leg extension take over and drive the pedal down through the power phase.
This cue works because it activates the glutes and quads slightly earlier in the revolution, pre-loading the muscles before the main power phase begins. The measured effect is small — perhaps 1-2% improvement in gross efficiency — but unlike pulling up, it does not come at the cost of extra metabolic demand. It simply shifts the timing of muscle activation forward by a few degrees.
Let cadence self-select
Stop chasing a specific rpm number and start paying attention to what your body chooses. Ride at your natural cadence for most of your training. Allow it to vary with terrain and intensity — you will naturally spin faster when you are fresh and at moderate intensity, and grind slower on steep climbs when the gradient demands it.
The exception is targeted cadence work: short intervals at deliberately high or low cadences to expand the range your body is comfortable with. A few minutes of 110+ rpm spin-ups teach neuromuscular control at speed. A few minutes of low-cadence force work at 55-65 rpm build muscular endurance for steep terrain. But these are targeted tools for specific demands, not daily practice.
Use your cadence data to see your distribution over a training block. If it sits between 80 and 95 for the majority of your riding time, you're in the zone. If you are consistently below 80 on flat terrain, it is worth experimenting with shifting one gear easier and letting the cadence rise — not because 80 is wrong, but because slightly higher cadences tend to be more sustainable over longer durations for most riders.
Ride more
The least exciting but most evidence-supported way to improve your pedal stroke is to ride your bike. The body's neuromuscular system refines movement patterns through repetition. Thousands of hours of pedalling create deep-seated motor programmes that no amount of conscious cuing can replicate.
Jim Martin's data consistently shows that the pedal stroke patterns of trained cyclists — riders with years of accumulated volume — are already close to the biomechanical optimum. They did not get there by doing drills. They got there by riding. The pattern self-organised over time, shaped by the feedback loop between muscular effort, metabolic cost, and performance output.
This does not mean beginners should not think about pedalling. A rank novice who has never used clipless pedals will benefit from a few sessions of attention to the top-of-stroke transition and the basic motor pattern. But beyond that early learning phase, the returns from deliberate pedal stroke work diminish rapidly while the returns from structured training — intervals that stress your FTP zones, your VO2max, your muscular endurance — continue to compound.
The Hierarchy of Pedalling Efficiency
If you want to pedal more efficiently, here is where to put your time and money, in order of impact:
1. Bike fit. Saddle height, fore-aft, cleat alignment. The biggest lever. Get it done professionally and revisit it if your flexibility or riding position changes.
2. Ride volume. The body self-organises toward efficient pedalling patterns. More riding creates better motor programmes. There is no shortcut.
3. Cadence selection. Find your natural cadence and stop fighting it. Use targeted cadence work for specific race demands, not as a daily obsession.
4. Crank length. If you match the criteria above — masters age, limited hip flexibility, aggressive position — consider going 5mm shorter. It changes the geometry permanently, no concentration required.
5. Conscious technique. Last on the list for a reason. The top-of-stroke forward push is the only cue with consistent research support. Everything else — scraping mud, pulling up, making circles — either does not work or costs more energy than it saves.
The science has finally caught up with what the best bike fitters and coaches have known for years: you cannot think your way to a better pedal stroke. You can fit your way there. You can ride your way there. And you can stop wasting energy on cues that sound good in a YouTube video but fall apart under a force profiling sensor.
Your pedal stroke is probably closer to efficient than you think. The things holding you back are more likely your saddle height, your training consistency, or your fuelling strategy than the shape of the force curve on your Garmin screen.
That is fixable. All of it.
If you want to talk through your own pedalling data, cadence numbers, or bike fit questions with riders and coaches who have been through the same process, the Roadman Cycling community on Skool is where that conversation happens. Real data, real experience, no myths.