The 50 Percent Rule: Low-Load Isometrics Build Stronger Tendons
Half effort, held long enough, loads the whole tendon. Max effort only loads the part that was already strong.
Ask most people to do an isometric and you get the same thing: brace as hard as humanly possible, shake for five seconds, drop, repeat. That is a fine way to train your nervous system. It is a poor way to train the tissue that actually transmits the force.
The work of Dr. Keith Baar, a physiologist at UC Davis who studies connective tissue and previously worked as a strength coach in college football, keeps landing on an unintuitive conclusion. For tendon, the useful dose is not maximal. It is moderate, and it is long.
Why five-second max holds miss the tendon
A tendon is not a uniform rope. It is a bundle of fibres at slightly different lengths and stiffness. When you go hard and fast, the stiffest, shortest, strongest fibres take the load first, and because the whole thing happens in a couple of seconds, they take almost all of it.
The weaker fibres never get stressed. They also never get the signal to remodel. Over months of explosive training you end up reinforcing the strong part of a structure whose failure point sits in the part you keep skipping.
Explosive movement makes this worse in a second way. Rapid force application creates high jerk, the rate at which acceleration changes. Jerk concentrates stress at the tissue interfaces and drives the wear-and-tear side of the ledger without adding much to the building side.
What 30 seconds does that 5 seconds cannot
Hold a moderate load and the tendon does something a rope does not: it creeps. Under sustained tension the stiffest fibres slowly lengthen. As they give, the slack fibres beside them come taut and start carrying their share.
Somewhere past the 20 to 30 second mark, most of the structure is genuinely under load. Baar describes imaging work where short explosive contractions light up only narrow bands of tendon cells, while a long moderate hold lights up the whole tissue.
That is the entire mechanism in one sentence: time under moderate tension is what recruits the parts of the tendon that effort alone will not reach.
The dose-response surprise
The part that reliably raises eyebrows in our coach meetings is the comparison between maximal dynamic work and moderate holds at matched time under tension. Both produced similar anabolic signalling. The tissue outcome was not similar at all.
At full effort with dynamic, high-jerk contractions, the trained tendon finished weaker than the untrained control. The breakdown signal ran ahead of the growth signal. At around 50 percent with sustained holds, the trained tendons came out stronger.
More is not better here. More is a different outcome.
Low load and heavy work are additive, not competing
This is where a lot of the online summarising goes wrong. Nobody is telling you to stop lifting heavy.
In Baar's work with rock climbers, three groups trained grip for a month. Low-load hangs alone produced about a 3 percent strength gain. Heavy max hangs alone produced about 3 percent. Doing both produced about 6 percent. The two methods were not redundant, because they were solving different problems: one builds contractile strength, the other builds the transmission system underneath it.
| Heavy slow resistance | Low-load long-duration isometrics | |
|---|---|---|
| Main target | Muscle strength and tendon stiffness | Whole-tendon collagen remodelling |
| Typical dose | 3 to 5 sets, 6 to 8 reps, 3 sec down | 4 sets, 30 sec at ~50 percent |
| Next-day cost | Real fatigue and some soreness | Almost none |
| In-season fit | Needs careful placement | Fits anywhere, including race weeks |
| Frequency | 2x per week | 3x per week |
How to find 50 percent without a force plate
You do not need lab equipment. Three practical ways we use it in class:
- Load maths. If your best five-rep squat is 100 kg, a wall-supported Spanish squat or a paused position around 45 to 55 kg of equivalent tension is the neighbourhood. Precision is not the point.
- Effort dial. Push to what feels like a 5 out of 10. You should be able to hold a normal conversation through the full 30 seconds. If you are gritting your teeth at second 12, it is too much.
- The last-five-seconds test. Second 25 to 30 should feel like work but never like failure. If you cannot finish the set clean, drop the load rather than the clock. The clock is the active ingredient.
Five positions we actually use in Olympic Village
Every one of these is 4 rounds of 30 seconds, 3 seconds on, 3 seconds off, resting 60 to 90 seconds between rounds.
- Spanish squat wall hold for the patellar tendon. Band behind the knees, shins vertical, sit into about 60 degrees.
- Mid-range calf raise hold for the Achilles. Bent knee one session, straight knee the next, so you catch soleus and gastrocnemius both.
- Dead hang or half-bodyweight hang for the finger flexors, elbows, and shoulder connective tissue. Scale by standing on a box and taking weight off with your feet.
- Ring or bar support hold for the elbow and shoulder. Arms locked, shoulders down, breathing normally.
- Split-squat isometric for the hip and adductor tissue, especially useful for runners and HYROX athletes carrying sled volume.
Not sure which one to start with, run the tendon readiness check and let the answer pick your position for you.
Where it goes in the week
The best argument for this protocol is not the strength number. It is the recovery cost, which is essentially zero. Sustained isometrics produce very little delayed-onset soreness, so unlike heavy eccentrics they do not steal from tomorrow.
That makes them one of the few things you can genuinely add to a full week without subtracting something else.
One useful add-on from Baar: a sustained hold clamps down blood flow to the area. Following each set with 30 to 60 seconds of light, unloaded movement (a few Copenhagen raises, easy cable work, ankle pumps) acts as a pump and restores circulation between rounds.
What this does not fix
Low-load isometrics are a building tool and a useful early rehab tool. They are not a diagnosis.
- If pain is sharp, one-sided, or worse week over week, see a physiotherapist. Ten more sets will not out-work a structural problem.
- If the tendon is hot, swollen, or painful at rest, that is a reactive stage and needs professional guidance on load, not a self-prescribed protocol.
- Tendon remodelling is measured in months, not weeks. Collagen turnover is slow. Give any honest attempt at this 12 weeks before you judge it.
None of this is medical advice. Persistent joint or tendon pain deserves a conversation with a qualified clinician, and we are glad to work alongside yours.
Where to go next
If you want the wider picture, our complete guide to tendon training and joint health covers heavy slow resistance, eccentrics, plyometric progression, sleep, and the collagen and vitamin C timing question in full.
If you would rather just be handed a plan, the free 4-Week Tendon Loading Starter lays out the isometric, heavy slow resistance, and eccentric progression day by day.
And if you are in Vancouver and want a coach watching the positions rather than guessing at them from a screen, come try a class in Olympic Village. Bring the cranky knee. We have seen it.
Sources
- Baar K. (2017). Minimizing injury and maximizing return to play: lessons from engineered ligaments. Sports Medicine 47(Suppl 1):5-11. PubMed
- Baar K. (2019). Stress relaxation and targeted nutrition to treat patellar tendinopathy. International Journal of Sport Nutrition and Exercise Metabolism 29(4):453-457. PubMed
- Kubo K, Kanehisa H, Fukunaga T. (2001). Effects of different duration isometric contractions on tendon elasticity in human quadriceps muscles. Journal of Physiology 536(2):649-655. PubMed
- Rio E, et al. (2015). Isometric exercise induces analgesia and reduces inhibition in patellar tendinopathy. British Journal of Sports Medicine 49(19):1277-1283. PubMed
- Interview remarks and unpublished protocol detail attributed to Dr. Keith Baar are drawn from his public podcast and interview appearances discussing low load, long duration isometrics.