The Cycling Adaptation You Can't See: Tendon Stiffness and Why It Raises Your FTP

A rider I've been talking with finished eight weeks of a winter lifting block and sent me a slightly defeated message. "I think it didn't work. My legs look exactly the same. Scale hasn't moved. Quads aren't bigger." Then, almost as a footnote: "Although I did a twenty-minute test on Saturday and my FTP is up about eight watts. Probably noise." It wasn't noise. He'd spent two months building the adaptation that matters most for a cyclist, and the only place it showed up was on the bar — his back squat 5RM had climbed from 80 to 100 kilos — and on the power file.

This is the part of strength training for endurance riders that almost nobody talks about, because it's invisible. It has no Instagram surface area. But it's the single biggest reason heavy lifting raises your sustained power, and once you understand the mechanism, the weeks of "nothing happening" stop feeling like nothing.

The Mirror Lies, the Power File Doesn't

The visible part of strength training is hypertrophy: more cross-sectional area, bigger quads. For cyclists, that part is genuinely small. The work on well-trained riders is consistent — Rønnestad and colleagues' 2010 RCT, and the female-cyclist replication by Vikmoen et al. in 2016, both report quad CSA gains in the 4–7% range over twelve to sixteen weeks, with bodyweight essentially unchanged. Four to seven percent, in a leg you see every day in cycling kit, is a difference you cannot detect by eye. It looks like nothing.

What's happening underneath is different. Llanos-Lagos and colleagues' 2025 meta-analysis, pooling seventeen studies of heavy strength training in trained cyclists, found a clear effect on cycling economy and time-trial performance while VO2max barely budged. Those improvements are too big to be explained by 4–7% more muscle. Something else is doing most of the work.

That something else is mostly the tendon.

What Tendon Stiffness Actually Is

A tendon is a spring. Not a metaphorical spring — a literal one. When your quadriceps contract, the force travels through the patellar tendon, which stretches a tiny amount under load and returns that stored energy when the muscle shortens. The stiffness of that spring — how much force per unit of stretch it transmits — is a measurable, trainable property.

A stiffer tendon does two things that matter on a bike. First, it transmits muscle force to the pedal with less loss; a floppy spring eats some of the force you produce, a stiff one delivers it. Second, the stretch-shortening cycle on each pedal stroke stores and returns elastic energy more efficiently. Each stroke gets a small "free" contribution from the recoil. Multiply by sixty to ninety strokes per minute for four hours, and the integrated savings are not small.

Aagaard and Andersen's 2010 review in Scandinavian Journal of Medicine & Science in Sports lays out the mechanism: heavy resistance training increases patellar tendon stiffness, which improves force transmission and changes the way the quadriceps fascicle operates during contraction. Bohm et al.'s 2024 work in Scientific Reports, run on well-trained triathletes, gave us the cleanest recent evidence: maximal-strength training drove measurable improvements in muscle-tendon properties and increased tendon matrix remodeling. Not in beginners. In already-trained endurance athletes — the same population that says "I tried lifting, nothing happened."

The same body of work points at a second piece: rate of force development. Aagaard's earlier work reported RFD improvements on the order of 15% after roughly fourteen weeks of heavy training. On the bike, RFD shows up as the snap at the top of the pedal stroke, where you're driving past the dead spot. A faster, sharper contraction means less time spent in the inefficient part of the stroke.

Why This Raises Sustained Power

Here's the bridge to FTP.

At any given power output below maximum, your body recruits a mix of muscle fibers. Slow-twitch type I fibers do most of the steady-state work; faster, more fatigable type II fibers get called in as intensity rises or as type I fibers tire. Aagaard and Andersen describe the consequence of heavy strength training cleanly: trained type I fibers can produce a given absolute force while operating at a smaller fraction of their own maximum. The same 250 watts that used to require a higher percentage of your slow-twitch fibers' max capacity now requires less. You delay the moment when type II fibers have to be recruited. You burn glycogen slower. You stay aerobic longer. The hour-four leg lasts longer.

This is durability — the property that decides who still has watts in the final climb of a five-hour race. It does not photograph. It doesn't show up on a one-rep max either. It shows up on the power file at hour three, and on the same Strava segment ridden after a long day instead of fresh.

The tendon stiffness piece sits underneath all of this. Better force transmission per contraction means each fiber is doing a touch less work to produce the same wattage. Better elastic energy return means the muscle has to do less concentric work on each stroke. The effects compound. This is why the meta-analytic picture — economy up, time-trial power up, VO2max essentially flat — is the shape it is. The aerobic engine isn't getting bigger. The drivetrain between the engine and the pedal is getting better.

The Timeline Problem

The frustrating part of all of this is timing.

Muscle hypertrophy is relatively fast — you can see measurable cross-sectional area changes inside six to eight weeks. Tendon remodeling is slow. The collagen matrix turns over on a longer timescale, and stiffness changes typically take 8–12 weeks of consistent heavy loading to show up clearly. This is why most amateur lifting experiments quietly die at week four or five: nothing has visibly happened, the lifter assumes the program isn't working, and it gets abandoned right before the adaptation that actually matters would have started showing.

If you understand this, the early weeks stop feeling like a waste. The first month is mostly neural — the nervous system learning to recruit motor units more efficiently, which is why your 5RM climbs quickly without much visible change. The second month is where muscle CSA starts moving and tendon adaptation begins. The third month is where stiffness changes become measurable and the FTP creep starts to show. The visible stuff happens last; the invisible stuff happens first.

One of the reasons we built Dorsi to track absolute load over rep counts was exactly this — the adaptation you can't see is the one that matters most for endurance work, and the only honest proxy for whether you're driving it is whether the weight on the bar keeps climbing.

Why Light Weight Wastes the Window

The most common mistake cyclists make when they finally start lifting is doing it light. Three sets of twenty at fifty percent feels reasonable, feels like it respects the endurance background, feels like "circuit work." It accomplishes almost nothing for the adaptation we just described.

Tendon stiffness is driven by high mechanical strain. The collagen has to be loaded hard enough to trigger the remodeling cascade — fibroblasts laying down new, more organized collagen, the matrix reorganizing under repeated high-tension cycling. That signal lives above roughly 70% of one-rep max, and gets cleaner above 80%. Below that, you're producing metabolic stress your bike rides already provide in surplus, and you're not producing the mechanical signal lifting is uniquely good for. Rønnestad and Mujika's 2014 review states this plainly: cycling economy improves with heavy strength training, and does not improve with explosive or muscular-endurance protocols. There is a threshold, and circuit work sits below it.

This is why "I lift, but I keep it light because I'm a cyclist" is one of the more expensive errors in the sport. You're booking the time and the gym sessions, and you're not collecting the adaptation those sessions exist to produce.

The Quiet Part of Getting Faster

Cycling is, more than almost any other sport, an output-honest game. The power meter does not care how you look, what you weigh, or how the program is supposed to be progressing. It tells you what you produced. The mirror is irrelevant. The number on the bar matters only as a leading indicator of the number on the power file.

That's the fairness of it. The adaptation that decides whether your FTP creeps up over a winter block lives mostly in a structure you can't see, on a timeline that frustrates you, in response to loads heavy enough to be uncomfortable. If you understand what you're actually buying — a stiffer spring, a faster contraction, a slow-twitch fiber pool that carries more of the workload at any given wattage — the lifting block stops being a thing you do despite being a cyclist. It becomes the part of the year where the engine and the drivetrain quietly get rebuilt.

The rider who sent me the defeated message has now finished another four weeks. His FTP is up another six watts. His quads still look, in his words, "completely normal." He has stopped checking the mirror. He's checking the bar and the power file. Those are the right two places to be looking.