A coach can tell players to switch play more often, or can make the pitch wider and award a point for reaching the far touchline. The second approach changes behaviour more reliably, and the reason is worth understanding.
Instructions compete with the game in front of the player
A player under pressure processes the situation he can see, and a remembered instruction from a meeting is a weaker input than an opponent two metres away.
Instructions therefore tend to be followed in the calm moments and abandoned in the ones that matter.
Constraints work differently because they change what the situation rewards, so the desired behaviour becomes the solution the player would have chosen anyway.
The constraint has to target the right variable
Pitch dimensions, player numbers, touch limits, scoring rules and time limits each push behaviour in a distinct direction.
A narrow pitch produces tight combinations and quick feet; a wide one produces switches and longer passing. Neither is generally better, and each teaches something the other cannot.
Designing a session therefore begins with the behaviour wanted rather than with a drill, and the drill is derived from it.
Scoring rules are the strongest lever
Players optimise for whatever wins the practice, so changing how points are earned changes decisions immediately and without any explanation.
Awarding extra value for a goal scored after a switch, or for a possession that breaks a line, redirects the whole session in one sentence.
The effect fades if the same rule is used repeatedly, because players learn to exploit the scoring rather than to play the behaviour, so rules are rotated.
Over-constraining produces a different game
A practice with a touch limit, a minimum pass count, restricted zones and a time constraint stops resembling football at all.
Players then learn to solve the exercise rather than the game, and the behaviour does not transfer to a match where none of those rules apply.
The usual guidance is one primary constraint at a time, with the rest of the practice left as close to the real game as possible.
The coach's role shifts to observation
If the constraint is working, intervention is largely unnecessary, and the coaching consists of watching which players are finding the solution and which are not.
Where nobody finds it, the constraint is wrong rather than the players, and the adjustment is to the design rather than to the effort level.
That reframing is the practical value of the approach: a session that is not producing the behaviour is a design problem the coach can fix.