3 mistakes that destroy motor learning (and how to fix them)

1/4h LabO #238: Watch the episode on YouTube

In January, I stopped producing so I could rebuild my motor learning course from scratch. By the end of the month: almost 200 pages and a few certainties gone. Here are the three mistakes that make people repeat without ever learning, and three tests to spot them.

Why I stopped everything to rework motor learning

In January, I did something odd: I stopped producing. No new content, no new training courses, nothing. I just picked up a course, the motor learning course of level 2 LabO RNP. Taken apart entirely, rebuilt from zero. By the end of the month, I had almost 200 pages, something dense but structured. And above all, I had lost a few certainties. Things you hear everywhere in the field, that you've taken for granted for years, and that vanished along the way.

One observation struck me. Between what we call "learning" in the field, whether you're a coach, a physio or you work with kids, and what really changes in a lasting way, there is a chasm. And you find that chasm everywhere: with the strength and conditioning coach and their athletes, with the physio who wants to stabilize a movement pattern, with the educator teaching a child a movement. Same gap, same traps, same mistakes. Always the same one: repeating without learning.

Hence the uncomfortable question that opens this article, the one I suggest you keep in mind to the end. Do your sessions really produce learning, or just an improvement that collapses as soon as the context shifts?

I'm handing you three takeaways. Three filters that everyone who guides movement should keep in their pocket. A session that doesn't pass these three filters, no matter how beautiful, intelligent or well built it is, does not produce learning.

What is motor learning (and why we confuse it with performance)

Take the simplest working definition possible. Motor learning is a reorganization of the system that holds when the context shifts. In the 200 pages that served to rebuild the course, one sentence keeps coming back: you only learn what withstands a change of context. That's all. You change the context and it holds, there may have been learning. You change the context and it collapses, you have only installed something locally.

The problem is that almost all of us run on the same implicit logic: if performance goes up during the session, learning has happened. The athlete does better at the end than at the start, the patient moves better at the end of the session, the child succeeds at the exercise, so they have learned. Not necessarily.

Immediate performance and learning are two distinct things. Performance can climb because you guide well with your feedback, because the environment doesn't change, because the task is simple, or just because the system has settled into a solution. Learning, for its part, appears only when the system is forced to reorganize, when it can no longer do otherwise. Keep this performance / learning distinction at hand, because the three mistakes that follow all stem from it.

Mistake #1: taking immediate performance for motor learning

First takeaway, and it stings a little: tiring a system does not force it to learn. In the field, we take the improvement visible during the session for a lasting change. But performance can go up for a thousand reasons unrelated to learning: you guide, the environment is stable, the task is simple, the system has found a local solution and clings to it.

Learning plays out at another level. It is triggered only when the system is forced to reorganize. As long as it gets by without changing, it doesn't change. That's why you can have a session that "works" from start to finish, and zero retention the next day.

The truth test is the change of context. You change the context, it holds: maybe learning. You change the context, it collapses: nothing built, just something set down locally. This filter applies to the strength and conditioning coach with their athletes, to the physio who wants to stabilize a pattern, to the educator teaching a child a movement. The same everywhere.

The four traps that look like learning

In the field, four situations pass for learning without being it:

- systems that repeat without learning ;
- systems that automate compensations ;
- systems that stabilize fragile solutions ;
- systems that progress while becoming less adaptable.

Seen from the outside, these four look alike, and all of them disguise themselves as progress. In the coach, the strength and conditioning coach, the physio, the teacher: exactly the same. The last one is the sneakiest, because progress that makes the system less adaptable really does look like a win in the moment.

The test to run at the end of the session

Concretely, at the end of your next session, one single question: what can this person do now that they would still be able to do if I changed one key constraint?

And you change a constraint, a real one. You change the rhythm. You cut the feedback, in other words you stop the guidance. You modify the environment. You remove your presence. Then you watch. If the skill you were working on disappears as soon as the constraint shifts, there was no learning, just a local installation.

Mistake #2: explaining too much prevents calibration

Second takeaway, and I know it ruffles a few feathers: the more you explain, the more you prevent calibration. We all believe that learning goes through understanding, good instructions, precise feedback. The famous "if you explain well, it sinks in". The opposite is often what happens.

The system doesn't learn when you hand it the solution. It learns when it detects a mismatch, that is, a gap between what it predicts and what it receives. The chain is simple: no mismatch, no surprise; no surprise, no recalibration; no recalibration, no plasticity. Nothing.

This is where excessive guidance becomes vicious. Throwing out lots of instructions is not inefficient in itself. The problem is that it short-circuits the very trigger of learning: you give the answer before the system has had time to search. The result: the session looks great, the person does well, the patient reproduces, the child succeeds, and retention is zero. You have built a dependence on guidance, and you are the one who created it.

How to recreate mismatch

At your next session, keep it simple. Cut 50% of your corrective feedback. Wait three repetitions before giving your comment. Deliberately space out your feedback. And let the error become perceptible, so that the system feels it by itself.

Then observe. If the person can't correct themselves alone, look closely at what is going on: a problem of dependence on guidance and feedback, not a problem of motivation. Your role as a guide is first to create learning situations, not to supply the answer. Charlie Francis said it very well: "Let the drill make the speech." Let the exercise speak in your place.

Mistake #3: a clean movement can hide non-learning

Third takeaway, the most counterintuitive. You watch a movement, you find it clean, fluid, nice, you validate it. You may have just validated exactly the opposite of what you wanted. Because visually, two movements can be identical while functionally, they are opposites.

Take both sides. The first movement is guided by a calibrated perception: the system picks up the relevant information and it adapts. The second is held by costly conscious control, or by a compensation. To the eye, it looks the same. Under constraint, the second collapses.

The reason is mechanical: we don't learn movements, we learn to exploit information. The movement is only a consequence, what emerges from the coupling between the organism, the task and the environment. I move to perceive, I perceive to move. If the person hasn't learned to perceive, they have learned nothing, even when the movement is beautiful.

Change a single parameter

After an exercise that you judge technically successful, something clean and validated, change just one parameter. Only one, in the individual-task-environment matrix: uncertainty, timing, pressure, or spatial constraint.

Then watch. If the system explores, adjusts, reorganizes: learning is under way. If it freezes, compensates or regresses sharply: the solution was stabilized, not learned. That small parameter change tells you which of the two movements you really had in front of you.

What really matters: retention and transfer (and the place of variability)

At bottom, two things matter: retention and transfer. Does it hold tomorrow, does it hold elsewhere. The rest is decor.

This month of work did not teach me to run more complex sessions. It taught me to run them more precisely. Less control, fewer explanations, less premature validation. In exchange, you go looking for more conditions, more relevant variability, and more respect for how a system really learns.

Variability is not there to complicate things. It is there because a system only reveals what it has really learned when confronted with changing conditions. It is what serves transfer, and it is exactly what the three tests from the takeaways check: change a constraint, recreate mismatch, modify a parameter. Three ways of asking the same question.

The only question to ask at the end of a session

Today, I no longer ask myself "was the session a success?". I ask myself: will the system still be capable tomorrow, without me, in another context?

This is the continuum that this month of work allowed me to lay down. Replace the question of the moment with the question of duration and context, and most of your session decisions change by themselves.

Because at bottom, the body often already knows how to do it. What learning has to rebuild is the ability to perceive, decide and adapt. Those were my three takeaways. Do what you want with them.

FAQ

What is motor learning?

A lasting reorganization of the system, one that withstands a change of context. The benchmark is simple: you only learn what withstands a change of context. If you change the context and the skill holds, there was learning; if it collapses, you had only installed a local solution.

What is the difference between performance and motor learning?

Immediate performance can go up because you guide well, because the environment is stable, because the task is simple, or because the system has settled into a solution. Learning, for its part, appears only when the system is obliged to reorganize. Performance that improves during the session therefore does not prove that learning has happened.

Should you cut feedback to learn better?

Above all you should stop short-circuiting the system's search. Excessive guidance gives the solution before the system has searched, which removes the mismatch and therefore the recalibration. In practice: cut 50% of your corrective feedback, wait three repetitions before correcting, space out your comments and let the error become perceptible.

Does a clean movement mean learning has happened?

No. Two movements that are identical to the eye can be functionally opposite: one guided by a calibrated perception, the other held by costly conscious control or a compensation. To the eye it is the same, but under constraint the second collapses. We learn to exploit information, the movement is only its consequence.

How can you check that motor learning has really taken place?

Change a key constraint and observe. Modify the rhythm, cut the feedback, change the environment, remove your presence, or change a single parameter after a successful exercise (uncertainty, timing, pressure, spatial constraint). If the system explores and adjusts, learning is under way. If it freezes, compensates or regresses, the solution has only been stabilized.

What criticisms can be made of the approach based on guidance and verbal instructions?

Three pitfalls come up again and again. Excessive guidance creates a dependence on feedback and manufactures performance without retention. Over-explaining removes the mismatch the system needs to recalibrate. And premature validation of a "clean" movement makes you validate a fragile solution instead of real learning. The fix fits in one sentence from Charlie Francis: "Let the drill make the speech."

Related articles

• Motor learning: the myth of coordination and 5 reflexes that slow progress

• Variability and motor learning: why repeating the same movement caps your progress

• Motor learning: the link between technique and physical qualities

• Constraints-based motor learning: why your body tries to survive before it performs

Discover the Neuro-Postural Reprogramming training