Restraint

Restraint means stopping an opponent's pawn — or an opponent's plan more generally — from ever advancing, rather than dealing with the consequences once it does. The clearest version is controlling the square directly in front of an enemy pawn so firmly that the pawn can never safely move there, which freezes not just that pawn but often the entire structure around it, since a locked pawn chain can't be reorganized once its lead pawn is permanently restrained.

Controlling the square in front, not just the pawn itself

A pawn is easy to attack directly but hard to permanently stop, since it can often just be defended by another piece or pawn. Controlling the square in front of it is a more durable form of pressure: as long as that square is covered by enough of your pieces or pawns, the enemy pawn simply can't advance there without being captured for nothing, which achieves the same practical result as winning the pawn — it's frozen — without needing to actually win any material at all.

A blockaded pawn stops defending its own structure

Restraining one pawn often has effects well beyond that single pawn, because pawns that were counting on it to advance and support a later break can no longer do so. A pawn chain built around eventually pushing a specific pawn forward loses its entire long-term plan once that pawn is permanently restrained, which is why targeting the one pawn whose advance the opponent's whole structure depends on is usually a far more efficient use of restraint than trying to freeze every pawn individually.

Restraint applies to plans, not only to pawns

The same underlying idea — control the square or line something needs before it gets there, rather than reacting once it arrives — extends to pieces and whole plans as well. Occupying the one square a knight would need to reach before it can maneuver into an active outpost, or controlling the file a rook wants before it gets doubled there, is restraint in the same sense as blockading a pawn, just aimed at a different kind of target.

The piece assigned to restrain has to earn its keep

A piece parked purely to prevent a single pawn advance can become a liability of its own if it has nothing else to do and the opponent finds a way to attack it directly, forcing it away and releasing exactly the advance it was there to stop. The strongest restraining setups use a piece that's doing something useful even while it restrains — controlling the key square as a side effect of also being well placed for its own purposes — rather than one whose entire job is sitting still.

When letting the pawn advance is actually fine

Restraint isn't automatically correct just because a pawn advance is available to stop. If the pawn's advance would create its own weaknesses — leaving squares behind it uncoverable, or simply putting the pawn somewhere easier to attack than where it started — letting it move forward and dealing with the resulting weaknesses can be the more efficient plan than tying a piece down to prevent an advance that would have backfired on the opponent anyway.

Blockading a pawn that's already passed the point of restraint is covered separately on passed pawns, and restraint as a component of a broader habit is part of prophylaxis.