Constraints: Changing Movement Without Telling Someone How to Move
A constraint is a boundary you put on a task so the athlete has to solve a movement problem instead of being handed the answer. You change the conditions so the old stroke stops working and a better one has to emerge. That’s the whole idea, and it rests on a specific theory about where movement comes from.
Where movement comes from
The constraints-led approach grows out of ecological dynamics, which treats the moving body as a self-organizing system rather than a computer running a stored program. The claim is that a movement pattern isn’t retrieved from memory and sent out as commands. It emerges, in the moment, from the interaction of three things. This taxonomy:
1. Individual constraints. The performer. Structural ones change slowly: height, limb length, wingspan, strength, mass. Functional ones change fast: fatigue, anxiety, attention, motivation. A swimmer with long forearms and one with short forearms will not, and should not, hold an identical catch.
2. Environmental constraints. Everything physical and external. Gravity, water density and temperature, the lane, the light. Also sociocultural: what the group expects, who’s watching. Mostly outside your control in the pool, with some exceptions.
3. Task constraints. The goal, the rules, the equipment, the space, the information available, the boundaries you set. This is the category a coach actually manipulates. Most coaching-by-constraint is task-constraint design.
Change any one of the three and the movement reorganizes, because the solution the body lands on is specific to the combination of all three. That’s the mechanism. You’re not installing a technique. You’re changing the problem so a different technique becomes the natural answer.
Why a constraint can beat an instruction
Two mechanistic reasons, both connecting to things I’ve written about before.
First, attention. Instructions usually point attention at the body. “High elbow, rotate the hips” is an internal focus, and internal focus tends to disrupt automatic control (the constrained action hypothesis). A constraint changes what the water is doing to the athlete and lets them feel their way to the solution, which keeps attention external. Researchers have argued a constraint can change technique in skilled athletes more effectively than attentional cues alone, because it forces exploration of the movement space instead of conscious control of one joint. I’d treat that as a supported argument, not a closed case.
Second, the attractor idea. A habitual stroke is a stable state. The system keeps falling back into it because it’s deep and familiar, even when it’s not the fastest option available. Telling someone to change rarely moves them out of a deep habit. Perturbing the system, changing the task so the old pattern literally cannot succeed, is what forces the reorganization. A paddle balanced on the head doesn’t tell the swimmer to stop lifting it. It makes head-lifting fail. The correction is in the physics, not the instruction.
A good constraint makes the wrong movement stop working, then lets the athlete find the right one on their own. What they find tends to fit their own body and tends to stick, because they built it instead of copying it.
What the evidence says
I’ll be straight about the state of it, because this approach gets sold harder than the data supports.
The theory is coherent and well grounded. The mechanism, self-organization under constraints and exploration of the movement landscape, lines up with a lot of what we know about coordination and attention. And there are positive studies. Constraint manipulations reliably change movement patterns, which is not in question. Some studies show learning and motivation benefits. A physical education study on hurdling reported higher intrinsic motivation and autonomy under a constraints-led lesson than a traditional one. A baseball study reported that a constraints-led group improved opposite-field hitting more than a group given prescriptive instruction.
But the hard proof that it beats good traditional coaching for skill retention and transfer is thinner than the enthusiasm around it. A 2018 systematic review of constraints-led training in interceptive sports said it plainly: the benefits appear obvious, but the evidence is not clear, and the quality of the underlying studies was often low. Most of the work is short-term, task-specific, and not always well controlled. So: strong theory, promising and growing evidence, not a settled case. Treat anyone claiming it’s simply superior with some caution, me included.
What this looks like in swimming
Task constraints you can actually set. Examples, not a program.
• Equipment that forces a position. A paddle on the head to stop head-lifting. A snorkel to remove breathing from the problem so you can isolate something else
• Numerical limits. A stroke-count cap per length forces distance per stroke. A target time forces pace. A tempo trainer forces rate.
• Space and timing. Swimming to a marker, no-breath zones into and out of the wall, underwater-distance targets off each wall.
• Rules that make the old pattern impossible rather than just discouraged. This is the core design principle. A good constraint removes the escape route. If the athlete can still cheat the intended solution, the constraint is too loose.
The design skill is picking a constraint that makes exactly one thing hard, the thing you want to change, without breaking the rest of the stroke. Too tight and you distort everything. Too loose and they find a way around it and nothing changes.
