Your neighbour's child learned at four. Yours is six, the bike with stabilisers has been in the garage for two summers, and every time you try to take them off, your child gets off and says they do not feel like it. You stand there with a spanner in your hand: is it too early? too late? is it that your child cannot get there?
Here are the two things that really matter, because they change everything else.
The first: the age at which a child learns depends largely on which bike you put them on, much more than on how "coordinated" they are. The second, which is the uncomfortable part: stabilisers train everything except the thing that has to be learned. They are not dangerous and they are not a fault: they simply keep balance, the only truly difficult part, out of the exercise.
The real ages, and how wide they are
The clearest figure comes from a survey conducted in Portugal, with 2,005 responses collected from adults and parents about the route by which they had learned to pedal. Children who had started with a balance bike (a bike without pedals, the one the child propels with their feet) were pedalling on their own at an average of 4.16 years (standard deviation 1.34). Those who had moved on from a bike with stabilisers got there at 5.97 years (SD 2.16). Those who had used only a traditional bike, with no intermediate step, at 7.27 years (SD 3.74). The differences were statistically significant (p < 0.001), and the authors add a practical detail that is not trivial: the balance bike works better when it is the first bike, not when it arrives halfway through to make up for lost ground.
Before reading those numbers as a score, look at the standard deviations. They are enormous. With stabilisers, an average plus or minus one standard deviation means a range from 3.8 to 8.1 years; with only a traditional bike it runs from 3.5 to 11 years. A child who learns at seven is not behind: they are within the normal range, and are probably just paying the price of the route by which the bike was offered to them.
An independent Dutch study asked the same question of 173 parents of children aged 4 to 6 and reached the same direction: children who had used a balance bike had started practising earlier, had practised for less time and pedalled on their own sooner than those who had used stabilisers or both. The authors attribute the advantage to the fact that the bike without pedals puts postural control under demand from the very first minute.
One limitation needs saying clearly: both of these studies are retrospective questionnaires, not experiments. They reconstruct from memory, and parents who choose a balance bike may differ in other ways from those who choose stabilisers. They show a strong and consistent association across two countries, not an experimental demonstration.
Why stabilisers do not teach balance
The reason is mechanical, and once you see it you cannot unsee it.
Riding a bike is made of two things the child has to learn together: pedalling and staying balanced on two wheels, constantly correcting the lean with the steering and the body. With stabilisers, the second one simply is not practised: the bike stays upright on its own, so the error to correct almost never arises and the correction is never learned. The child becomes very good at turning the pedals and never touches the real problem. Then you remove the stabilisers and the real problem arrives all at once, at an age when the child has already learned that the bike stays up by itself.
The bike without pedals does the opposite: it removes the pedalling and leaves only balance, with the feet on the ground as a safety net that is always available. The child can make a hundred mistakes without falling.
Here there is also a measurement, not just reasoning. A Portuguese group followed 23 children of about 6 years old divided into two training routes (12 with the bike without pedals, 11 with stabilisers), with four inertial sensors on head, trunk, frame and handlebars, measured before the sessions, straight after and two months after the switch to a normal bike. The result: the bike without pedals produces greater and functional movement variability, meaning a richer repertoire of adjustments, which then translates into more adaptive responses when the child moves to a real bike. It is a small study, but it points in the same direction as the questionnaires and explains the why.
Another work by the same group, on 12 children followed over six sessions with the bike without pedals, filmed and classified seven different movement patterns. After training, both the average number of patterns and the average and maximum speed had increased, and, a detail you will need shortly, as the speed rose, the feet touched the ground less and less and longer flight phases appeared. Speed, in other words, is not an accessory: it is the lever that brings balance out.
Finally, there is a figure that concerns the general picture. Among 97 nursery school children assessed both on fundamental motor skills and on balance bike ability (two timed courses, one straight and one curved), the ability on the bike without pedals was significantly correlated with all three sub-components of motor skills (locomotion, object control and stability), but could not be reduced to any one of them. The authors' conclusion is that riding a bike is a fundamental motor skill in its own right. This is the technical way of saying what every parent knows: no other game teaches it in your place. If you are interested in the overall picture of how these skills develop, we discuss it in Motor games for children: what to offer age by age.
What to look at before trying
There is no clinical checklist, and none of the studies we cite sets a threshold to pass before getting on the saddle. There are, however, three things you can observe in the yard that tell you where the child is.
Standing on one foot for a few seconds. It is the simplest way to see how postural control is organising itself on one support. But here one thing needs saying that rebalances expectations: a study using force platforms and motion capture on 48 children aged 3 to 10 compared with 11 young adults found that postural control in single-leg stance matures around age 9 in the acceleration and stable balance phases, and that at 10 it is not yet at adult level in the deceleration phase. That is: balance is still being fully built in the very years when children learn to ride. Do not wait for it to be finished, because it never will be.
Keeping the trunk steady while the legs do something else. Watch the child running downhill, kicking a moving ball, going down the stairs alternating feet: if the upper body stays upright and the head stays oriented forward, the piece is there. The milestones of these movements are in When a child starts to jump, run and climb stairs.
Being able to brake. This is the part almost everyone postpones and it is the one that causes the most harm. In an Austrian study of primary school children assessed with practical tests, overall skills were good, but the three things that proved most difficult were exactly turning left, slalom and emergency braking. The same study showed a second finding that matters more for you than for the child: there were significant discrepancies between the objective assessment and self-assessment, with many children, and girls in particular, overestimating their abilities. And among the factors examined (level of physical activity, school performance, training), the only one significantly associated with bike skills was specific bike training.
The method that works in practice
It is not complicated: it is only counterintuitive compared with how we were taught.
1. Remove the pedals rather than buying something. On most bikes the pedals come off with a spanner. Bike without pedals, saddle lowered until the child, seated, has the whole sole of both feet flat on the ground. The same criterion applies if you buy a real balance bike.
2. Find a lawn with a very slight slope. Grass forgives falls, and the slope gives speed without the child having to push. A descent you barely perceive is enough.
3. Three stages, in order: walk, push, glide. First they walk while sitting on the saddle. Then long strides and pushes. Then, when the speed rises, the feet lift off by themselves and stay up for a second, then three, then ten. Do not ask them to lift their feet: ask them to go faster. That is exactly the mechanism measured in the study of movement patterns: speed is the lever, and the flight phases are the consequence.
4. Do not hold the saddle and do not hold the handlebars. If you hold the bike, you are doing the balancing for them: you have become a human stabiliser. If you really must, hold their shoulders or hips, very lightly, and let go early.
5. Eyes far ahead. "Look at the gate at the end", not "look where the wheels are going". A downward gaze makes the steering jerky.
6. Brakes before pedals. Teach them to stop before the pedals go back on: a braking game, stopping on a chosen spot, repeating until it is automatic. Given the difficulty with emergency braking found in the Austrian study, this is the part to train most.
7. Short, frequent sessions. Fifteen or twenty minutes, and you stop while they are still happy: there are no studies on the ideal duration, it is common sense. On frequency, however, there is a figure, small and in another population: in a study of 11 children and adolescents aged 9 to 16 with neurodevelopmental disorders, followed after a bike-learning camp, the frequency of practice was correlated both with cycling skills (0.72; p = 0.01) and with real participation on the bike (0.86; p < 0.001). It is a correlation over a few participants, not proof that the number of sessions causes the result.
8. Put the pedals back when they glide. Not before. When they can stay balanced for ten metres with their feet up, the pedalling arrives almost by itself: it is the only piece the stabilisers had really taught.
The helmet: here the numbers are solid
On many aspects of motor development the evidence is weak, and we say so. On this one, no.
A systematic review with meta-analysis selected 43 studies, analysed 40, and drew on data from over 64,000 injured cyclists. Among cyclists involved in a fall or an accident, helmet use was associated with reduced odds of head injury (OR 0.49; 95% CI: 0.42-0.57), serious head injury (OR 0.31; 95% CI: 0.25-0.37), facial injury (OR 0.67; 95% CI: 0.56-0.81) and fatal head injury (OR 0.35; 95% CI: 0.14-0.88). The protection was greatest precisely for serious or fatal injuries. For neck injuries, the objection heard most often, no clear association emerged (OR 0.96; 95% CI: 0.74-1.25), and the authors describe them as rare and not associated with helmet use. It should be read for what it is: a null result, not a demonstration that the helmet cannot have any effect on that site. The authors found neither temporal trends nor signs of publication bias.
An older Cochrane review, on five well-conducted case-control studies, reported reductions in the risk of head, brain and severe brain injury between 63% and 88%, with equivalent protection in motor vehicle accidents (69%) and in all other cases (68%), and a 65% reduction in upper and mid-face injuries. A more recent meta-analysis on maxillofacial injuries found the same sign (OR 0.682).
Honesty requires adding that none of these studies is randomised, which would not be possible anyway, and that people who wear a helmet might also ride differently from those who do not. There is also a second limitation to keep in mind in an article about children: these analyses concern cyclists of all ages, and the specific benefit in childhood is not estimated separately. That said, the consistency of the results across dozens of studies, populations and decades makes this one of the strongest pieces of evidence we have on the effectiveness of the bicycle helmet.
And a helmet alone is not enough. In the New Zealand study of 293 children aged 8 to 12, 41% of the 214 helmets checked by bike shop staff were judged unsafe and 41% of the 205 bikes unfit for the road; among the 127 children who cycled to school, 59% had an unfit bike or an unsafe helmet. The eleven- and twelve-year-olds were significantly better than the eight-to-ten-year-olds in the practical test (p < 0.0001), and the authors' conclusion deserves to be quoted as it stands: there is no precise age from which a child can be said to be "safe" alone on an open road.
When difficulty is a signal
In the great majority of cases a child who does not learn needs more opportunities with the right equipment, not an assessment: change the bike, change the method, and in three afternoons it happens.
It is, however, worth looking more closely when the difficulty is not only about the bike. The signal is not the single missing movement: it is the overall picture. Difficulties that also show in other everyday movements (tying shoelaces, using cutlery, catching a ball, writing, jumping), that last over time, that interfere with daily life and that started early: these are the markers used to think about developmental coordination disorder, which has dedicated international clinical practice recommendations, drawn up by a multidisciplinary panel with literature reviews and formal consensus. If this picture sounds familiar, the dedicated article is Clumsy child: when coordination is a disorder.
On how effective the interventions are in these cases, the honest answer is: encouraging but uncertain. A Cochrane review of 15 studies and 649 children aged 5 to 12 found, in a meta-analysis of six studies with 169 children, an advantage of task-oriented interventions over no intervention (mean difference -3.63; 95% CI: -5.88 to -1.39; p = 0.002), but with very low-quality evidence; in a second meta-analysis of only two randomised studies and 51 children the effect was not significant. The authors write that they have very little confidence in the estimate. On safety: two studies explicitly reported no adverse effects, and the authors of nine others confirmed by correspondence that no injuries had occurred; for the remainder the information is not reported.
There is, however, one thing that is known: learning to ride remains possible even when it is difficult. In a pilot study of 15 children and adolescents with autism spectrum disorder aged 7 to 16, after a one-week camp (five 75-minute sessions, with an individualised, task-oriented programme) all of them showed skill acquisition on the bike and nine pedalled on their own for at least about twenty metres. It is a pilot study with no control group, and it should be read for what it is: it does not demonstrate the effectiveness of the method, it shows that the goal is not out of reach.
Red flags
There are, on the other hand, signals that do not belong to this picture and that should be looked at straight away, and by a doctor:
- use of only one side of the body, or a clear asymmetry between right and left;
- pain that appears with movement, or a swollen, warm joint;
- above all, a rapid change or the loss of a motor skill the child had already acquired.
This is not extra caution. The international recommendations on developmental coordination disorder apply by definition to stable, non-progressive difficulties, and they list among the elements to assess in the neurological examination precisely "a rapid change or deterioration of motor function"; the clinical examination serves to rule out other causes: cerebral palsy, hemiplegia, muscular dystrophy, juvenile arthritis, neuromuscular disorders. The same recommendations say that at that point the picture is no longer the one we are discussing here.
If you want an outside view
Sometimes it is enough for someone to watch the child move and tell you where the missing piece is: whether it is balance, fear or simply the wrong bike. With Ri-Hub Junior you start from a free twenty-minute introductory consultation by video call with a physiotherapist on the Italian professional register who works with children and adolescents: you show us your child walking, jumping, standing on one foot, and we think together about what to offer them. For children up to 10 years old sessions last 50 minutes and cost 59,90 €; from 11 to 18 years old the standard session at 29,90 € applies.
One limitation stated clearly: by video call, balance, coordination, fear and method can be worked on well. Complex neurological and neuromuscular conditions are neither assessed nor followed online: if the signals in the list above emerge, the reference point is the paediatrician.
On scooters the data say different things depending on age: Scooters and children.
Sources
- Learning to Cycle: From Training Wheels to Balance Bike: International Journal of Environmental Research and Public Health, 2022
- Mastering balance: The use of balance bicycles promotes the development of independent cycling: British Journal of Developmental Psychology, 2022
- Learning to Cycle: Why Is the Balance Bike More Efficient than the Bicycle with Training Wheels? The Lyapunov's Answer: Journal of Functional Morphology and Kinesiology, 2024
- Learning to Cycle: Is Velocity a Control Parameter for Children's Cycle Patterns on the Balance Bike? Children, 2022
- Quantifying cycling as a foundational movement skill in early childhood: Journal of Science and Medicine in Sport, 2020
- Development of postural control during single-leg standing in children aged 3-10 years: Gait & Posture, 2019
- On the Road to Safety: Examining Children's Cycling Skills and Physical Activity Levels: Children, 2024
- Bicycle injuries and helmet use: a systematic review and meta-analysis: International Journal of Epidemiology, 2017
- Helmets for preventing head and facial injuries in bicyclists: Cochrane Database of Systematic Reviews, 2000
- Effect of helmet use on maxillofacial injuries due to bicycle and scooter accidents: a systematic literature review and meta-analysis: International Journal of Oral and Maxillofacial Surgery, 2024
- Child cyclists: A study of factors affecting their safety: Journal of Paediatrics and Child Health, 2017
- International clinical practice recommendations on the definition, diagnosis, assessment, intervention, and psychosocial aspects of developmental coordination disorder: Developmental Medicine & Child Neurology, 2019
- Task-oriented interventions for children with developmental co-ordination disorder: Cochrane Database of Systematic Reviews, 2017
- Accelerating Motor Skill Acquisition for Bicycle Riding in Children with ASD: A Pilot Study: Journal of Autism and Developmental Disorders, 2020
- Promoting Cycling Skills and Participation with Two-Wheel Bike Riding for Children with Neurodevelopmental Disorders: Physical & Occupational Therapy in Pediatrics, 2024
Last source check: September 2026.
