How Long Should I Be Able To Stand On One Leg? (The Real Chart, Not The Viral One)

Under 60, you should clear 40-plus seconds per leg with your eyes open. That is what the published normative data says — men aged 18 to 39 averaged 44.4 seconds on their best of three attempts, women 45.1, out of a possible 45, because the trial was capped there. The famous 10 seconds is not a benchmark for a healthy 35-year-old at all. It is a pass/fail line from a mortality cohort aged 51 to 75, where the only thing recorded was whether you completed 10 seconds. And the seconds-by-age chart filling page one — nearly a minute in your 30s and 40s, 45 seconds in your 50s, 26 seconds in your 70s — traces back to an unnamed "Duke Health study" that I went looking for and could not find. Two different tests, two different populations, one mashed-up chart.

How long should the average person be able to stand on one leg?

With the eyes open, a healthy adult under 60 stays above 40 seconds on the best of three, both legs. After that it falls fast.

The reference set is Springer and colleagues, 2007, "Normative Values for the Unipedal Stance Test with Eyes Open and Closed," in the Journal of Geriatric Physical Therapy: 549 healthy subjects aged 18 and up — 258 female, 291 male — in six age groups, three trials per condition. Inter-rater reliability was 0.994 eyes open and 0.998 eyes closed — two testers, two stopwatches, same answer.

Eyes open, best of three then mean of three, in seconds:

• 18-39. Men 44.4 (4.1), mean 43.2 (6.0). Women 45.1 (0.1), mean 43.5 (3.8).
• 40-49. Men 41.6 (10.2), mean 40.1 (11.5). Women 42.1 (9.5), mean 40.4 (10.1).
• 50-59. Men 41.5 (10.5), mean 38.1 (12.4). Women 40.9 (10.0), mean 36.0 (12.8).
• 60-69. Men 33.8 (16.0), mean 28.7 (16.7). Women 30.4 (16.4), mean 25.1 (16.5).
• 70-79. Men 25.9 (18.1), mean 18.3 (15.3). Women 16.7 (15.0), mean 11.3 (11.2).
• 80-99. Men 8.7 (12.6), mean 5.6 (8.4). Women 10.6 (13.2), mean 7.4 (10.7) — a spread wider than the mean itself, which is why "the average 85-year-old" is close to a useless idea.

One footnote on those brackets: Springer's table labels them standard error, but the magnitudes only work as standard deviations — a standard error of 13.2 on 22 people is impossible.

Over 60 there is a second, independent body of data. Bohannon's 2006 meta-analysis pooled 22 studies and 3,484 participants aged 60 to 99, eyes open, for a mean of 15.7 seconds: 27.0 (95% CI 20.4-33.7) at 60-69, 17.2 (11.6-22.8) at 70-79, 8.5 (1.0-16.1) at 80-99.

Is there a real one leg balance test chart by age?

Yes — and it is almost certainly not the one you found. The real chart is Springer's, and the single most important thing about it never gets copied across: the trial was capped at 45 seconds.

That changes what the numbers mean. A male 18-39 best-of-three of 44.4 out of a possible 45 does not say men in their twenties balance for 44 seconds. It says nearly all of them stood there until the tester stopped the watch. It is a pass mark, not a measure of capacity: nobody knows how long those men could have gone.

Arm position is the other rule, and not optional if you want to compare yourself to these numbers. Springer's subjects stood barefoot with their arms crossed over the chest, eyes on a spot at eye level, and the timer stopped the moment they uncrossed. Araujo's kept the elbows extended and the arms at the sides instead. Pick one and stay in it — windmilling your arms means you are not running either test.

Bohannon's 27.0 seconds at 60-69, from 851 people, lands a tenth of a second off Springer's 60-69 mean of three, 26.9 — agreement that close is what none of the viral charts can show you. He has his own ceiling problem: the 80-99 sample was only 146 people, he used each study's best performance, which inflates the values, and most of the component studies capped trials at 30, 45 or 60 seconds, which "can result in the underestimation of average performance."

Now the chart you saw. AARP writes that "in one study led by researchers at Duke Health, adults in their 30s and 40s could balance on one leg for nearly one minute. Adults in their 50s were able to stand on one leg for about 45 seconds, and those in their 70s for 26 seconds." No journal, no year, no authors, no DOI, and no paper I could find matches those numbers. I am not calling it fabricated — I am telling you I could not verify a single number in it.

One thing I can rule out: it cannot come from the Mayo Clinic study cited alongside it, because that study enrolled nobody under 50 and capped its trials at 30 seconds. It cannot produce "nearly a minute in your 30s." Same pattern as the dead hang age charts — a clean number per decade is easy to publish and impossible to source, so it outranks the real data.

What does the 10-second balance test actually measure?

Whether you can hold one supervised position for ten seconds, recorded as yes or no. That is the entire measurement, from a study never designed to grade a healthy 35-year-old.

Araujo and colleagues, 2022, in the British Journal of Sports Medicine, tested 1,702 adults aged 51 to 75 — 68% men, mean age 61.7 ± 6.8. The protocol is eyes open, which nearly every fitness site gets wrong: barefoot, the non-support foot resting against the back of the opposite lower leg, arms at the sides, gaze fixed two metres away at eye level, up to three attempts.

348 people, 20.4%, failed, and failure climbs steeply: 4.7% at 51-55, 8.1% at 56-60, 17.8% at 61-65, 36.8% at 66-70, 53.6% at 71-75. Over a median follow-up of 7 years (IQR 4.16-9.41), 123 people died, 7.2% — 17.5% of the fail group against 4.6% of the pass group, p<0.001. Vital status came from the official regional registry, so nobody was lost to follow-up. Age-adjusted hazard ratio 2.18 (95% CI 1.48-3.22); fully adjusted 1.84 (95% CI 1.23-2.78), p=0.003, after age, sex, BMI, coronary artery disease, hypertension, diabetes, obesity and dyslipidaemia.

Here is the part nobody republishes: the authors tested whether the ten-second result adds anything to a model that already knows your age, sex, BMI and comorbidities. Harrell's C-index moved from 0.7990 to 0.8090 — a gain of 0.0100 (95% CI −0.0005 to 0.0205), p=0.06. Continuous net reclassification improvement was 17.50% (95% CI −1.46 to 36.45), p=0.07. Two measures did clear it: the integrated discrimination improvement at 0.0143 (0.0019-0.0267), p=0.024, and a −2 log likelihood comparison at p=0.002 — the two the authors lean on to conclude the test adds real prognostic information. The honest summary: the value beyond what a doctor already knows about you is real but small. Diabetes was also three times more common in the fail group (37.9% vs 12.6%), so part of what this test reads is peripheral neuropathy. And the cohort came from one private exercise-medicine clinic in Rio de Janeiro, mostly self-referred and higher-income, which the authors say should make you cautious about extrapolating.

Then the clinical framing. America's official falls screen, the CDC's STEADI 4-Stage Balance Test, runs four eyes-open positions of 10 seconds each — feet side by side, instep to big toe, tandem heel-to-toe, single-leg stance — stopping at the first failure. Its threshold: "An older adult who cannot hold the tandem stand for at least 10 seconds is at increased risk of falling." That cut-off is at stage three, the tandem stand. The CDC uses single-leg stance as its hardest stage and publishes no risk threshold for it at all. "10 seconds on one leg is the clinical danger line" is borrowed from a Brazilian mortality cohort, not from the falls screen people assume it came from.

Why is standing on one leg so much harder with your eyes closed?

Because vision is doing most of the work, and you don't notice until you switch it off. In Springer's data, the same men who held 44.4 seconds eyes open managed a best of 16.9 (13.9) eyes closed, with a mean of 10.2 (9.6). Women went from 45.1 to a best of 13.1 (12.3), mean 8.5 (9.1). That is two thirds to three quarters of your time, gone, from closing your eyes.

Balance runs on three inputs: vision, the inner ear's read on head position, and what the receptors in your feet and ankles report. Eyes open, a fixed point at eye level is a continuous error signal and the ankle corrects before you feel anything. Close your eyes and you are on the vestibular system and proprioception alone, a much noisier loop.

It drops further with age: women aged 60-69 posted a best of 3.6 seconds (2.3), mean 2.5 (1.5), and at 80-99 a best of 2.1 (1.1), mean 1.4 (0.6). Read that before you test your parents and scare everyone — under three seconds eyes closed is the published normal range for a healthy woman over 60. It is a different test with different norms, and the one people most often fail and then misread.

Does balance really decline faster than strength as you age?

In the only study I can find that measured both in the same people, yes — balance fell about twice as fast as strength. It is also a study of 40 people. Both halves of that sentence matter.

Rezaei, Kaufman and colleagues, 2024, in PLOS ONE, tested 40 healthy independent adults over 50 in Rochester, Minnesota — split evenly between 50-to-64 and 65-plus, and evenly by sex — on a force plate, eyes open, with a 30-second maximum. In standard deviations of change per decade: unipedal stance −0.62 non-dominant and −0.53 dominant, against grip strength at −0.34 and knee extension at −0.26. Stance was the measure most affected by age (p ≤ 0.001), falling 2.2 per decade non-dominant and 1.7 dominant in the paper’s normalised units. Walking speed and stride length showed no significant age-related decline, which drove the headlines. The limits belong in the same breath as the figures: n=40, cross-sectional — different people at different ages, nobody followed — one Minnesota cohort, and a 30-second ceiling that healthy people hit. "Balance is the best marker of aging" is a reasonable read of a 40-person cross-section and an unreasonable thing to state as fact. What survives is the direction, which Springer and Bohannon back: stance time collapses between 50 and 80 while nobody's grip does anything that dramatic.

How do you actually improve your one leg balance?

You train on one leg, under load, on purpose — then test yourself barefoot with your arms crossed on your chest, which is a different thing from practising the test.

Most of my own leg volume is still unilateral, and not because of balance. Two-legged lifts hide imbalances; train the legs separately and the weakness has nowhere to hide. My view is that asymmetries are where knee and hip problems tend to start. I'm 37, lifting heavy four days a week, running half marathons, zero joint issues — unilateral work is a big reason why. Here's the single leg workout I've been running recently, move by move. In a session:

• Split squats and Bulgarian split squats, 2-3 hard sets per leg — where the ankle and hip learn to stabilise against real force.
• Step-ups and single-leg RDLs. The RDL puts the stance ankle under a slow, long lever, the same demand the test reads.
• Barefoot work, eyes-closed holds last. A stiff sole stabilises you, which is why the test is barefoot.

A strong lower body does not hand you this for free. A 315 squat predicts approximately nothing about how long you stand on one leg — it is a bilateral, mid-range, braced lift with a bar fixing your torso. My own tell is yoga: my warrior 3 transition into one-leg tree is genuinely sub par, and I do yoga daily. Lifting heavy makes you strong. It does not make you steady on one foot. For the size and strength side of the same session, see building bigger legs.

Single-leg work only happens if it has a slot in the week, and it is the first thing cut when you're rushed. That's a programming problem. The Momentum Method is the four-day split I run; its template leg day is built on bilateral lifts, so put a split squat first in that session, where it can't get dropped. Then retest every few months: barefoot, arms crossed, both legs, best of three. Two minutes, and it reads something no barbell is asking about.