1. RNP
  2. Motor Learning
  3. Error, uncertainty, variability
  4. Noise: When Variability Speaks but Says Nothing

Motor LearningSeries 4/13

SeriesError, uncertainty, variability4 of 13

Noise: When Variability Speaks but Says Nothing

Noise is the parasitic variability that teaches nothing, and it has to be told apart from usable signal. Learn to read one without confusing it with the other.

LabO RNP·5 min read·reviewed on

In short. Noise is the part of variability that carries no usable information. It is a parasitic fluctuation with no functional gain. Instead of exploring, it disorganizes. Signal teaches. Noise destabilizes.

You're watching an athlete repeat a movement. With every rep, the trajectory shifts a little. Your hand wants to correct it, and your eye wants to clean it up. This is the moment that matters. What you see wobbling under your fingers might be learning in progress. It might also be a system falling apart. The two look alike. Noise is what tells them apart.

Noise isn't the absence of variability, it's the empty version of it

For a long time, every fluctuation was filed under the same label: noise, an imperfection, an error to correct. In that logic, you chase repeatability, eliminate deviations and standardize execution. The problem is that a perfectly repeatable system stops learning, stops adapting, and becomes fragile.

The historical mistake was to treat all motor variability as noise. Modern thinking sees variability as the medium of exploration: it tests solutions and reveals how sensitive the system is to constraints. True noise is something else. It is variability spinning its wheels: lots of movement, zero information.

You can hear the difference as two sounds. An instrument being tuned produces variations that converge on a note. A badly tuned radio produces static that goes nowhere. On the surface, the agitation looks the same. The meaning is the opposite.

Telling usable signal from parasitic fluctuation

RNP draws a distinction your eye can't make on its own: functional variability versus noisy variability. The first is informative exploration and progressive adjustment. The second is sensory instability, overload and defense. A poorly calibrated system can produce a huge amount of variability without learning anything.

What counts isn't the amount of variability but its informational quality. If you can't read that difference, you'll handle every fluctuation the same way. Either you'll try to reduce them all, or you'll let them all run wild. You'll never support the variability that teaches while containing the variability that destabilizes.

This way of reading is a close cousin of how you read error. Not every error needs correcting, and not every variation needs reducing. In both cases, what you're tracking is whether usable information is present or absent.

The RNP training space, free for life

Get access to RNP training, no strings attached.

Welcome! Introduce yourself

SecureNo cardNo password

By continuing, you agree to our Terms of Use and our Privacy Policy.

Spotting noise in the field, under your fingers

Take two athletes. The first is in a learning phase. Their trajectories differ slightly from rep to rep, but overall efficiency holds. That's exploration. The second is fatigued and shows erratic variability with no gain at all. You see both of them wobble. In the first case, the system is exploring. In the second, it's collapsing. Treat them the same way and you get it wrong.

Kids show the same thing. One tries several ways to get over an obstacle, and their variability is rich and structured. Another, sensorily overwhelmed, keeps switching strategies without ever improving. One is learning. The other is surviving. Your job is to know which one is in front of you. Rehab works the same way. One client varies their foot placement slightly as they progress. Another sways because they lack sensory reference points. The observation looks the same, and the meaning is radically different.

On the ground, the test is as much about feel as about sight. Signal has a texture that converges and has a direction. Noise has a texture that slips, leaving you with a jittery sense of drift, with no heading.

Where noise comes from: a system that can no longer tolerate variation

Noise isn't fate. It's a state. A defensive system can show too much disorganized variability because it can no longer stabilize anything. A sensorily saturated system reads change as danger. In that case, forcing variability only reinforces rigidity or amplifies the noise.

That's why you can't simply decide to add variability. If the system can't tolerate it, you create overload, defensiveness and stiffening. Here, intra-individual variability is your indicator. Its absence points to rigidity, and its disorganized excess points to noise. The zone where the system learns lies between the two.

Sensory calibration comes first. Variability only helps a system that is available for it. As long as the sensorimotor loop registers danger where there is only variation, it will produce noise, not signal.

From noise to robustness: what the reading is for

Reducing noise has never meant eliminating variability. It means restoring the system's ability to turn fluctuation into information. Once the system regains that tolerance, the same variability that was producing noise becomes usable again. It then feeds dynamic stability: a solution that holds under perturbation and adjusts without falling apart.

That's where robustness comes from: the capacity to maintain function under fatigue, stress and change. Robustness is a product of learning, not of mechanical repetition. A system that has learned to read its own fluctuations as information doesn't collapse when the context shifts. Noise is the symptom of a system that doesn't have that reading yet, or has lost it.

“It's not the amount of variability that counts, but its informational quality.” A poorly calibrated system can produce a lot of variability without learning anything.

In one sentence: noise is variability that churns without saying anything. Your job isn't to silence it but to understand why the system can no longer make it usable.

Field test. Next time you see a movement wobble from rep to rep, don't correct it right away. Ask yourself one thing: is this variation converging toward something, or is it slipping in place? If it's converging, it's signal, so let it do its work. If it's slipping, it's noise. Look at fatigue, sensory saturation or defensiveness before you touch the movement.

Reading noise means you stop cleaning up what's learning. It's the first thing I want to pass on to the professionals I train.

Scope: a functional, non-medical approach. It complements health care and does not replace it.

€0Your free RNP PathsStart