← Longevity & Supplement Guides
Eccentric training: the half of the rep most people throw away
Watch any gym for ten minutes and you will notice something odd: nobody counts the way down. The lift is the performance — the grimace, the exhale, the number that goes in the app. Then the weight comes back down because gravity is available and free, and the rep is declared over.
Half of every repetition is being thrown away. And it happens to be the half in which your muscle is strongest.
Lowering is not the rest of the rep
When a muscle shortens under load, that is a concentric contraction — the curl, the push, the stand. When it lengthens under load, holding something back rather than moving it, that is an eccentric contraction. Sitting down slowly. Walking downstairs. The lowering half of a push-up.
Eccentric contractions are physiologically strange in a useful way. A muscle can generate more than 20% more force while lengthening than while shortening or holding still, and it does so at a lower metabolic cost — less oxygen consumed, lower heart rate, for the same external workload. Kazunori Nosaka, who has spent a career on this, laid the whole picture out in a January 2026 review in the Journal of Sport and Health Science.
Put plainly: your capacity to absorb force is larger than your capacity to produce it. And almost all of our training is aimed at the smaller number.
Three seconds a day, and the number that came back
The study that made this idea travel is deliberately absurd in its smallness. In 2022, Sato and colleagues took young sedentary adults, sat them at an isokinetic dynamometer, and had them perform one three-second maximal contraction of the elbow flexors, once a day, five days a week, for four weeks. Thirteen people per group. One group did the contraction concentrically, one isometrically, one eccentrically. A fourth group just showed up to be measured.
Total training time across the month: about a minute.
The eccentric group increased maximal isometric torque by 10.2%, concentric torque by 12.8%, and eccentric torque by 12.2%. The concentric group improved on exactly one measure, isometric torque, by 6.3%. The isometric group improved on one measure, eccentric torque, by 7.2%. The control group did not change.
One detail rarely survives the retelling: muscle thickness did not change in any group. The headlines that say three seconds a day builds muscle are describing a study that specifically measured muscle size and found none. What it found was strength — which is largely a nervous-system story in the first month of anything.
Figure — the dose that actually worked
One 3-second maximal eccentric contraction a day, elbow flexors, 4 weeks
Change in maximal voluntary contraction torque after 4 weeks. Bar widths are the reported percentages on a shared 0-15% axis: +12.8%, +12.2%, +10.2% (5 days a week, Sato et al. 2022); +3.9%, +2.5% (3 days a week, Yoshida et al. 2024). The 2-days-a-week group is drawn as an empty track because it produced no significant change — an absence, not a small number. Untrained young adults, elbow flexors only, no change in muscle thickness in any group.
Take two days away and it mostly disappears
This is the part that should have travelled with the headline and did not.
The same research group went back and asked the obvious follow-up question: does the three seconds still work if you do it less often? In a 2024 paper in the European Journal of Applied Physiology, twenty-six adults did the identical single three-second maximal eccentric contraction, but at two or three days per week for four weeks.
The two-day group showed no significant change in any strength measure. The three-day group improved — concentric torque by 2.5%, eccentric torque by 3.9% — statistically real, and roughly a quarter of what the five-day protocol produced.
So the effective dose was never "three seconds." It was three seconds, on nearly every day of the week. The variable doing the work was frequency, and the viral version of the finding quietly dropped it. This is the ordinary shape of exercise news: the striking number survives, the condition that made it true does not.
Twenty-seven studies later, the advantage narrows
If eccentric contractions are stronger and cheaper, the tempting conclusion is that eccentric training is simply better training. The pooled evidence says something more precise and less exciting.
In August 2025, Spudić and Nosaka published a systematic review in Sports Medicine — Open covering 27 randomised trials and 162 separate results, all comparing eccentric-only against concentric-only isokinetic training in healthy adults. Eccentric-only training was clearly superior for one thing: eccentric strength, which rose 27% versus 10%, a large effect. For concentric strength the two were indistinguishable — 13% versus 14%. For isometric strength, likewise — 18% versus 17%.
Read that again, because it is the honest centre of the topic. Eccentric training does not make you generally stronger than conventional training does. It makes you specifically, substantially better at the thing it trains: controlling load while lengthening. That is specificity, not magic, and specificity is worth knowing about rather than being disappointed by.
What "better at lowering" is actually worth
Ask what fails first, and the case reassembles itself.
Almost nothing that goes wrong in an ordinary body goes wrong on the way up. People do not typically collapse while standing from a chair; they land in it. The stumble that becomes a fall is a failure to decelerate. Going down the stairs is harder on the knees, and worse for the confidence, than going up. Carrying a suitcase down to the platform is the part people plan around.
Deceleration is eccentric work, and it is the capacity we neither train nor measure. Our post on muscle power versus strength as you age makes an adjacent point: the number that predicts how well someone moves through the world is rarely the number they lift.
The applied evidence is modest but consistent. Nosaka's review notes a 12-week trial in older women in which descending stairs twice a week produced considerably larger gains in maximal isometric strength than ascending them — 34% against 15% — with the descent introduced gradually enough to avoid meaningful soreness. Same staircase, opposite direction, different stimulus.
Soreness is a toll, not a fee
The reason eccentric work has a reputation problem is that unfamiliar eccentric work makes you ache. Delayed onset muscle soreness peaks between one and three days afterwards, and for most people that is the entire association.
Two things about that soreness are worth internalising. First, it is largely a one-time toll. The repeated bout effect means a single session of eccentric exercise confers protection against damage from the next similar session, and that protection lasts weeks to months. Second — and this genuinely surprises people — the soreness is not a readout of torn muscle. It is a sensitisation of nociceptors in connective tissue, driven by signalling molecules such as nerve growth factor. Severe soreness can occur with little fibre damage, and meaningful adaptation can occur with little soreness.
Which means damage is not the mechanism and soreness is not the scoreboard. Nosaka's practical conclusion is that the toll can be made small by starting at low intensity, low volume and slow speeds and building from there — the same unglamorous advice that governs how tendons adapt. Usefully, older adults appear to experience less eccentric muscle damage than younger adults, not more.
How to add it without redesigning anything
The good news about this literature is that it asks for a change of attention rather than a change of programme. You already do eccentric contractions in every set you perform. You are simply performing them badly, at speed, without looking.
- Count three on the way down. Same exercise, same load, roughly one second up and three seconds down. In the taxonomy this is "prolonged eccentric loading," and it costs nothing but attention.
- Aim for most days, not long days. The frequency study is the clearest practical finding here: two days a week did nothing, three days a week did something. Small and often beat large and rare.
- Use the stairs downward. If you take the lift up and the stairs down, you have chosen the eccentric half on purpose. It pairs neatly with what we wrote about stair climbing and longevity.
- Bodyweight is enough to start. Slow chair squats, wall push-ups lowered under control, heel drops off a step. These are the movements used in home-based eccentric programmes, and they need no equipment.
- Progress control before load. Technique and tempo first, then range, then weight or speed. Expect a sore week when you introduce something new to a limb; expect it not to repeat.
And keep the caveats where you can see them. The three-second studies used a dynamometer, one small muscle group, and untrained young people — a machine that fights back is not the same as a dumbbell, and beginners improve at almost everything. The pooled review is isokinetic training, not barbells. None of this has been shown to beat a sensible weekly strength habit; it is a way to get more out of one.
The bottom line
Eccentric training is not a shortcut, and the three-seconds-a-day version of the story is missing its most important word, which is "daily." What the evidence supports is narrower and more durable: the lowering half of a movement is a real training stimulus, your body is strongest there, it costs less metabolically than the lifting half, and almost nobody is doing anything deliberate with it.
You do not need a new programme. You need to stop letting gravity finish your sets. Count to three on the way down, most days, in whatever you already do — and notice, a few weeks later, whether the stairs feel different going the other way.
