Can the chair challenge really tell men from women using mechanics?

A centre-of-mass model shows the chair challenge barely separates men from women, and shoes or momentum matter more.

Direct answer

Mechanics explains why the chair challenge is hard, but not why it should be a sex test. Modelling the body as an articulated chain of rigid segments, the paper finds a static limit of 2.36 foot lengths for its male reference model and 2.49 for its female one, just short of the customary 2.5 — and once real body proportions from 6068 US Army survey participants are used, the sex gap shrinks to 0.03–0.05 foot lengths, about half to one European shoe size.

1sources cited

This article was generated with WisPaper-powered search and paper analysis.

The question is where the centre of mass can be

The chair challenge asks a person to stand several foot lengths from a wall, bend forward until the head touches the wall, lift a chair to the chest and stand up without moving the feet or pushing off the wall. Viral videos present the outcome as a sex test: women succeed, men fail, supposedly because women carry a lower centre of mass or because men's larger feet place them further back when the distance is paced in foot lengths.

The paper treats this as a problem in static equilibrium rather than a claim to be taken at face value. A slow rise is possible only if the combined centre of mass of body and chair can be brought directly above the base of support, between heel and toe tip. Because the body is articulated, the relevant question is not where the centre of mass sits in a standing posture but how far back the joints and grip can move it while the head still touches the wall.

The authors therefore build a planar model of the body as a chain of rigid segments with standard biomechanical parameters, then search the postures allowed by the rules for the one that places the combined centre of mass furthest back. No one was asked to lift a chair: every participant in the study is either simulated or was measured by the US Army for other purposes.

The best posture is legs straight and trunk horizontal

At the customary toe–wall distance of 2.5 foot lengths, the set of attainable centres of mass is small and lies mostly in front of the toes: from +3 to +29 cm for the male reference model and from 0 to +27 cm for the female one. The search converges on the same posture in both cases — legs straight and vertical, trunk horizontal, head in line with the trunk, and the chair held close under the abdomen above the feet.

That posture follows from the geometry. Taking the hip joint as reference, a horizontal trunk with the head in line puts the top of the head one trunk-plus-head length in front of the hip. Since the head must touch the wall, the hip then sits exactly that length from the wall; a more upright trunk or a bent neck would bring the hip closer. In this posture the combined centre of mass lies a horizontal distance in front of the hip, so the toes can be at most the trunk-plus-head length minus that distance from the wall.

This gives the paper's rule of thumb: the largest toe–wall distance for a static rise is approximately the trunk-plus-head length minus the distance from hip to combined centre of mass. With the scaled reference segments, that is 86.1 − 21.1 = 65.0 cm, or 2.36 foot lengths, for the man and 82.7 − 20.0 = 62.7 cm, or 2.50 foot lengths, for the woman. The full posture search, which also lets ankles, knees and neck move, returns 65 cm and 62 cm — within 0.01 foot lengths of the rule of thumb.

Centimetres and foot lengths rank the sexes differently

The unit decides who appears to have more room. In centimetres the male reference model has the larger limit, 65 cm against 62 cm, because his trunk and head are longer. Only when the distance is expressed in the person's own foot lengths does the female model come out ahead, at 2.49 against 2.36. At two foot lengths both succeed; at three, neither does.

The paper is explicit that the limit is set mainly by trunk-plus-head length relative to foot length, not by the height of the centre of mass. The horizontal distance from hip to combined centre of mass in the bent posture is about 12% of stature in both reference models — 11.7% for the man and 12.0% for the woman — so if anything it works to the woman's disadvantage. The popular "lower centre of mass" explanation is therefore true of a standing posture and irrelevant to the bent one the trick requires.

The geometric limit also assumes the calf muscles can hold the whole body balanced with the centre of mass above the toe tip. The required ankle torque is about 180 N m for the man and 145 N m for the woman, roughly 60% of the maximum isometric plantar-flexion torque of young adults. Moving the edge of support back to the ball of the foot, a quarter foot length behind the toe tip, lowers the practical limit to 2.05 and 2.19 foot lengths. The paper's reading is that the problem is balance, not strength.

Real body proportions shrink the sex gap

The reference segments come from a dataset based on 100 men but only 15 women, so the authors replace the two ratios that matter most — trunk-plus-head length and foot length relative to stature — with values from the 2012 US Army anthropometric survey ANSUR II: 4082 men and 1986 women with a mean age of 30 years. That survey contains external body dimensions and body mass but no segment masses, so the mass distribution still comes from the reference dataset.

In ANSUR II the trunk plus head is not relatively longer in women: 48.1% of stature against 48.7% in men. Women's feet are relatively smaller, 15.1% against 15.4% of stature. Running the full model with these proportions gives limits of 2.37 and 2.40 foot lengths for all participants, and 2.40 and 2.45 for the subgroup of European ancestry. The sex difference drops from 0.13 to 0.03–0.05 foot lengths, and what remains comes from relative foot size.

In shoe-size terms, one European size corresponds to 2/3 cm of last length, so pacing 2.5 feet moves the start position about 1.7 cm per size. The reference-model difference of 0.13 foot lengths is about two shoe sizes; the survey difference is half to one. The standard deviation of foot length within each sex in the survey is 1.2–1.3 cm, about two sizes — already larger than the average gap between the sexes.

Shoes, age and momentum outweigh the sex difference

A sensitivity table changes one assumption at a time for both reference models. Chair mass, ankle, knee and neck ranges, hip flexibility and the position of the ankle joint shift the limit by at most 0.03 foot lengths; holding the chair tightly against the chest costs 0.06. A ±3% change in trunk-plus-head length and a ±5% change in foot length matter most, while stature itself does not, because everything scales with it. The sex difference persists in every variant.

Several factors outside the baseline model matter more than that residual difference. Breast mass, computed with the simplified fixed best posture, shifts the limit by at most 1.6 cm in the unfavourable direction. Moving 3–5 kg of fat from hips to belly costs 0.03–0.05 foot lengths. Age matters more: the loss of about 5 cm of stature in women and 3 cm in men by age 70, mostly in the spine and hence in trunk-plus-head length, lowers the limit by 0.13 and 0.07 foot lengths.

Shoes matter most. A shoe is roughly 2.5 cm longer than the foot, so pacing 2.5 shoe lengths places the toes about 7.5 cm further from the wall — 0.27–0.30 foot lengths, more than twice the largest sex difference the authors found. With shoes, neither average adult passes the test statically. Because the extra length is about the same for everyone, shoes penalise small feet relatively more and slightly reduce the sex difference.

A weak classifier of individuals, not of averages

Judging the trick as a sex test requires the spread between individuals, not just the averages. Using the calibrated rule of thumb with each survey participant's own stature, trochanterion height and foot length, the authors compute an individual static limit for all 6068 participants. The spread within each sex, with a standard deviation of about 0.12 foot lengths or 3 cm, is several times larger than the difference between the sexes.

At the customary 2.5 foot lengths, 13% of men and 22% of women would pass statically — 16% and 33% in the European-ancestry subgroup, and 10% and 46% with the reference proportions. For equal numbers of men and women, predicting "woman" after a success and "man" after a failure is right in 54% of cases, rising to 58% and 68% for the other datasets; even the best possible toe–wall distance raises this only to 56–72%. A fair coin scores 50%. Stature or foot length alone, measured with a tape and compared with the best threshold, identifies sex correctly in 83% of cases.

At three foot lengths, a distance also found in the instructions, not a single one of the 6068 participants passes statically; the largest individual limit is 2.82 foot lengths. The paper's own explanation for why so many videos show women succeeding is offered as speculation: momentum, since a brisk extension of hips and back can push the head against the wall and drive the body backwards; loosely paced distances, sometimes in shoes; and a selection effect in which videos matching the expected pattern are more likely to be recorded and shared. A linearised inverted-pendulum estimate shows that a small backward velocity — about 0.18 m/s, or an impulse of roughly 15 N s, for a 5 cm offset at 0.80 m height and 88 kg — suffices to carry the centre of mass to the toe line.

What the model does not cover

The model is two-dimensional with a rigid trunk, so a curved spine would enlarge the attainable set somewhat. Mass distribution is taken from one reference dataset for all individuals, and the proportions used are those of Russian students and US soldiers; other populations were not examined. The individual limits rely on the calibrated rule of thumb rather than the full posture search.

Most importantly, the model says nothing about how people actually perform the trick. It states what is statically possible, not what happens when a real person with real balance, footwear and momentum attempts it. The momentum argument in the discussion is explicitly speculative, and the authors propose that a systematic analysis of existing videos — with two independent raters recording toe–wall distance, footwear, whether the person rose slowly or with a jerk, and apparent sex — could separate the effects.

The tape measure beats the party trick

The paper's own summary is that the chair challenge is a fine demonstration of static equilibrium and a lousy sex predictor. The comparison that carries this judgement is not between men and women but between three classifiers: the challenge outcome at 54–68% correct for equal numbers of men and women, a coin at 50%, and a tape measure at 83%.

The reason is structural rather than a matter of measurement precision. The limit is set by trunk-plus-head length relative to foot length, and the within-sex spread of about 0.12 foot lengths is several times the between-sex difference of 0.03–0.05 foot lengths in the survey data. A real average difference does not make a usable individual test when the distributions overlap that heavily.

What the paper does establish is a priority order for anyone thinking about the trick. The unit matters: centimetres favour the male model, own foot lengths favour the female one. Footwear matters more than sex, adding 0.27–0.30 foot lengths against a residual gap of half to one shoe size. Age and momentum matter more than the residual sex difference too. The chair challenge is worth teaching as an articulated-body equilibrium problem with a scaling lesson built in; as a way to tell men from women, it loses to a tape measure by a wide margin.

Sources used in this answer

1

The chair challenge: can mechanics really tell men from women? A centre-of-mass analysis of a viral party trick

A static centre-of-mass model gives limits of 2.36 foot lengths for the male and 2.49 for the female reference model, but with ANSUR II proportions the sex difference shrinks to 0.03–0.05 foot lengths, and predicting sex from the outcome is right in only 54–68% of cases versus 83% for a tape measure.